Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

An evaluation of BASIC computer language as a prerequisite to university computer science Ellis, David Norman 1989

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-UBC_1989_A8 E35.pdf [ 10.75MB ]
Metadata
JSON: 831-1.0097772.json
JSON-LD: 831-1.0097772-ld.json
RDF/XML (Pretty): 831-1.0097772-rdf.xml
RDF/JSON: 831-1.0097772-rdf.json
Turtle: 831-1.0097772-turtle.txt
N-Triples: 831-1.0097772-rdf-ntriples.txt
Original Record: 831-1.0097772-source.json
Full Text
831-1.0097772-fulltext.txt
Citation
831-1.0097772.ris

Full Text

A N E V A L U A T I O N OF BASIC COMPUTER L A N G U A G E AS A PREREQUISITE TO UNIVERSITY COMPUTER SCIENCE by David Norman Ellis B. Sc., The University of British Columbia, 1968 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF ARTS in THE F A C U L T Y OF G R A D U A T E STUDIES MATHEMATICS & SCIENCE EDUCATION We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA September, 1989 © David Norman Ellis, 1989 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of MATH 1L epur-/frycA; The University of British Columbia Vancouver, Canada Date HS^ OCTT. II DE-6 (2/88) ABSTRACT This study was undertaken to determine how prior knowledge of BASIC computer language affects the achievement in introductory computer science courses at university. It looked at comparisons of achievement in introductory computer science courses at the University of British Columbia (U.B.C.) among groups who have learned BASIC, or who have learned other languages, or who have learned no languages. It investigated compar-isons of achievement among demographic factors: gender, age, Faculty, and major. Achievement differences in first year FORTRAN courses, in first year Pascal courses and in a second year Pascal course, among groups of students with different backgrounds were also examined. The study investigated the effect on achievement of "how well", "when", and "where" either BASIC or Pascal had been learned. Finally, it identified factors that are the best predictors of success in introductory computer science. A questionnaire was distributed in six introductory computer courses at U.B.C. during the school year 1985-1986. Marks were collected in these courses at the end of the year. After matching by student number a sample of 1194 students was analyzed using the analysis of covariance and the multiple linear regression routines of the Statistical Package for the Social Sciences (SPSS). It was found that students who have taken BASIC language do better in university introductory level computer science courses than those who have no prior knowledge of a computer language, and they do as well as those who have prior knowledge of some other computer language. It was discovered that students who have previously learned a com-puter language have better achievement than those who have not learned a language. Stu-dents with a knowledge of at least two languages have an even higher achievement. The order of learning BASIC language was observed not to be significant in subsequent com-puter science achievement. Males had higher achievement than females in the surveyed i i courses. The younger students tended to have higher achievement than the older students in these courses. Achievement differences were found among the Faculties involved. Stu-dents who were majoring in mathematics outperformed those who were non-mathematics majors within the Faculty of Arts. Achievement in the second year course did not appear to be dependent upon the computer language background prior to entering university. Stu-dents who were able to write complex BASIC programs outperformed others with a limited familiarity with BASIC language in introductory computer science courses. For students who had prior knowledge of Pascal language the level of how well the language had been learned did not appear to be a factor in introductory computer science achievement. The age at which BASIC or Pascal language was first learned was a critical factor. Those who first learned the language in the 13 to 18 age range outperformed others who first learned the language at an older age. The place where either BASIC or Pascal language had been learned did not appear to be a critical factor for achievement in introductory computer science. Only the overall year percent in all other courses taken and the variable as to "how well" BASIC language had been learned, proved to be significant factors in predicting success in introductory computer science. The writer concludes that the results of this study provide sufficient evidence to support the argument that BASIC language should continue being taught in pre-university computer science courses. i i i CONTENTS Page LIST OF TABLES vii LIST OF FIGURES ix ACKNOWLEDGEMENT x Chapter 1. INTRODUCTION 1 Purpose of the Study 3 Overview 4 Organization of the Following Chapters 6 2. REVIEW OF THE LITERATURE 7 Predicting Computer Science Achievement 7 Predicting Achievement in General 27 Introductory Computer Science Curriculum 28 Introductory Computer Programming Language 32 3. METHODOLOGY 40 Sample Selection 40 Course Descriptions 41 Questionnaire 43 Pilot Testing 46 Course Grades 47 Questions 47 BASIC Language Backgrounds 47 Backgrounds (non-computer language-related) 48 Different Language Backgrounds 48 iv Chapter Page BASIC and Pascal B ackgrounds 49 Achievement Predictors 49 Methods of Analysis 49 4. RESULTS 52 Descriptive Analyses 52 Achievement in Introductory Computer Science Courses 66 Inferential Analyses 67 Results on BASIC Language Backgrounds 68 BASIC versus No Language 68 BASIC versus Other Languages 69 BASIC Only versus BASIC and Other 70 BASIC First versus Not BASIC First 71 Number of Computer Languages (0-4) 72 Results on Demographic Backgrounds 73 Gender 73 Age 74 Faculty 75 Major 76 Results on Different Language Backgrounds 78 Different Backgrounds in First Year 78 Different Backgrounds in Second Year 82 Results of BASIC and Pascal Backgrounds 83 "How Well" Language was Learned 83 "When" Language was Learned 86 "Where" Language was Learned 88 v Chapter Page Results of Achievement Predictors 90 Achievement Predictors 90 5. DISCUSSION 92 Summary and Conclusions 92 BASIC Language Backgrounds 92 Demographic Backgrounds 95 Different Language Backgrounds 97 BASIC and Pascal Backgrounds 98 Achievement Predictors 100 Implications 100 Limitations 101 Suggestions for Further Research 102 BIBLIOGRAPHY 104 APPENDIX A. Letter to Instructors 117 B. Coding Scheme 119 C. Statistical Tests 123 D. Personal Communications 141 vi LIST OF TABLES Table Page 1 Number of Sections Surveyed 41 2 Number of Responses to the Questionnaire 41 3 Distribution of Students in Computer Courses 53 4 Distribution of Students by Year of Registration 54 5 Distribution of Students by Faculty 54 6 Order of Learning BASIC for Students with Prior Knowledge of BASIC Language 56 7 Order of Learning Pascal for Students with Prior Knowledge of Pascal Language 58 8 Distribution of Students by Number of Computer Science Courses Taken 62 9 Comparison of Marks in Computer Science Course with Averages in all Other Courses 67 10 Comparison of Results of Students Who Learned BASIC as the First Language and Students Who Knew No Languages 69 11 Comparison of Results of Students Who Had Learned BASIC and Students Who Had Learned Another Language 70 12 Comparison of Results of Students Who Had Learned Only BASIC and Students Who Had Learned BASIC and Another Language 71 13 Comparison of Results of Students Who Had Learned BASIC First and Students Who Had Learned BASIC Other than as a First Language 72 14 Comparison of Results of Students with Various Number of Computer Languages Learned 73 15 Comparison of Students Results by Gender 74 16 Comparison of Students Results by Age 75 17 Comparison of Students Results by Faculty 76 18 Comparison of Students Results for Selected Majors 77 19 Number of Non-mathematics Majors in the Faculty of Arts 78 vii Table Page 20 Comparison of Results Based on Computer Language Background in the FORTRAN (CPSC 101 and CPSC 151) Courses 80 21 Comparison of Results Based on Computer Language Background in the Pascal Language (CPSC 114 and CPSC 118) Courses 81 22 Comparison of Results Based on Computer Language Background in the Second Year (CPSC 210) Course 83 23 Comparison of Results Based on "How Well" BASIC Language had been Previously Learned 85 24 Comparison of Results Based on "How Well" Pascal Language had been Previously Learned 86 25 Comparison of Results Based on "When" BASIC Language had been Previously Learned 87 26 Comparison of Results Based on "When" Pascal Language had been Previously Learned 88 27 Comparison of Results Based on "Where" BASIC Language had been Previously Learned 89 28 Comparison of Results Based on "Where" Pascal Language had been Previously Learned 90 viii LIST OF FIGURES Figure Page 1 Courses Involved 42 2 The Questionnaire 44 3 Distribution of Students with Prior Knowledge of a Computer Language 55 4 Distribution of Students with Prior Knowledge of BASIC Computer Language 56 5 Distribution of Students with Prior Knowledge of Pascal Computer Language 57 6 First Learned Computer Language 59 7 BASIC, Pascal, LOGO Languages Learned 60 8 Languages First Learned by 835 Students with Prior Knowledge of a Computer Language 61 9 Order of Learning each Computer Language for Students with Prior Knowledge 62 10 Distribution of Students with Prior Knowledge of a Computer Language by Course 63 11 Distribution of Students with Prior Knowledge of BASIC Language by Course 64 12 Distribution of Students with Prior Knowledge of Pascal Language by Course 64 13 Distribution of Students with Prior Knowledge of LOGO Language by Course 65 14 Distribution of Students with Prior Knowledge of Other Languages by Course 65 15 Percentages by Gender with Prior Knowledge of Computer Languages 66 ix ACKNOWLEDGEMENT I would like to thank the members of my thesis committee, Dr. Marv Westrom, Dr. James Sherrill, and Dr. Hugh Dempster for their guidance. It has been a privilege for me to work with them. I am endebted to the efforts of the members of the Eric Hamber School Data Processing Service Club for doing much of the typing of this document. Thanks also, to their sponsor, Mr. Tim Ireland. I would like to express my gratitude to the U.B.C. computer science students who agreed to participate in the study and to their instructors for granting me permission to interrupt their classes. Finally, I would like to thank my family for their encouragement during the study, especially my wife Brenna for her patience and constant support. x Page-1 Chapter One Introduction This is an information age, with computer technology being responsible, in part, for recent increases in world productivity. Thus, as Andrew Molnar (1978) said "...computer literacy is a prerequisite to effective participation in an information society and as much a social obligation as reading literacy" (p.37). In order for students to participate effectively in the information age one must introduce computer science curricula into schools and post-sec-ondary institutions. The most common computing course, or the largest single topic within a computing course in the school system, at present, and over the past twenty years, is "computer programming". In the school system this "computer programming" course is taught using BASIC language in a majority of situations (Woodhouse, 1983). Charles Schulz (1984) claimed that one of the most important computer-related backgrounds desired for entry into college or university programs in the U.S.A. is a background in BASIC, Pascal or FORTRAN. Currently, there is considerable debate about the merits of teaching BASIC as the intro-ductory computer programming language. (Hereafter, the writer will use the term "language" to refer to "computer programming language".) It is acknowledged that BASIC is widely used and comparatively easy to learn. Malkit Sail (1986) wrote "BASIC is the only language that comes free of cost and is resident in almost all personal computer sys-tems" (p.49). It was developed in 1965 by John G. Kemeny and Thomas E. Kurtz of Dartmouth College primarily as a language for introductory computer science courses. However, as noted below, some computer educators claim that it is a mistake to teach BASIC as a first language. Alfred Bork (1982b) stated BASIC is the junk food of modern programming. ... It is becoming clear that students who learn BASIC as their first computer language will in almost all cases acquire a set of bad programming habits. These habits are very difficult to over-come, so BASIC programmers have difficulty writing readable and maintainable code (p. 12). Introduction Page 2 Another distinguished computer scientist, Edsger Dijkstra (1982) claimed that "It is practi-cally impossible to teach good programming to students that have had a prior exposure to BASIC: as potential programmers they are mentally mutilated beyond hope of regenera-tion" (p. 14). Roy Atherton (1982a) (author of a text on COMAL programming language (1982b)) wrote an article entitled "BASIC damages the Brain" in which he explained how a student of his was insistent on programming in BASIC and was reluctant to use more modem techniques. He added "Her brain was damaged and it proved difficult, though not impossible, to effect a cure" (p.16). Borge Christensen (an author of COMAL - a pro-gramming language which combines "...the simplicity of BASIC with the power of Pascal" (1982b, p.6)) stated that "BASIC was seriously and fundamentally defective" (1982a, p. 18). A common complaint of BASIC language is that it is not structured. Structured programs have fewer errors, are readable and easy to modify (Bork, 1981). Kernighan and Plauger (1974) claim that "...even the techniques of structured program-ming do not ensure that code will be good; 'structured' programs can be just as bad as their unstructured counterparts" (p.303). They indicate that in any programming language "...the presence of bad features is not an invitation to use them, nor is the absence of good features an excuse to avoid simulating them as cleanly as possible. Good programming languages are nice, but not vital" (p.319). Shapiro (1980) obtained results that suggest that too great an emphasis on structured programming leads to poor programming style and incorrect implementation of straightforward algorithms. Ralston (1971) states that "...whatever training the student in a first course receives in good programming technique should be almost independent of the language used for implementation.... choice of language is very much a second order effect compared to quality of instruction" (p.28). He adds that if a "perfect" language were used in a first year course, students might be misled into thinking that all is beauty and elegance in computer science, and not see its imperfec-tions. Ralston claims that a beginning computer student may be unable to grasp significant Introduction Page 3 concepts in depth. Thus, is too much emphasis being placed on the programming language employed and not enough on good teaching, or other aspects of the language? Whether BASIC should or should not be taught, it is slowly being replaced by other com-puter languages in many educational institutions. Agee (1985) was asked if "...we should downgrade the use of BASIC in favour of another language". He concluded that because BASIC is so widely used in the industrial/business community and is provided with nearly every available microcomputer that "...students need to learn BASIC for the real world of the 1980's and probably into the 21st century" (p. 13). He pointed out that it is important that students learn what they need and not what educators want them to learn. Tesler (1984) concurs that BASIC "...has fallen from favor somewhat in the academic world" (p.72). But he finds it impossible to select a best programming language because of the great diversity of languages. He concludes that a language must be chosen for the purpose intended. P u r p o s e o f t h e S t u d y The purpose of this study is to determine how the prior knowledge of BASIC language affects the achievement in introductory computer science courses at university. To this end the study compares the achievement in introductory computer science courses of a group with a prior knowledge of BASIC language with that of other groups who have learned other languages, or have no prior knowledge of any language. It also compares achieve-ment among groups with different backgrounds, such as gender, age, Faculty and major. The study assesses the differences in achievement in first year FORTRAN and Pascal courses and in a second year computer science course among groups with different com-puter language backgrounds. It ascertains the impact on achievement of "how well", "when" and "where" a previous language was learned. Finally, the study identifies the factors that are the best predictors of success in introductory courses of computer science. Introduction Page 4 The implications this study holds for the computer curriculum in secondary schools is its raison d'etre. If those students who first learned BASIC have lower achievement than the group who did not learn BASIC first, then the school system using B A S IC as the first language may need to reconsider its use. At the same time, the study's results may be of interest to computer software developers for future undertakings. This study may provide valuable information to universities who are presently admitting into computer science programs a large number of students, with widely diversified backgrounds. The universi-ties might use the information for counselling purposes, for course designing, or for limit-ing enrolments. If those students who first learned BASIC do not have lower achievement than other groups, the school system does not need to change from instruction in BASIC language to one that is structured, or free from the "bad" features of BASIC, without more extensive studies. This could potentially save educational systems and school districts a great deal of money in software, hardware and in the costs of retraining teachers. O v e r v i e w Questions this study will attempt to answer include: 1 How does introductory computer science achievement compare between the group of students whose first computer language was BASIC and the group who knew no languages? 2 Does achievement differ between the group with prior knowledge of BASIC and the group who had prior knowledge of computer languages other than BASIC? 3 Does achievement differ between the group with prior knowledge of BASIC only and the group who had prior knowledge of BASIC and another language? 4 Of all students with prior knowledge of BASIC does achievement differ between the group who had learned B A S IC first and the group who learned another language before learning BASIC? Introduction Page 5 5 Are there differences in achievement among the groups of students who have already learned 0, 1,2, 3, or 4 computer languages? 6 Is there a difference in achievement by gender? 7 Are there differences in achievement among various age groups? 8 Are there differences in achievement among the Faculties represented? 9 Is there a difference in achievement between students who are majoring in mathemat-ics and students who are majoring in various fields within the Faculty of Arts (i.e. non-mathematics majors)? 10 Is there a difference in achievement among students with different computer language backgrounds enrolled in the first year FORTRAN (CPSC 101 or CPSC 151) or Pascal (CPSC 114 or CPSC 118) courses? 11 Is there a difference in achievement among students with different computer language backgrounds enrolled in the second year course (CPSC 210)? 12 Are there differences in achievement among the groups that respond in five different ways regarding "how well" a previous language (either BASIC or Pascal) was learned? 13 Are there differences in achievement among the groups that respond in three different ways regarding "when" either BASIC or Pascal was learned? 14 Are there differences in achievement among the groups that respond in four different ways regarding "where" either BASIC or Pascal was learned? 15 Which factor(s) under investigation is/are the best predictor(s) of success in univer-sity introductory courses in computer science? Organization of the Following Chapters A review of the literature pertaining to the experimental questions, a description of the methodology, the results of statistical analysis and a discussion of the findings of the study are found in the following chapters. The review of the literature is presented in Chapter Two. Chapter Three contains the details of the sample selection, the courses involved, the development and administration of the questionnaire, the pilot testing, other data used, the statement of the questions, and the methods of data analysis. The research hypotheses and the results of the descriptive and inferential analyses are discussed in Chapter Four. The conclusions are formulated in Chapter Five together with a discussion of the findings and their implications. This final chapter also includes limitations of the study and areas for further research. Page~7 Chapter Two Review of the Literature Few studies have looked at the effects of secondary school computer science programs on the ensuing achievement in introductory computer science courses at university. Do stu-dents with secondary school experience in computer science perform at different levels, i.e. either higher or lower levels of achievement in university computing courses than students that have taken no formal secondary school computing courses? Predicting Computer Science Achievement Over the last twenty years, many studies have been conducted to predict academic achieve-ment in an introductory computer science course at the secondary school, college or university level (Alspaugh, 1970 and 1972; Barker and Unger, 1983; Bateman, 1973; Bauer, Mehrens and Vinsonhaler, 1968; Buff, 1972; Butcher and Muth, 1985; Cafolla, 1987; Campbell and McCabe, 1984; Capstick, Gordon and Salvadori, 1975; Cheney, 1980; Correnti, 1969; Dey and Mand, 1986; Dixon, 1987; Fowler and Glorfeld, 1981; Gathers, 1986; Glorfeld and Fowler, 1982; Gray, 1974; Greer, 1986; Guinan and Stephens, 1988; Hostetler, 1983; Howerton, 1988; Hunt and Randhawa, 1973; Konvalina, Wileman and Stephens, 1983; Kurtz, 1980; Leeper and Silver, 1982; Lemos, 1981; McGee, Polychronopoulos and Wilson, 1987; Mazlack, 1976; Mazlack, 1980; Mussio and Wahlstrom, 1971; Newsted, 1975; Nowaczyk, 1983; Oman, 1986; Petersen, 1976; Petersen and Howe, 1979; Plog,1980; Ramberg and Van Caster, 1986; Sauter, 1986; Schroeder, 1978; Sharma, 1987; Sorge and Wark, 1984; Stephens, Wileman and Konvalina, 1981; Stephens, Wileman, Konvalina andTeodoro, 1985; Stevens, 1983; Szymczuk and Frerichs, 1985; Tillman, 1974; Werth, 1986; Whipkey and Stephens, 1984; Wileman, Konvalina and Stephens, 1981; Wileman, Stephens and Konvalina, 1982; Yaney, 1970). One purpose of this chapter is to consider the findings of these studies. Review of the Literature Page 8 Many different factors have been used as predictor variables of the success in introductory post-secondary computer courses. There is, however, an apparent absence of articles in which university computer science course achievement has been predicted using secondary school computer science performance as one of the predictors. Bauer, Mehrens and Vinsonhaler (1968) performed some preliminary research for the development of a test battery for the selection of computer programmers. They examined the validity of the IBM Aptitude Test for Programmer Personnel (ATPP) and the Strong Vocational Interest Blank (SVIB). Two other predictor variables used were the College Qualification Test (CQT) and grade point averages (GPA's). The best single predictor was GPA (r=0.68). They concluded that numerical reasoning (CQT Numerical) and spatial reasoning (ATPP Part U-figure series) appear to be two of the most important cognitive abilities needed for success as a computer programmer. Correnti (1969) investigated characteristics of students enrolled in computer programming courses to determine whether any of them could be useful for predicting success in pro-gramming courses. His study used as independent variables programming aptitude, study habits and attitudes, scholastic aptitude, occupational level of the parents, and academic achievement as measured by students' previous GPA and their college entrance examina-tion results. He concluded that these variables are significant indicators of programming success. The "above-average" achievers were significantly different from the "below-average" ones on the variables studied. Yaney (1970) also did a study to predict programming performance. He developed a paper and pencil test that was administered both before and after a training program. He found that programming performance could be predicted accurately at both times, with the post-test yielding a high(er) degree of confidence. Review of the Literature Alspaugh (1970) investigated the relationships between selected student characteristics and proficiency in two different programming languages (Basic Assembly Language and FORTRAN). She found that the mathematics background of computer programming stu-dents appeared to be the major influencing component for achievement in programming. The more successful programming student possesses a low level of "impulsiveness" and "sociability" and a high level of "reflectiveness". Since Alspaugh had similar findings in the two computer languages, she concluded that computer programming aptitude probably doesn't vary from computer language to computer language. Since most of the variance in programming proficiency was not explained she stated that there are obviously other pre-dictors yet to be identified. Alspaugh (1972) concluded that the placement of computer programming classes within the mathematics curriculum is justified. However, "it would not be realistic to assume that all students who are talented in mathematics will also be talented in computer programming" (p.98). Mussio and Wahlstrom (1971) analyzed the predictive validity of a number of tests in order to determine which one(s) predict(s) most accurately success in a course for computer pro-grammers. From the Computer Programmer Aptitude Battery (CPAB), the Thurstone Test of Primary Mental Abilities (PMA), the SVIB, and a revised form of Russell's Scale of Motivation for School Achievement, the best single predictor of academic performance was the diagramming test of the CPAB, correlating 0.49 with the course grade. The best com-bination of predictors, using stepwise regression procedures, was the chagramming test, the programmer key and age-related interest scale of the SVIB, and Russell's test of moti-vation, correlating 0.67 with the criterion measure, the authors conclude that "reasoning ability is the single most important qualification for programmers" (p. 34) in addition to certain personality measures. Buff (1972) used ten independent variables to provide measurements of aptitude, achieve-ment, and socioeconomic status in order to develop a method for prediction of academic Review of the Literature Page 10 achievement in FORTRAN programming courses. He found that achievement scores cor-related more highly with the dependent variable than did the aptitude scores, almost without exception. The GPA proved to be the best single predictor. Of the socioeconomic vari-ables, Buff found that the father's occupation and the father's education correlated nega-tively with the dependent variable. The hypotheses that student achievement can be pre-dicted from aptitudes, abilities, prior academic performance, and other socioeconomic characteristics, were upheld. Hunt and Randhawa (1973) attempted to ascertain the relationship between 12 cognitive variables, and the final mark in a university introductory computer science course. They used raw scores from seven tests taken from the Kit of Reference Tests for Cognitive Factors and five subtests of the Watson-Glaser Test of Critical Thinking (CT) as the pre-dictor variables. They found using a stepwise regression analysis that the Locations, Hidden Patterns, and Seeing Problems from the reference tests, and Deduction from the CT test accounted for only 23.1% of the variance of the criterion variable. Bateman (1973) used 16 independent variables in order to predict a course grade in an introductory computer course, for the purpose of counselling and/or screening students. Among the independent variables that he used were prior programming experience, an I.B.M. aptitude test, Scholastic Aptitude Test (SAT) verbal and mathematics scores, total credit hours accumulated, GPA, high school rank, and program major field. He found that prior programming experience was moderately correlated with the final computer course grade (r=0.7). His predictor variables accounted for 60% of the variance of the course grade. Gray (1974) too, was interested in the guidance and counselling of post-secondary stu-dents. He identified and analyzed selected variables which included: age, sex, Revised Programmers Aptitude Test (RPAT) scores, ATPP scores, and scores on seven areas of the Review of the Literature Page 11 General Aptitude Test Battery (GATB). He found that Intelligence, Verbal, and Spatial scores on the GATB and ATPP scores were valid predictors of success in the data processing technology program, but none were found to be reliable enough to determine acceptance or rejection of prospective students. The ATPP score was determined to be the best single indicator of success, and was useful as a counselling guideline. Gray also con-cluded that regression equations can be used effectively to predict achievement level for placement purposes. Tillman (1974) conducted a study to predict successful computer programmers using the Biographical Information Blank (BIB), the School and College Abilities Test (SCAT) and the SVIB as predictor variables. The SCAT Math section demonstrated validity as a pre-dictor of success; while the BIB and SVIB variables were not significantly correlated to success. Newsted (1975) used 14 independent variables in attempting to predict both grades in an introductory university programming course and students' self-perceived ability in this course. He found that college GPA was the only independent variable that correlated pos-itively with the programming course grade. He found that programming experience and career orientation predicted self-perceived programming ability. None of the personality variables correlated significantly with either criterion variable. None of the personality variables entered into either of the regression equations with the two dependent variables. Of the behavioural variables the "time spent on the course and working with other students correlated negatively with the dependent variables" (p.89). That is, although poorer stu-dents may spend much time and ask many questions of their instructors and fellow stu-dents, it does not improve their grade or their ability perception. Furthermore the 14 pre-dictor variables used in this study fail to account for the majority of the variance in either the course grade or the self-perceived ability variables (i.e. only for 41% and 49%, respectively). Review of the Literature Page 12 Capstick, Gordon and Salvadori (1975) investigated the use of ATPP scores as predictors of course grades in two different computing courses. The ATPP scores showed no significant relation to course grades in the COBOL course, but were significantly related to course grades in the FORTRAN course. The arithmetical reasoning component of the ATPP was the only significant component that could be used to predict grades in the FORTRAN course. Mazlack (1976) investigated the correlations between a student's gender, his academic pro-gram, his semester in school and his grade in the computing course. He found negligible correlations between each of these three independent variables and computer science achievement. He concluded that there is no reason to segregate students from differing semesters or from various academic programs because of concerns about differences in learning ability when these concerns are based on academic discipline bias. Over several years Mazlack (1980) found that academic discipline, sex and academic expe-rience have low correlations with success in an introductory programming course. He also found that Programmer Aptitude Tests (PAT) had low predictive value when administered to people at college level. Mazlack concluded that the longer a student is in school the less uniform is his application to the study of the introductory computer science course. As in his later study, Mazlack (1976) concluded that those beginning their university experience and those well along in their academic careers seem to succeed or fail in much the same manner. Petersen (1976) gathered biographical information, temperament information from the Thurston Temperament Schedule, and general aptitude information from the GATB as independent variables in predicting achievement in the course "Introduction to Computers". His criterion variables were the midterm examination mark, the programming grade, the final examination score and the final grade. He concluded that the criterion variables were Review of the Literature Page 13 correlated with the independent biographical variables high school rank, the number of mathematics and science courses taken in high school, the number of semesters of mathe-matics and science courses taken in high school, high school GPA, college GPA, high school mathematics and science GPA, and a self-predicted grade in the introductory com-puter course. The criterion variables were found to be correlated with the temperament variables "impulsive" and "sociable". The dependent variables were correlated with the general aptitude variables "verbal", "numerical", "clerical", and "general intelligence". The college GPA although a "late" predictor variable was the best predictor for the program-ming grade, the final examination score and the course grade. That is, the students with the highest GPA also had the highest programming grades, the highest final exam scores and the best course grades. Later, Petersen and Howe (1979) extended this research. They were among the first to undertake to predict success in a "general education type of introductory computer course" (p. 183). Their data included fifteen items from an autobiographical survey, the Thurstone Temperament Schedule results, and the GATB scores, which included a general intelli-gence score. Of these predictors only college GPA and general intelligence were found to contribute significantly to predicting an introductory computer course grade. The authors state that in advising pre-college students on the "Introduction to Computers" course, "an able high school student successful in mathematics and science will probably be a success-ful student in computer science" (p. 190). Schroeder (1978) conducted a study to determine if spatial reasoning ability, mathematical reasoning ability and Piagetian formal thought ability were significantly correlated with achievement in computer programming classes. His results revealed that mathematical reasoning ability had the greatest correlation with computer programming course achieve-ment. Further analysis indicated that of the three independent variables both mathematical reasoning and Piagetian formal reasoning scores made significant contributions to explain-Review of the Literature Page 14 ing the variance of achievement scores. Thus, he concluded that these two factors in com-bination are significant in predicting success in computer programming, but spatial reason-ing ability is not important. Cheney (1980) studied the relationship between cognitive style and a student's ability to program. He found that analytic decision makers tend to perform better than heuristic decision makers on programming examinations. He concluded that cognitive style can be used as a predictor of student success in learning a computer programming language. Kurtz (1980) conducted a study to investigate the relationship between intellectual devel-opment and performance in an introductory programming class. He found that intellectual development did not vary with sex, class level, and previous coursework. He found that the "late concrete" level and the "late formal" level of intellectual development were strong predictors of "poor" and "outstanding" performance, respectively, in an introductory pro-gramming course. It was also found that the intellectual development level predicts performance on tests better than it predicts performance on student-written programs. Plog (1980) conducted a study to investigate the relationship of selected variables in pre-dicting academic success in computer programming using community college introductory computer programming students. She found no significant relationship between total apti-tude test scores and the degree of academic success in an introductory computer program-ming course. She found that both the verbal ability scores and the quantitative ability scores were higher for those students successfully completing the computer programming course than for those who were unsuccessful. She concluded that the SCAT total aptitude scores do equally well for both sexes in predicting final course grades in the introductory course. Older students have higher final course grades and higher total aptitude scores than younger students. Caucasians' scores were significantly higher than Negroes' scores on final course grades and total aptitude scores. She found that the amount of previous train-Review of the Literature Page 15 ing and experience in data processing made no significant difference in the total aptitude test scores and in the degree of academic success in an introductory computer programming course. Stephens, Wileman and Konvalina (1981) did a study to identify group differences in computer science aptitude of beginning computer science students at university. They used various student background factors which included age, sex, college year, hours worked, previous computer experience, estimated high school performance, and estimated current college performance, as well as aptitude variables which included reading compre-hension, sequence completion, logical reasoning, algorithmic execution and alphanumeric translation, in order to investigate any group differences. However, of these factors only two of them had a significant impact on performance - the estimated current college perfor-mance and the estimated high school performance. "Students who performed well in high school and in college generally performed well on the aptitude pretest" (p.94). The com-ponents of the pretest that were most significant in differentiating students (i.e. sequences, logical reasoning and algorithmic execution) relate to the mathematical reasoning ability of the students. Thus, these researchers concur with the findings of Petersen and Howe (1979) that students who do well in high school mathematics and science in particular, do well in an introductory computer science course. A significant group difference was found on the algorithmic execution component of the pretest between age groups and between groups differing in previous computer experience. Older students, i.e. 35 or older, and those students with no prior computer experience performed at a lower level on the execu-tion of algorithms. The reading comprehension component did not produce significant differences in scores for any of the factors. Thus, the authors claim that their research indicates that reading comprehension does not seem to be useful as a predictor of computer science aptitude. Review of the Literature Page 16 Wileman, Konvalina and Stephens (1981) used reading comprehension, alphabetic and numeric sequences, logical reasoning, algorithmic execution, alphanumeric translation, age, high school performance and the number of hours of work per week as factors in pre-dicting success in a beginning university computer science course. When correlated with final exam scores for the computer science course, only reading comprehension, alphabetic and numeric sequence completion, logical reasoning, and algorithmic execution were found to be significant. This is a much stronger result than that obtained by Mazlack (1980) reported earlier in this chapter. Although these variables reinforce one another, they do not provide independent sources for prediction. In addition, a stepwise multiple regression was completed to determine the most important factors in predicting success. When all eight factors were included in the prediction equation only 25% of the variance in final exam scores was explained. Stephens, Wileman, Konvalina, and Teodoro (1985) found that there was a direct relationship between performance on this placement test and on computer course performance at a foreign university. Wileman, Stephens and Konvalina (1982) examined the relationships between mathematics competency and computer science achievement in order to advise and to place prospective computer science students. They concluded that there is a strong relationship between mathematical competencies and probable success in beginning computer science courses. Konvalina, Wileman, and Stephens (1983) compared students who withdraw from an introductory university computer science course with those who do not withdraw from these courses with respect to such demographic factors as age, estimated high school performance (Students chose among the categories "excelled", "above average", and "average or below".), hours worked per week, prior computer education, prior non-programming computer work experience, prior programming work experience, years of high school mathematics, and number of college mathematics courses, and with respect to such predictor factors as number and letter sequences, logic-type questions, calculator Review of the Literature Page 17 simulator, algorithmic questions, and high school algebra word problems. A consistent result that was found is the "important role of mathematical reasoning ability and mathe-matical background to potential success in computer science" (p.377). The nonwithdraw-ers were significantly older; their high school performance was significantly better, they had significantly more previous educational experience involving computer science; they had more mathematics background (high school and university combined). Between the two groups "there were no statistically significant differences in the number of hours worked per week, previous work experience involving computers, and the number of years of high school mathematics" (p.378). On the five predictor components the scores for the nonwithdrawers were significantly higher for the sequences, the logic-type questions, the calculator and word problems components, in addition to the predictor total. The authors are using the results of this research in advising and placing students. Fowler and Glorfeld (1981) have developed a model expected to yield a 75% correct classification rate of either high or low aptitude in an introductory university computer course (that is evenly divided between programming and other computer topics). They used personal, academic and aptitude variables as predictors, including sex, age, academic major, university year, number of mathematics courses, GPA, final grade in an introduc-tory computer science course, the SAT mathematics and verbal scores, and the Wolfe Programming Aptitude Test (WPAT) percentile scores. Of these factors the authors found that "Grade Point Average (GPA) is the most important single variable in the model, indi-cating that a student's past academic performance is the best indicator of future academic performance" (p. 108). Also, they found that the development of problem-solving skills is important to success in computing as indicated by the inclusion of the SAT mathematics score and the number of mathematics courses variables. They found that age is of marginal importance but younger students can be predicted to have a greater likelihood of success in introductory computing. One reason the authors suggest for this is "it is also possible that Review of the Literature Page 18 the younger students have had more exposure to computing in their secondary education" (p. 108). Glorfeld and Fowler (1982) followed up their previous study (Fowler and Glorfeld, 1981) with a validation study of the classification model developed earlier. The results of this second study reveal that an approximate 75% correct classification rate would be expected in actual application of the new model developed from the combined sample data set. The authors state that the age variable does not appear to be making a real contribution to the model. However, GPA is seen as the most important component of the model. The inclusion of the number of mathematics courses and the SAT mathematics score as pre-dictors agrees with the findings of Alspaugh (1972), Petersen and Howe (1979) and Stephens, Wileman and Konvalina (1981), that mathematics background is related to the likelihood of success in introductory computing. Lemos (1981) performed an experiment using a pretest score and age as covariates which compared scores on an examination that tested proficiency with BASIC language between university students who were business majors and others who were not. All of these stu-dents were enrolled in an introductory computer course. The results showed no significant difference between the scores of the two groups. The ability to understand programming languages is independent of the student's academic direction. "Thus, the assertion that the general university student will not do well if programming is stressed does not appear to have an empirical basis" (p. 157). Leeper and Silver (1982) developed a predictor of success in an introductory programming course for use as a screening instrument to restrict enrollment. They used the number of units of each of high school English, mathematics, science, and foreign language completed, the student's high school rank and the SAT verbal and mathematics scores and predictor variables, the SAT verbal and mathematics scores correlated the strongest with Review of the Literature Page 19 the course grade. All of these independent variables explained at most 26% of the variation in the grades. Approximately 64% of those students for whom failure was predicted did fail. Barker and Unger (1983) shortened the Kurtz (1980) test of abstract reasoning to use it to predict success in an introductory computer science class and to segregate students into accelerated and non-accelerated sections of an introductory computing course. The results failed to yield the significant correlation between intellectual development predictor and final course grade as Kurtz (1980) found (0.12 versus 0.80 for Kurtz). However, this tool successfully segregated the advanced students from average and below average students. Nowaczyk (1983) used factors which included past academic performance, personality factors and performance in problem solving to predict success of college students in intro-ductory computer prograiriming courses. He found that past academic performance in mathematics and English courses, the amount of previous computer experience, the expected grade in the course, and performance on selected logic and algebraic word problems were significantly related to course performance. Stevens (1983) found a significant difference between cognitive styles and academic success of education students enrolled in instructional computer courses. Field indepen-dent students (those whose "perceptions are analytical") had significantly higher achieve-ment scores than field dependent students (those whose "activities and perceptions are global and tend to focus on the total environment") in the computer courses (p. 229). She concludes that educators must acquire new and different skills to meet the needs of students. Hostetler (1983) examined to what extent certain cognitive skills, personality variables, and past academic achievement can be used to predict success in an introductory computer pro-gramming course. Of the 20 independent variables used only the diagramming and Review of the Literature Page 20 reasoning tests of the CPAB and a student's GPA correlate significantly with a student's final course score. No personality trait variables and mathematics background correlated significantly with the dependent variable. A multiple regression equation of the five highest correlating predictor variables explained approximately 43% of the variance in the final course score. Sorge and Wark (1984) conducted a study to determine which factors can be used to pre-dict success or lack of success as a computer science major. As predictor variables they considered such precollege data as SAT verbal and mathematics scores, number of semesters and average high school grades in mathematics, science, English, and high school rank. The authors found that "to have a reasonable chance of making satisfactory progress" (p.44) in the computer science majors program students should have significant mathematical and verbal skills (SAT mathematics score of at least 560 and SAT verbal score of at least 500). They observed a high attrition rate even by better students. These dropouts said that time, attention to detail, systematic thinking and the impersonal inter-action with machines were undesirable activities of the program. The authors conclude that there are factors other than academic ability involved in being successful as a computer science major. Campbell and McCabe (1984) were concerned with identifying factors that influence success in the first year of a computer science major. They used the SAT mathematics and SAT verbal scores, the rank in high school graduating class, the size of high school gradu-ating class, the number of semesters of each of high school mathematics, science and English courses, the average grades in each of these three high school subject areas, and sex as the predictor variables. They found that the SAT mathematics score, the average grade in high school mathematics and sex were the best predictors of success. Review of the Literature Page 21 Whipkey and Stephens (1984) used SAT verbal and mathematics scores, a Quality Point Average (QPA), the Myers-Briggs Type Indicator (MBTI) scores of Extroversion-Introversion, Sensing-Intuition, Thinking-Feeling, and Judgement-Perception to predict beginning programming success. They found that the QPA was correlated the highest (r=0.73) with the target variable, a programming ability score (PAS). The best statistically significant predictor model uses QPA and SAT mathematics scores (r=0.75) to predict PAS. The personality traits measured by MBTI failed to add significantly as predictors. Butcher and Muth (1985) performed a study to predict college performance in introductory computer science courses using only high school transcript data and American College Testing Program (ACT) test scores. They found that performance can be predicted based on this information. However, only 37% of the variation in final course grade was accounted for by this relationship. Szymczuk and Frerichs (1985) investigated the predictive value of several standardized aptitude instruments for use at the high school level. They administrated the Iowa Test of Educational Development (ITED), Science Research Associates' CPAB and the WPAT. They concluded that mathematical ability is an important factor influencing success in a secondary school computer science course. Werth (1986) studied the relationship between the student's grade in introductory computer science and his/her sex, age, high school and college academic performance, number of mathematics courses, work experience, cognitive development, cognitive style, and personality factors. She used the measure developed by Kurtz (1980) to measure intellec-tual development, the Group Embedded Figures Test (GEFT) as a measure of cognitive style, and the MBTI as a measure of personality type. Significant relationships were found between letter grade and college grades, the number of hours worked, the number of high school mathematics courses, the GEFT scores, and the Piagetian intellectual development Review of the Literature Page 22 measure scores. Although no relationship was found between grade and personality type, computer science students "were found to be far more introvertive, intuitive and thinking than the population as a whole" (p. 141). Dey and Mand (1986) investigated the relationship between the level of the student's math-ematics background and his performance in introductory BASIC, Pascal, and COBOL courses, and the effect of the student's performance in one computer language course on the learning of another. Their results indicated significant correlations between a respon-dent's reported average grade in high school and college mathematics courses and his expected grade in a current introductory computer science course. They found that students with some computer language background are more likely to succeed in a subsequent com-puter science course than students who have had no previous computer language exposure. For students exposed to more than one computer language, they found that there was a stronger complementary relationship between COBOL and Pascal than between either BASIC and COBOL or BASIC and Pascal. That is, students with prior exposure to COBOL or Pascal had higher grade expectations in a subsequent Pascal or COBOL course than students with no prior computer language experience. Whereas students with prior exposure to BASIC seemed to have no higher grade expectation in a subsequent Pascal or COBOL course than students with no prior computer language experiences. Greer (1986) examined the relationship between high school computer science course experience and subsequent introductory university computer science course achievement. The students were grouped according to both the amount of high school computer science instruction and the degree of emphasis on structured programming in high school courses. No significant differences in overall achievement in the university course were found among the groups. Students with more high school experience were found to be less likely to withdraw from the course. Greer concluded that higher grades in university computer science do not result from high school computer science experience. Review of the Literature Page 23 Oman (1986) investigated to what extent previous computer experience (the number of time-sharing systems previously used, the number of microcomputers previously used, the number of programming languages previously used), the number of years since high school graduation, and SAT mathematics and verbal proficiency scores have in predicting how successful a person will be in an introductory computer science course. Mathematics proficiency was the most highly correlated to the dependent variable, final course grade. The SAT mathematics score alone accounted for 65% of the grade variation. All six inde-pendent variables accounted for 82% of grade variation. The author concludes that this model can be used as an advising tool for predicting success in introductory computer science courses. Gathers (1986) was concerned with the identification and use of factors which could effec-tively predict success in a first computer science course in university. His study included as factors: high school GPA, ACT scores, standardized Nelson Denny Reading scores, and the University of Tennessee at Martin (UTM) Math Placement Test score. Of these placement factors the ACT English score and the UTM Math Placement Test score were found to be the best predictors of success in the first course for computer science majors. This research indicated that verbal skills are just as important as mathematical skills in predicting success. Ramberg and Van Caster (1986) examined the differences in scores on a placement test [the same test used as in Konvalina, Wileman, and Stephens, (1983)] between a group of stu-dents who eventually withdrew from an entry-level programming course and a group who finished the course. They found that those students with prior programming experience scored higher on the placement examination than both those with no prior programming work and those who did not finish the course. The amount of mathematics background was directly related to the placement examination score of the finishers. The authors inter-viewed many "successful" students who attributed almost unanimously their good grade to having enrolled in a beginning computer programming class in high school. Sauter (1986) assessed the relative roles of mathematics aptitude and language aptitude in the development and enhancement of programming skill. In an introductory COBOL course she found that programming ability is related to both mathematical aptitude and language aptitude. However, only 20% and 17% of the variation in students' perfor-mances on a Syntax Test and a Logic Test, respectively, could be explained. Cafolla (1987) conducted a study to deterrnine if success in writing computer programs is related to the level of cognitive development, verbal reasoning, mathematical reasoning, and grade point average. He found, using multiple regression analysis, that the level of cognitive development and verbal reasoning were significant predictors of success in computer programming. The best single predictor was verbal reasoning (r=0.62). Dixon (1987) investigated factors which may contribute to success or failure in university computer science courses. She found no significant differences in achievement for those who started computing studies prior to high school, in high school, and at college. A sig-nificant positive correlation was detected between SAT mathematics scores and computer science course achievement. A positive correlation existed between achievement and the length of programs written. Dixon discovered no significant differences in achievement between students with different initial programming language experience. She found no significant differences in learning style preference and achievement. However, Dixon did detect significant differences in learning style preference for students selecting computer science as a major compared to those students not majoring in computer science. McGee, Polychronopoulos and Wilson (1987) investigated the influence of BASIC lan-guage knowledge on performance in two introductory Pascal language courses. In one course no differences in final course grades were found based on prior programming lan-Review of the Literature Page 25 guage background. In the other course students who had done prior programming in BASIC had significantly higher final grades that those who had no programming background. Sharma (1987) analyzed many research studies concerning the identification of components of computer science aptitude responsible for success in computer science courses. He was interested in whether cognitive styles, psychological types, and background variables are factors which in addition to mathematical ability may have a bearing on programming abil-ity. The background variables include academic program, semester in school, gender, college GPA, and general intelligence tests. Sharma agrees with Stevens' (1983) finding that the learners' cognitive styles and psychological characteristics should be considered in the planning of programs and strategies to maximize their academic success in computer science courses. Further research is required to investigate the relationship between both cognitive styles and psychological types and performance in computer science courses. Howerton (1988) administered a post-treatment knowledge retention test to students who had completed an introductory college computer programming course. He was attempting to determine whether any of three levels of pre-college exposure to computers and com-puting would have a significant effect on student performance in this course. He found that for all three levels of pre-college computing exposure the mean scores of those who had been exposed were significantly higher than the mean scores of those who had not been exposed. He concluded that students with pre-college exposure to computers and computing at any level can be expected to achieve higher scores in introductory computer programming courses than students who have not had that exposure. In general these studies indicate: 1) that students with a strong background in mathematics and science are more likely to perform well in introductory computer science courses (Alspaugh, 1970 and 1972; Review of the Literature Page 26 Petersen, 1976; Schroeder, 1978; Petersen and Howe, 1979; Stephens, Wileman and Konvalina, 1981; Konvalina, Wileman and Stephens, 1983; Fowler and Glorfeld, 1981; Glorfeld and Fowler, 1982; Nowaczyk, 1983; Campbell and McCabe, 1984; Szymczuk and Frerichs, 1985; Dey and Mand, 1986; Gathers, 1986; Ramberg and Van Caster, 1986; Werth, 1986; Dixon, 1987; Sharma, 1987), 2) that students' GPA is the best single predictor of an introductory computer science course grade, of the many variables studied (Bauer, Mehrens and Vinsonhaler, 1968; Correnti, 1969; Buff, 1972; Newsted, 1975; Petersen, 1976; Petersen and Howe, 1979; Fowler and Glorfeld, 1981; Hostetler, 1983), and 3) that programming aptitude scores can be used in predicting programming success (Correnti, 1969; Mussio and Wahlstrom, 1971; Bateman, 1973; Hunt and Randhawa, 1973; Gray, 1974; Tillman, 1974; Capstick, Gordon and Salvadori, 1975; Plog, 1980; Leeper and Silver, 1982; Hostetler, 1983; Sorge and Wark, 1984; Campbell and McCabe, 1984; Whipkey and Stephens, 1984; Szymczuk and Frerichs, 1985; Oman, 1986; Sauter, 1986). Other studies have included many demographic variables that proved to be somewhat sig-nificant in predicting computer course success. Correnti (1969) included study habits and attitudes and the occupational level of the parents as independent variables. Bateman (1973), Nowaczyk (1983), Dey and Mand (1986), Oman(1986), Ramberg and Van Caster (1986), McGee, Polychronopoulos and Wilson (1987) and Howerton (1988) successfully used previous computer experience. In his investigation of cognitive style Cheney (1980) found that analytic decision makers perform better than heuristic decision makers on pro-gramming exams. Kurtz (1980) found that intellectual development predicted performance. Stevens (1983) concluded that field independent students had significandy higher scores in computer courses than field dependent students. Campbell and McCabe (1984) concluded that SAT verbal scores and gender were predictors of performance in computer science. Werth (1986) found that the amount of work experience had a significant correlation with computer science course grade. Cafolla (1987) used cognitive development and verbal reasoning as predictor variables. However, demographic variables have not proved to be Review of the Literature Page 27 as good in predicting computer science success as have the academic or aptitude variables. Tillman (1974) learned that biographical and vocational interest variables were not corre-lated to successful programming. Newsted (1975), Hostetler (1983), Whipkey and Stephens (1984) and Werth (1986) found that none of the personality variables used corre-lated with programming grades. Plog (1980) found previous data processing training and experience were not significant in predicting success in a computer programming course. Mazlack (1976 and 1980) concluded that a student's sex, and his academic program, his semester in school have negligible correlations with computer science course achievement. Lemos (1981) ascertained that student's academic discipline was independent of his ability to understand programming languages. Barker and Unger (1983) learned that intellectual development did not have a significant correlation with the final course grade in an intro-ductory computer science class. Greer (1986) concluded that the amount of high school computer science course experience has no effect on computer science grades at university. Sharma (1987) suggested that further study is required to determine what effect cognitive styles and psychological characteristics have on performance in computer science courses. Pred ic t ing A c h i e v e m e n t in G e n e r a l Many studies have been conducted to predict academic performance in general from various factors for student counselling purposes (e.g.: Cronbach, 1949; Horst, 1957; Lavin, 1965; Fudge, 1970; Hunt, 1977). College level ability tests correlate about 0.50 to 0.55 with GPA (Cronbach, 1949). However numerous personality and social factors are also involved and should not be overlooked in academic performance prediction. The content of predictor batteries varies depending on the particular course for which grades are being predicted (Horst, 1957). However, Cronbach (1949) contradicts this Review of the Literature Page 28 stating that multifactor tests of abilities add litde to the prediction of performance in particular courses beyond what the general intelligence factor will predict. Lavin (1965) states that the "single best predictor of performance on the college level is the high school academic record" (p.57). Biographical information greatly aids in accurate prediction of academic performance in college (Fudge, 1970). These data, either taken alone or in combination with aptitude test scores, provide better estimates of academic performance than do aptitude test data, either taken alone or in combination with high school rank. The prediction of academic performance is a complicated exercise involving many factors, whether the task be to predict success in computer science or in some other subject area. If one is to develop in students an ability to adjust to the world of work one should be con-cerned with other outcomes besides academic achievement. Lavin (1965) says that "we need to develop additional dimensions of student performance and to find out how they relate to various aspects of life after the completion of school" (p. 168). These dimensions will have implications for what one should include in a computer science curriculum. Introductory Computer Science Curriculum What should be taught in an introductory computer science course? Up to the present, the "typical" introductory course teaches "the students to become proficient in a single pro-gramming language" (van Dam et al, 1974, p. 175). Furugori and Jalics (1977) presented the results of a questionnaire returned from 44 universities. They found similar results to those found by van Dam's survey. "The first course in Computer Science is definitely a programming course for almost all the schools" (Furugori and Jalics, 1977, p.l 19). Brookshear (1985) stated that this traditional approach of learning to program in the first year and later learning other subjects, produces students with a narrow, vocationally-Review of the Literature Page 29 oriented view of computer science. The first course in computer science should not give "the unfortunate impression that Computer Science = Programming" (Ralston, 1984, p. 1002). As for the future, many suggestions have been made as to what should be included in the curriculum for an introductory computer science course (Austing and Engel, 1973; Austing et al, 1978; Bork, 1982; Cashman and Mein, 1975; Cherniak, 1976; Gibbs, 1977; Gries, 1974; Hyde, Gay and Utter, 1979; Kimura, 1979; Poirot, 1976; Ralston, 1984; Ralston and Shaw, 1980; Schneider, 1978; Singhania, 1980; Solntseff, 1978;"Task Force", 1985). Computer science has developed so much during the past twenty-five years that people are keen to examine its future development, which is almost certain to lead away from the present state within the next few years. Problem solving is a necessary skill for introductory computer scientists (Cashman and Mein, 1975; Grady and Gawronski, 1983; Gries, 1974; Hyde, Gay and Utter, 1979; Poirot, 1976; Solntseff, 1978). This skill should be taught early in the introductory course preceding computer programming (Cashman and Mein, 1975). Poirot (1976) stated that in one of his eight areas of study "the student's requirements as to problem-solving ability and programming understanding far exceed his need for proficiency in a particular language. ... problem-solving approaches should be language independent" (p.45). Teachers should emphasize problem solving and algorithm development techniques that are applicable regardless of the programming language (Grady and Gawronski, 1983). Hyde, Gay and Utter (1979) reported that an important ingredient that contributes to the success of their introductory computer course is "the utilization of a simple-to-learn interactive language (BASIC) at the beginning of the course, which allows the instructor to concentrate on teaching PSP [problem solving process] and allows the students to solve problems on the computer early in the course" (p.58). Review of the Literature Page 30 Algorithmic solutions to problems should also be taught in the introductory course (Gries, 1974; Schneider, 1978; Solntseff, 1978). Schneider (1978) believed "the single most important concept in a programming course is the concept of an algorithm" (p.108). Other important ideas worth inclusion in an introductory course involve the student devel-oping programming reading skills (Kimura, 1979) and using previously written programs (Cherniak, 1976) before being taught program writing skills. A first year computer science course needs more discrete mathematics content (Ralston, 1984). Mathematical reasoning ability and mathematical maturity are necessary for success in computer science (Ralston and Shaw, 1980). They explained that because computer science is an evolving field, specific skills learned today will soon become obsolete. However, with a knowledge of the principles that underlie these skills, a student will be free from obsolescence. In secondary school computing a broader course designed to meet the needs of all students is required, since all students will be affected by the computer (Crawford, 1978). Singhania (1980) agreed with this in his college business computing science curriculum course. Students enter this course with a tremendous variation in background, motivation, expectations and analytic skills. Thus no single emphasis will likely meet the needs of the majority of students. Self (1983) stated that programming skills should be taught as one of the five main parts of an ideal computer literacy course to children in order to narrow the socioeconomic spread between them and to give the students a higher level of self esteem and accomplishment. He said also, that skills learned in programming have carry-over value into other academic areas. Using the results of a survey involving returns from over 3500 Minnesota secondary school teachers nearly 1200 of which use computers in their classrooms, Hansen, Klassen, Anderson and Johnson (1981) have reported information on what teachers think every high Review of the Literature Page 31 school graduate should know about computers. The majority of the teachers (nearly 84%) felt that "every secondary school student should have some minimal understanding of computers" (p.468), and (nearly 93%) that "every secondary school student should learn about the role that computers play in our society" (p.468). However there was a lack of consensus among teachers on the third and final statement, where just under 29% agreed that "every secondary school student should be able to write a simple program" (p.468). Solntseff (1978) stated that an introductory course should provide a firm foundation on which the student can build his understanding of the computer as a problem-solving tool. "Probably the most important objective of an introductory programming course should be the development of positive attitudinal outcomes in students" (Lemos, 1978, p.301). Lemos used some team activities within the class in order more effectively to achieve this objective. Cole and Hannafin (1983) found that high school students often choose to enroll, or not to enroll, in introductory computer science based on inaccurate perceptions of the course. More attention needs to be directed to the clarification of the expectations of the introductory computer science coursework. Among its many recommendations a task force working under the Association for Computing Machinery (A.C.M.) proposed that there be four courses offered at the sec-ondary level: two full year courses offering an introduction to computer science, and two half year general education courses. One of these courses is to be about programming; the other, to include information about the ways computers are used and their impact on people's lives ("Task Force", 1985). These ideas must be given serious consideration when revisions are made of existing intro-ductory computer science courses at secondary and post-secondary levels. These courses should be in a constant review process because computer technology is ever-changing. Review of the Literature Page 32 Introductory C o m p u t e r P r o g r a m m i n g L a n g u a g e There is considerable debate about which computer programming language should be taught in an introductory computer science course (Bauer, 1979; Blaisdell and Burroughs, 1985; Bork, 1971; Braswell and Wadkins, 1984; Brookshear, 1985; Chanon, 1977; Charmonman and Ralston, 1975; Cherniak, 1976; Citron, 1983; Curtis, 1983; Lemos, 1979; Lucas and Kaplan, 1976; Papert, 1980; Poirot, 1979; Schneider, 1978; Self, 1983; Solntseff, 1978; "Task Force", 1985; Ulloa, 1980; Van de Riet, 1975; Williams, 1982). Charmonman and Ralston (1975) believed that "language ... is important but not nearly so important as some leading computer scientists would have you believe" (p.965). They also wrote that "the language used for teaching in a first course in computer science is much less important than how the course is taught" (p.965). Brookshear (1985) wrote that "teaching programming is more general than teaching a programming language. The former is to teach students organizational skills while the latter is to teach a particular way of organiz-ing" (p.24). The organizational skills taught do not dictate the use of a particular programming language. One of the major problems in teaching programming is trying to implant good program-ming habits in students (Ulloa, 1980). However, "no matter how careful one is to introduce clean programming habits (restricted control flow, top-down design, meaningful data names, completeness in specifications and documentation, etc.) at the very beginning of students' programming experience students too easily fall into 'quick and dirty' programming" (Cherniak, 1976, p.65). One language that is highly acclaimed as an excellent introductory programming language is Pascal, a well-structured, high-level language. (Pascal is used in introductory computer science courses at the University of British Columbia - CPSC 114, CPSC 116, CPSC 118, and CPSC 210.) After using Pascal in an introductory computing course Bauer (1979) said that his students were able to do longer and more interesting assignments than those who Review of the Literature Page 33 completed the course prior to the introduction of Pascal. He stated that students have no greater difficulty in learning Pascal than any other introductory language. Schneider (1978) had a set of ten basic principles which formed the skeletal outline of an introductory programming course. One of his principles was to "choose a programming language that enhances the learning process" (p. 109). It must be both rich in those con-structs needed for introducing fundamental programming concepts, and simple enough to be learned in an introductory course. Schneider felt that Pascal best achieves both of these objectives. Since Pascal was specifically designed for use in the teaching environment, it "is probably the strongest contender for the title of 'state-of-the-art pedagogic language'" (Solntseff, 1978, pp. 122-123). He wrote that because Pascal is a good data and action structuring language (according to him BASIC, FORTRAN, PL/1 are not), it is also usable in higher level courses. Pascal has been selected as the language for the Advanced Placement (AP) program Computer Science course. The test development committee felt that it was important for the language of the AP Computer Science course to have "characteristics that facilitate struc-tured prograniming and a high degree of modularity" (Braswell and Wadkins, 1984, p.10). Another language that is widely used, especially with microcomputers, is BASIC. It is extremely popular as the programming language first-taught at the secondary school level, both here in British Columbia, and across the continent. The National Science Foundation (N.S.F.) in a report based on the results of a survey, entitled "Computing Activities in Secondary Education" (C.A.S.E.), stated that in 1975, 62.4% of the secondary schools teaching computer science used BASIC as the programming language. "With the recent microcomputer development, this percentage has certainly increased" (Poirot, 1979, Review of the Literature Page 34 p. 103). (The writer would concur with Poirot now, ten years later.) According to Self (1983), BASIC is becoming "the universal language of the masses" (p.9). He felt that "Because of its popularity, BASIC is likely to survive until programming in the English language becomes commonly available" (p.9). Papert (1980) (author of the LOGO lan-guage) claimed that BASIC language is synonymous with programming "despite the exis-tence of other computer languages that are demonstrably easier to learn and are richer in the intellectual benefits that can come from learning them" (p.34). Some leading computer scientists are opposed to the teaching and learning of BASIC lan-guage. "Informally a great many people in computer science departments will tell you that it (BASIC) is a disaster." (Bork, 1983b). Atherton (1982b) wrote that BASIC has "serious deficiencies which have become painfully apparent" (p.9). Bork (1982b) said "Many of the people learning to program in junior high school and high school cannot overcome the initial bad habits which have often been instilled in them when they come to the universities... The main culprit is BASIC" (p. 12). "Many universities now recognize that students who learned BASIC programming in high school have a difficult time mastering the computer science curriculum" (Bork, 1985, p.31). He added that as in many other pursuits "If students are introduced to programming initially in such a way as to build up a set of bad habits associated with poor programming style, it is very difficult or even impossible to change these bad habits at a later stage " (Bork, 1983a, p.24). The Computer Science Department of the United States Air Force Academy concurred with Bork that "students who have first learned BASIC find it more difficult to learn structured languages such as Pascal" (McCann, 1983). Thus, it is critical that the proper habits be established at the outset. Bork (1983a) stated that "if students have not seen correct prograrnming style, they perhaps will not be able to continue with computers at the college level" (p.27). One might ask what is the "correct" programming style? According to Bork, the language must be "structured". He said that it is impossible to structure some versions of BASIC. He Review of the Literature Page 35 claimed that BASIC has no adequate procedures; the single letter or single letter plus single digit used for variables in some versions of BASIC prevents one from giving meaningful names to variables, which makes later program revision very difficult. Atherton (1982a) described several examples of poor features of BASIC~"spaghetti-like" structure, multiple statement lines and short variable names. Papert (1980) stated that BASIC programs acquire the structure of a labyrinth and only the brightest children are able to write non-trivial programs. Dijkstra (1968a) wrote a powerful letter to the editor on the "Go To" statement entitled "Go To Statement Considered Harmful". He stated that the "quality of programmers is a decreasing function of the density of 'go to' statements in the programs they produce" (p. 147). Dijkstra said the "use of the 'go to' statement has such disastrous effects,... the 'go to' statement should be abolished from all 'higher level' programming languages" (p. 147). BASIC language does have a "Go To" statement. In structured pro-gramming the "Go To" statement is considered to be superfluous. Dijkstra wrote further that "The 'go to' statement as it stands is just too primitive; it is too much an invitation to make a mess of one's program" (p. 147). Why then would BASIC have incorporated the "Go To" statement? Since BASIC was developed at a time when there was little experience with programming particularly for large, complex activities a "number of rules have been discovered, violation of which will either seriously impair or totally destroy the intellectual manageability of the program" (Dijkstra, 1972, p. 863). Dey and Mand (1986) wrote that there is "a widely held belief that BASIC, because of its lack of structured constructs, does not contribute significantly to learning in other computer science courses, and may in fact be detrimental" (p. 146). Some computer scientists feel that learning an unstructured language, such as BASIC, "permanently affects an individual's ability to learn a structured language; these computer scientists recommend that no unstructured languages be taught" ("Task Force," 1985, p.272). Review of the Literature Page 36 A second, and perhaps more important, reason why some leading educators in this field hold BASIC language in disfavour is because of the methods used in teaching the lan-guage. Bork (1982b) said "The major problem is the way programming is taught" (p. 12). Bork (1983a) claimed that "only a very small percentage of courses in BASIC are taught so as to promote good habits" (p.27). He (Bork, 1982b) wrote that in most instances BASIC is taught antithetically to everything else known about programming. The process of pro-gramming is very different from what it was twenty five years ago. Bork (1983a) stated that many high school courses are not teaching programming as it is being practised today, but are teaching strategies which have long since been discredited for practical program-ming. He believed that students should be taught more modern techniques. Bork (1987) wrote that BASIC can be taught in structured fashion, but it is almost never done. Christensen (1982b) stated "that as a programming language BASIC is hopelessly obsolete. It never was a very good one, but seen from a modem point of view it is a disaster. People who start to learn programming using BASIC may easily be led astray and after some time, find themselves fighting with problems that could be solved with almost no effort if they used a programming language more suited to guiding human thinking" (p.7). Carney (1983) wrote that BASIC'S poor reputation "is based on teachers who have no formal background in computer science and are simply following the obsolete programming style in BASIC that is not modularized, uses too many GOTO's, and is not at all consistent with the structured programming languages" (p.8). A third reason to avoid BASIC language in the minds of many of the experts, is BASIC'S lack of standardization. Bork (1982a) stated that there is no widespread move to have existing BASICs or newly developed BASICs conform to a standardized version of the language, which is now underway. The standard version has departed greatly from the current implementations according to Bork. He added that BASIC is one of the least stan-dardized computer languages. Review of the Literature Page 37 As an alternative to BASIC programming language Bork (1982b) has suggested LOGO or Pascal or another structured programming language. "In the earliest exposures to pro-gramming it is important to work with a structured approach, a language which encourages structured use" (Bork, 1983a, p.27). Bork added "any programming course should focus on the issues of style and structure, and these should receive greater attention than any issues of grammar" (p.27). In conclusion the three reasons outlined by Bork (1982a) why the teaching of BASIC must stop are: 1. BASIC "does not lead easily to structured programming (and) tends to develop poor programming habits", (p.33) 2. "there are better languages for the student to begin with", (p.33) and 3. "the lack of standardization in existing BASICs" (p.33). For these reasons Bork (1985b) wrote that "the better universities, particularly in courses for computer science majors, avoid BASIC 'like the plague'" (p.28). Bork's condemna-tion of BASIC as the "junk food of modem programming... (and) junk food tends to destroy the body's desire for better types of food... (so) STOP TEACHING BASIC" (1982b, p. 12) has caused the writer to believe that some study is required to investigate the differences in achievement in computer science courses between those students who have some prior knowledge of BASIC and those who do not have any BASIC knowledge. Williams (1982) suggested that perhaps a language such as BPL ("BASIC-Pascal-Liaison, or Better Programming Language or simply a conversational programming language which was designed after APL" (p.289)) which was developed from the two languages BASIC and Pascal would be the best language to use in an introductory course. This two-tiered language which combines the advantages of BASIC and Pascal could be taught effectively Review of the Literature Page 38 to both the computer science student and the non-computer science student enrolled in the same introductory course. Lemos (1979) summarized ten approaches to programming language instruction. He added that all of these approaches lack empirical evidence of their pedagogical effectiveness and need to be tested in a wide range of experimental settings. One method of instruction is structured programming, - a key feature of Pascal (Dahl, Dijkstra and Hoare,1972; Wirth, 1973). "Although there seems to be a consensus that structured programming represents an improvement in programiriing practices, there also seems to be very little empirical evidence suggesting significant improvement in terms of programmer productivity, number or types of errors made, quality of the finished program, ease of learning, or ease of teaching. This is especially true for teaching programming at the college level" (Lemos, 1979, p. 175). The Association for Computing Machinery Task Force (1985) recommended that the specific language used in a first computer science course should encourage and facilitate modular, structured programming. Pascal language or Structured BASIC language could be used. On the other hand, Lucas and Kaplan (1976) demonstrated with an experiment that struc-tured programs were easier to modify. "Well-structured programming is closely linked to well-structured problem solving" (Van de Riet, 1975, p.954). Citron (1983) claims that the teaching of problem solving is more important than the actual computer language being taught. She said that structured programming is possible with many different languages, including BASIC. Hall and Kidman (1975) claim that "there is a great need to teach stu-dent programmers working in such languages to be consciously concerned with the struc-ture and organization of their programs" (p.313). However, just because structured pro-gramming may be preferable does not mean that it is best introduced in a beginning course (Klein, 1983). Beginners do not need to be encumbered with the complexity required of Pascal language (Blaisdell and Burroughs, 1985). Review of the Literature Page 39 There appears to be no agreement among those teaching the first course as to the best intro-ductory computer language. It is possible that every programming language is being taught somewhere in an introductory computer course. A number of criteria must be met in selecting the language to be used in an introductory computing course (Van de Riet, 1975). The application to which the programming language is being used is more important in determining its appropriateness, than what language is being used (Blaisdell and Burroughs, 1985). There are advantages for a student to learn two or more contrasting programming languages in an introductory course (Bork, 1971). Learning more than one language gives the stu-dent additional insight into programming not afforded with knowledge of just a single lan-guage. Bork said "when a student is familiar with only one language, he does not under-stand that language as fully as he might because he has nothing to contrast it with" (p.4). With the knowledge of two computer languages a student is able to choose a particular lan-guage for a particular problem (Bauer, 1979). The student can compare languages mean-ingfully. Learning two computer languages in an introductory course helps a student dispel the belief held by many students who have learned only one language, that the language he has learned "constitutes the 'center' of the computing universe" (Chanon, 1977, p.42). Whatever curricula and programming language(s) are taught in introductory computer science courses, it is desirable that instructors at secondary and university levels maintain close contact with one another and be in continuous communication with prospective employers (Austing et al, 1979). Certainly, the content and language(s) taught are worthy of a great deal of attention in our ever-changing technological society. Whether or not the syllabus and programming language taught in an introductory computer science course will affect the future achievement of students needs further investigation. For this reason this study proposes to attempt to determine the effects on student achievement of various programming languages used in introductory courses. Page 40 Chapter Three Methodology The present study was conducted in order to determine how the knowledge of BASIC pro-gramming language affected the learning of further computer science material including other computer languages. To obtain the results, fifteen questions were asked. To answer these questions, data had to be gathered. Chapter Three consists of a discussion of the data collected and the procedures used to collect that data. The discussion is organized around the sample, the courses, the questionnaire, the pilot testing, the course grades, the fifteen questions asked, and the methods of analysis. S a m p l e S e l e c t i o n The decision was made to survey first and second year computer science courses. By including both years, comparisons could be made between students with similar back-grounds in a first year course and those in their second university course. Some computer science instructors have suggested that differences between groups of students would be most apparent at the introductory level. The advantages or disadvantages of having previ-ously learned BASIC would likely diminish with each successive computer science course. A questionnaire was distributed to students enrolled at U. B. C. in first or second year Computer Science courses (numbered CPSC 101, 114, 118,151, 210), or in a Computer Education course (CSED 217) during a class session held in the second week of the course. Table 1 shows the number of sections of each course surveyed. Methodology Page 41 Table 1 Number of Sections Surveyed Term CPSC CPSC CPSC CPSC CPSC CSED 101 114 118 151 210 217 1 2 3 1 1 1 4 2 1 1 1 1 1 1 Total 3 4 2 2 2 5 The questionnaire was distributed by the writer, who provided a brief oral explanation. Completed forms were collected immediately. The questionnaire took less than five minutes to complete, and most students required only two or three minutes for this task. The length of time required to complete the form depended largely on how many computer languages the student had previously learned. Table 2 shows the number of questionnaires returned from each course surveyed: Table 2 Number of Responses to the Questionnaire Course Number of Responses CPSC 101 372 CPSC 114 371 CPSC 118 127 CPSC 151 251 CPSC 210 109 CSED 217 72 Total 1302 C o u r s e D e s c r i p t i o n s The courses in computer science at U.B.C. that were surveyed were one-term in length. A term course was comprised of approximately 37 hours of lectures and 13 hours of tutorials. In Figure 1 is a description of courses involved: Methodology Page 42 Figure 1 Courses Involved CPSC 101 Introduction to FORTRAN Programming - Practical introduction to computer use. Aspects of the FORTRAN language and some common algorithms and applications. Students will compose and implement several programs. Programming style will be emphasized. Students wanting a more comprehensive introduction to Computer Science should take Computer Science 114 and 116. CPSC 114 Principles of Computer Programming I - An introduction to the structure and use of digital computers. Concepts of algorithm, program and programming. Principles of program design using Pascal on terminals. Students will compose and implement several programs. In these exercises, emphasis will be placed on clarity and orderly development. CPSC 118 Principles of Computer Programming - Accelerated version of Computer Science 114 and 116, assuming some prior knowledge and experience of computer programming. Systematic study of structured programming in Pascal; data representation, algorithm design, introduction to computer organization. Omits introduction to FORTRAN. Students having previous experience with Pascal should take Computer Science 116 instead of 118. (Note: Computer Science 114 is a prerequisite for Computer Science 116.) CPSC 151 Introduction to FORTRAN Programming - A practical introduction to computer use. Aspects of structured FORTRAN and some common algorithms and applications. Introduction to Michigan Terminal System (MTS). Programming style will be emphasized. Intended for Applied Science students only. (Note: Credit is given for only one of Computer Science 101,114, and 151.) CPSC 210 Computer Program Design I - Programming techniques of intermediate sophistication. Information structures and algorithms which operate on them. Students will undertake a programming project. Prerequisite: Computer Science 116 or 118. (Note: All second year Computer Science students must take this course.) CSED 217 Microcomputer Programming I - Uses of microcomputers in education; introduction to programming in BASIC. Methodology Page 43 A l l of the specified courses were given during both terms. Note therefore, that a student could possibly have taken more than one course of computer science during the university year. For students who completed more than one questionnaire, only the first response was recorded. Q u e s t i o n n a i r e The students involved in the study completed a one-page questionnaire in order to provide the data required to answer the questions involved in this study. Figure 2 shows the ques-tionnaire used: Methodology Page 44 Figure 2 The Questionnaire cs Sec LEARNING COMPUTER LANGUAGES This questionnaire i s part of a study to analyze the trends in programming languages. If you choose to participate your mark in this course w i l l be obtained from the registrar as part of t h i s study. Responses w i l l be held c o n f i d e n t i a l . Tour instructor w i l l see a summary of the results obtained but w i l l not be able to iden t i f y individual students. If you do not wish to participate in t h i s study, check t h i s bcx £1 OBC Student Number facu l t y Year (1st, Majors_ 2nd, ...) Age If you have never learned a computer language, check t h i s box and hand in your questionnaire. You are finished. Three s p e c i f i c languages are l i s t e d below plus the category, "Other", to cover languages other than the three s p e c i f i e d . Please c i r c l e , in column (l), any of the languages you have learned and provide the following information for each language learned: In column (2): Has t h i s language the 1st language you learned? the 2nd? In column (3): How well do you know this language now: 1 - vaguely familiar 2 - capable of reading 3 - able to write simple programs (<60 lines) 4 - quite familiar 5 .- able to write complex programs <>200 lines) In column ( 4 ) : At what point(s) in your l i f e did you learn this language? In column (S): Where did you learn t h i s language? — r n — Computer language learned BASIC — m — Order of learning 1st [] 2nd [] 3rd [] l a t e r U — — m How well do you know? (check one) 1 [] 2 [] 3 [] 4 [ ] 5 [] TTi When did you learn? (check one or more) Age up to 13 [] (Elementary school) Age 13 to 13 [ ] (Secondary school) (5) Where did you learn? (check one or more) School, Univ. [ ] Home L ] Job [ ] Other [] Age 18 or more [I Pascal 1st t l 2nd [] 3rd t l l a t e r t ] 1 n 2 t i 3 [] 4 [ ] 5 [] Age up to 13 N (Elementary school) Age 13 to 18 U (Secondary school) School, Home Job Other Univ. [] [ j tl [] Age 18 or more' LOGO i s t n 2nd [] 3rd [] l a t e r t ] ( ] [ ] n Age up to 13 [ ] (Elementary school) Age 13 to 18 [ ] (Secondary school) School. Univ. [] Home [ ] Job [ ] Other t] Age 18 or more Other (specify) 1st [] 2nd [] 3rd [] latert 1 [ ] [ ] [ ] [ ] n Age up to 1 3 [ ] (Elementary school) Age 1 3 to l 3 [ ] (Secondary 3 c h o o l ) School, Home Job Other Univ [ ] [ ] [ ] [ ] Age 13 or more tl Thank you for your time. Methodology Page 45 Permission to distribute the questionnaire in class was requested, and it was granted by each of the ten instructors involved. This was followed by a letter (see Appendix A) with a copy of the questionnaire to be used. The first item on the questionnaire was used to obtain the consent of each student to be involved in the study. For those students who did not wish to be involved in the study nothing else was required to be written on the questionnaire. These students could then return the questionnaire to the writer. Otherwise, students were to write their university student number (in order to facilitate the matching of the questionnaire with the final mark), their year enrolled at university, their age, their sex, their Faculty and their major(s). This information was requested so as to make comparisons among different years, ages, Facul-ties and majors and between sexes. Those students indicating they had never learned a computer language were finished with the questionnaire. These could then be returned to the writer. Those students who had learned (a) computer language(s) were to indicate if the lan-guage )^ learned was (were) BASIC, Pascal, LOGO, or "Other", and to respond to the following questions about each language specified. 1) Students were to indicate the numerical order in which the listed languages had been learned - first, second, third, or later; this was included because the writer thought that there might possibly be some achievement difference between students who learned BASIC as their first language and those who learned BASIC other than as the first language. The data concerning the order in which the computer languages were learned were gathered because of the assertions of Bork and Dijkstra that learning BASIC first causes students to acquire bad programming habits which cannot be overcome. 2) The students were to check how well they knew each of these languages on a five-point scale from being (1) "vaguely familiar" to being (5) "able to write complex programs" of more than 200 lines in length. Such data were requested because the achievement might Methodology Page 46 possibly have been affected by how well previous languages had been learned. The criteria for the five-point scale was decided upon because of the distinct ability groupings. 3) Respondents were to answer how old they were when each language was learned. Three age ranges were given: age up to 13 (elementary school), age 13 to 18 (secondary school), and age 18 or more. They could make more than one selection for each language. The data were needed because the achievement might possibly have been affected by how long ago previous languages had been learned. These age ranges were used since it was felt to be the easiest classification for the respondent. 4) Students were asked where each language had been learned. Four selections were given for choices - at school or university, at home, on a job, or "other". If students chose "other" they were to indicate where the language had been learned. The "location" data were needed since there may have been an achievement difference due to the different envi-ronments in which the previous languages were learned. Pilot T e s t i n g Before the study began the questionnaire was tested on nine Grade 12 secondary school students enrolled in a Computer Science 12 course to see if the instructions on the ques-tionnaire were clear and unambiguous, and to learn how much time was required to com-plete the questionnaire. The pilot sample was chosen since it was readily available at the school where the writer taught. One disadvantage of the pilot test was that for the pilot students some of the items were inappropriate e.g. Faculty, majors. However, every stu-dent had learned BASIC language. The writer asked the students to specify which items (if any) were unclear to them; there were none. The time required by the students to complete the questionnaire was carefully recorded. The questionnaires were collected and reviewed by the writer. No changes were made in the questionnaire after the pilot testing. Methodology Page 47 Course Grades At the conclusion of Term 2 a computer file was set up containing information for each student who had been enrolled in the courses of this study. Two important fields of information in the file were the final computer science course grade converted to a percent-age (to permit easier comparisons and interpretations of results), and the average (in per-centage) for all courses completed during the two terms of the study. (The course mark in computer science was used as a measure of computer science achievement.) Other fields contained the number of courses in which the student was enrolled, the number of com-puter courses studied, the standing for the student in all courses taken and the current status of the student within the university. The course grades and year average (in percentage) were identified by student number so that they could be matched with the responses to the questionnaire. Individual student grades were kept in the strictest confidence. The responses from the questionnaire sorted by student number were coded and entered into a file on the U.B.C. Amdahl computer. The file containing the grades was then merged with the questionnaire file. Student num-bers with information missing were excluded from the study. The coding scheme is included in Appendix B. Questions The results of questionnaires returned were collected as data in order to help answer the following questions. BASIC Language Backgrounds Five questions pertain to comparisons among groups with various BASIC backgrounds. Methodology Page 48 1 How does introductory computer science achievement compare between the group of students whose first computer language was BASIC and the group who knew no languages? 2 Does achievement differ between the group with prior knowledge of BASIC and the group who had prior knowledge of computer languages other than BASIC? 3 Does achievement differ between the group with prior knowledge of BASIC only and the group who had prior knowledge of BASIC and another language? 4 Of all students with prior knowledge of BASIC does achievement differ between the group who had learned BASIC first and the group who learned another language be-fore learning BASIC? 5 Are there differences in achievement among the groups of students who have already learned 0, 1, 2, 3, or 4 computer languages? Demographic Backgrounds Four questions deal with comparisons among groups with different backgrounds, other than computer languages learned. 6 Is there a difference in achievement by gender? 7 Are there differences in achievement among various age groups? 8 Are there differences in achievement among the Faculties represented? 9 Is there a difference in achievement between students who are majoring in mathematics and students who are majoring in various fields within the Faculty of Arts (i.e. non-mathematics majors)? Different Language Backgrounds Two questions are concerned with comparisons among groups enrolled in an introductory computer science course with different computer language backgrounds, and among similar groups in a second year computer science course. Methodology Page 49 10 Is there a difference in achievement among students with different computer language backgrounds enrolled in the first year FORTRAN (CPSC 101 or CPSC 151) or Pascal (CPSC 114 or CPSC 118) courses? 11 Is there a difference in achievement among students with different computer language backgrounds enrolled in the second year course (CPSC 210)? BASIC and Pascal Backgrounds Three questions pertain to comparisons among groups with different BASIC language backgrounds, and among groups with different Pascal language backgrounds. 12 Are there differences in achievement among the groups that respond in five different ways regarding "how well" a previous language (either BASIC or Pascal) was learned? 13 Are there differences in achievement among the groups that respond in three different ways regarding "when" either BASIC or Pascal was learned? 14 Are there differences in achievement among the groups that respond in four different ways regarding "where" either BASIC or Pascal was learned? Achievement Predictors One question deals with the relationship between computer science achievement and a set of predictor variables. 15 Which factor(s) under investigation is/are the best predictor(s) of success in university introductory courses in computer science? Methods of Analysis In order to test the hypotheses used to answer the first fourteen questions stated in the study an analysis of covariance was used. The effect of certain variables on the computer science course mark (MARK) was investigated while the influence of the percentage obtained in all other courses (PERCENT) was statistically controlled. PERCENT was cal-Methodology Page 50 culated by removing the computer science course mark from the year average for all courses. The data were analyzed using the multivariate analysis of variance and covariance procedure (MANOVA) contained in the Statistical Package for the Social Sciences (SPSS). With a single dependent variable and covariate MANOVA was chosen instead of the analy-sis of covariance procedure (ANCOVA) because the interrelations among the independent variables were desired. The SPSS ANCOVA routine does not provide such post-hoc comparisons. Group means were adjusted for the covariate (PERCENT). Both the "Actual Mean" and the "Adjusted Mean" are reported for each test. If the independent vari-able tested had more than two levels then a square matrix of contrast coefficients for this factor was entered in order to observe the between-group comparisons. The matrix con-tained as many rows and columns as there were levels of the factor. The coefficients entered in a row indicated the desired comparison between the groups of the factor. For example, the contrast coefficients 10-1 mean that the effect of group 1 is to be com-pared with the effect of group 3 on the dependent variable. The contrast coefficients 1/2 1/2 -1/3 -1/3 -1/3 mean that the effect of the average of groups 1 and 2 is to be compared with the effect of the average of groups 3,4 and 5. Each set of contrast coefficients summed to zero, and each row was not a linear combination of any other rows. Only those between-group comparisons thought to be of interest for this study in the view of the writer were entered. On occasion in testing some hypotheses more than one matrix of coefficients was entered. Also, on some occasions superfluous comparisons were entered in order to build the required square matrix. These extra comparisons were not reported. For each MANOVA an omnibus F ratio was obtained. If this value was not significant, then the results of the between-group comparisons were not significant. If the omnibus F was sig-nificant, then the F ratio given for each complex contrast was compared with a critical F-table value (Kirk, 1968). A table containing the above information for each statistical test required to answer questions 1 to 14 is provided in Chapter Four. Methodology Page 51 To answer Question 15 a multiple linear regression equation was formed using the various factors as predictor scores and the computer science course grade as the criterion score using the SPSS subprogram REGRESSION. The best linear prediction equation and its prediction accuracy for calculating MARK were desired. To this end, a stepwise selection method of multiple linear regression was used to analyze the relationship between a set of 10 predictors and a single criterion variable. A significance level of 0.05 was chosen. Post-hoc comparisons were made using Scheffe's procedure (Kirk, 1968) with a significance level of 0.10. Ferguson (1981) and Scheffe (1959) both recommend a more liberal level of significance. Although the omnibus F significance level was set at 0.05 an experiment error rate greater than 0.05 existed since so many statistical tests were performed. £age 52 Chapter Four Descriptive Analyses Two types of data were collected in this study. First was the responses to the questionnaire described in Chapter Three. Second was the final course mark obtained by 1602 students enrolled in introductory computer science courses at the University of British Columbia during either the first or second term of the 1985-86 winter session. These marks were obtained via magnetic tape from the Office of the Registrar. The questionnaire was answered by 1302 students enrolled in introductory computer science courses. Responses to the questionnaire were not received from 365 students who had received a course mark. Of these missing responses: -44 students indicated that they did not wish to be involved in this study; -48 students had incomplete or missing student numbers recorded on the question-naire; -273 students did not return the questionnaire, were absent at the time of the question-naire distribution, or were not yet enrolled in the computer science course at the time the questionnaire was administered. The students enumerated above were not included in the study since no background infor-mation was known about them. On the other hand, 55 students had completed a question-naire but, for various reasons, no course mark was assigned for them: 21 of the 55 were females, 22 had learned BASIC language, and between 3 to 13 students inclusive, were from each of the six courses in the study. These were removed from the study once it was determined that no common characteristics could be found among them. Eight students completed a questionnaire in both first and second term. Only the first completed questionnaire was used because the study focuses on the first computer science course taken at the university. In addition, two students were enrolled in two first term computer courses; both completed two questionnaires. One of these students received marks in both Results Results Page 53 courses and was excluded from the study. The other student had a mark for only one of the two courses and was retained in the study. The two data types, questionnaire data and marks, were merged by matching student numbers resulting in 1236 records. Of the 1236 records, 42 students had a final mark recorded as 0 given in the computer science course. This indicates that they had unofficially dropped out of the course. These were excluded and a final sample of 1194 remained. In the sample the number of students enrolled in each of the surveyed courses is shown in the Table 3: Table 3 Distribution of Students in Computer Courses Course Number Number of Percent of Students Sample CPSC 101 (FORTRAN) 347 29.1 CPSC 151 (FORTRAN) 245 20.5 CPSC 114 (Pascal) 338 28.3 CPSC 118 (Pascal) 116 9.7 CPSC 210 (Pascal) 95 8.0 CSED 217 (BASIC) 53 4.4 Total 1194 100.0 There were approximately equal numbers of students in the FORTRAN courses numbered 101 and 151 (592 students, 49.6% of the sample) compared to the Pascal courses num-bered 114, 118, and 210 (549 students, 46% of the sample). Table 4 shows the years in which the students were enrolled. The majority were in first and second year (82.7%). Results Page 54 Table 4 Distribution of Students by Year of Registration Year Number of Percent of Students Sample 1st 555 46.5 2nd 432 36.2 3rd 104 8.7 4th 62 5.2 5th 28 2.4 6th 4 0.3 7th 3 0.3 8th 1 0.1 10th 1 0.1 Not stated 4 0.3 Total 1194 100.1" a Column totals more than 100.0 percent due to rounding. The median age of the respondents was 19 years. The mean age was 20.2 years. There were 329 female students (27.6%) and 859 males (71.9%) and six missing values (0.5%). The students were distributed across a number of Faculties. Table 5 shows the Faculties in which the students were enrolled. Table 5 Distribution of Students by Faculty Faculty Number of Percent of Students Sample Science 365 30.6 Applied Science 322 27.0 Commerce & Business Administration 199 16.7 Arts 196 16.4 Education 59 5.0 Agricultural Sciences 21 1.8 Forestry 5 0.4 Graduate Studies 3 0.3 Unclassified 20 1.7 Not stated 4 0.3 Total 1194 100.2* a Column totals more than 100.0 percent due to rounding. Results Page 55 The Faculties of Science, Applied Science, Arts, and Commerce and Business Ad-ministration accounted for more than 90% of the sample. Figure 3 shows the percentage of students in the sample who had learned one or more computer language prior to entering this course. Figure 3 Distribution of Students with Prior Knowledge of a Computer Language • No Languages • One or more Languages Figure 4 shows the percentage of students in the sample who had learned some BASIC language prior to entering this course. Results Page 56 Figure 4 Distribution of Students with Prior Knowledge of BASIC Computer Language • No BASIC H Knew BASIC Of the 757 students who had previously learned some BASIC language, 698 students had learned BASIC as their first computer language (92.2% of those knowing BASIC). For these students the order of their learning BASIC language is summarized in Table 6, below. Table 6 Order of Learning BASIC for Students with Prior Knowledge of BASIC Language Order of learning BASIC Number of Percent of Students Total First Language 698 92.2 Second Language 40 5.2 Third Language 12 1.6 Fourth Language 4 0.5 Not stated 3 0.4 Total 757 99.9a Column totals less than 100.0 percent due to rounding. Results Page 57 Most of the students who had learned BASIC indicated that they knew at least enough BASIC in order to "write simple programs" up to 60 lines (596 out of 757 students -78.7%) and that they learned BASIC in school (628 out of 757 students - 83.0%) and were between the ages of 13 and 18 when they first learned BASIC (625 out of 757 students -82.6%). BASIC was learned at home by 201 students (26.6% of 757 students). Of these, 105 had not learned any BASIC at school. The students' experiences with BASIC are summarized in Tables 1 - 4 of Appendix C. Figure 5 shows the percentage of students in the sample who had learned some Pascal lan-guage prior to entering this course. Figure 5 Distribution of Students with Prior Knowledge of Pascal Computer Language H Knew Pascal H No Pascal Of the 339 students who had previously learned some Pascal language, 178 students had learned Pascal as their second computer language (52.5% of those knowing Pascal). For these students the order of their learning Pascal language is summarized in Table 7, below. Results Page 58 Table 7 Order of Learning Pascal for Students with Prior Knowledge of Pascal Language. Order of learning Pascal Number of Percent of Students Total First language 54 15.9 Second language 178 52.5 Third language 85 25.1 Fourth language 19 5.6 Not stated 3 0.9 Total 339 100.0 Of the students who had learned Pascal, 103 out of 339 students (30.4%), indicated that they were able to "write complex programs" of more than 200 lines - the highest classifica-tion possible for the response to "How well do you know?" on the questionnaire. Of the "Pascal-knowledgeable" students, 188 out of 339 students (55.5%), had learned Pascal at age 18 or more and 309 out of 339 students (91.2%) at school/university . No student had learned Pascal before age 13. The students' experiences with Pascal are summarized in Tables 5 - 8 of Appendix C. Only 71 students (6% of the sample) had learned the LOGO language. Nine of these stu-dents had learned LOGO as their first language (12.7% of 71 students). The majority of LOGO learners indicated that they were not "quite familiar" with LOGO, nor "able to write complex programs" (63 out of 71 students - 87.7%). LOGO had been learned between ages 13 and 18 by 61 out of 71 students (85.9%) and 42 "LOGO-knowledgeable" students (59.1% of 71 students) had learned LOGO at school. The students' experiences with LOGO are summarized in Tables 9 - 12 of Appendix C. Figure 6 shows the number of students in the sample who knew a language and which lan-guage was learned first. Results Page 59 Figure 6 First Learned Computer Language Other 63 Figure 7 shows a Venn diagram of the number of students in the sample who had learned either BASIC, Pascal or LOGO languages or any combinations of these three. Of the 805 students who had previously learned one of these languages, 57 knew all three languages, and no student knew just LOGO language. Results Page 60 Figure 7 BASIC, Pascal, LOGO Languages Learned A language other than BASIC, Pascal, or LOGO had previously been learned by 299 stu-dents (25% of the sample). The "Other" computer languages known included the fol-lowing: FORTRAN, Assembler, Machine Language, PL/1, Forth, COBOL, DBase 2, APL, C, Comal and others. The most commonly known "Other" computer language for this group was FORTRAN (170 out of 299 students - 56.9%). This "Other" language had been learned as the first computer language by 63 out of 299 students (21.1%). Most of the "Other" language-knowledgeable students indicated that they knew at least enough of this language in order to "write simple programs" up to 60 lines (222 out of 299 students -74.2%). The majority of these "Other" language-knowledgeable students learned this lan-guage at "age 18 or more" (179 out of 299 students - 59.9%), and at school (233 out of 299 students - 77.9%). The students' experiences with "Other" languages are summarized in Tables 13 - 16 of Appendix C. Results Page 61 Figure 8 indicates which computer language had been learned first by the 835 students in the sample who had learned a language prior to entering this course. Figure 8 Languages First Learned by 835 Students with Prior Knowledge of a Computer Language • BASIC H Other Languages H LOGO 111 Pascal • Not stated Figure 9 shows the order of learning each computer language involved in this study for the 835 students who had previously learned a language. (A few students indicated that they had learned a language but they did not indicate the order of learning that language.) Results Page 62 Figure 9 Order of Learning each Computer Language for Students with Prior Knowledge 698 BASIC Pascal LOGO Other Computer Language Learned • 1st H 2nd IH 3rd • 4th From this sample of 1194 students the number of computer science or computer science education courses taken is shown in Table 8. Table 8 Distribution of Students by Number of Computer Science Courses Taken Current Number of Number of Percent of Computer Courses Students Sample 1 855 71.6 2 278 23.3 3 42 3.5 4 9 0.8 5 7 0.6 6 3 0.3 Total 1194 100.1" a Column totals more than 100.0 percent due lo rounding. Results Page 63 Of 1194 students only 61 students took more than two computer courses during the year, the equivalent of two terms or two half-year courses. The majority (71.6%) took only the one half-year course. The computer language background of students varied gready between computer science courses. Figure 10 shows the percentage of students in each surveyed course that had previously learned a computer language. Figure 10 Distribution of Students with Prior Knowledge of a Computer Language by Course 100% 75% Percent of Students 5 Q % with Prior Language 25% 0% All 95 students in CPSC 210 had learned Pascal because the prerequisite for this course is the completion of either CPSC 118 or CPSC 114 and CPSC 116. All 116 students in CPSC 118 had learned a language since prior knowledge of, and experience with pro-gramming was assumed as an entrance requirement. Figures 11-14 show the percentages 100% 100% CPSC 101 CPSC 151 CPSC 114 CPSC 118 CPSC 210 CSED 217 Course Results Page 64 of students in each course that had previously learned BASIC, Pascal, LOGO, and "Other" languages. Figure 11 Distribution of Students with Prior Knowledge of BASIC Language by Course 100% T 75% Percent of Students 50% Knowing B A S I C 25% 69% 65% , ' w j ! * ' <. < ' 1 85% 85% hi t i • plffll Iliill i M f Pitilll ' 43% CPSC 101 CPSC 151 CPSC 114 CPSC 118 CPSC 210 CSED 217 Course Figure 12 Distribution of Students with Prior Knowledge of Pascal Language by Course 100% 75% Percent of Students 50% Knowing Pascal 25% 0% 6% 25% 25% 100% I"'" -j, ' i i 55% I titlN & \ 25% CPSC 101 CPSC 151 CPSC 114 CPSC 118 CPSC 210 CSED 217 Course Results Page 65 Figure 13 Distribution of Students with Prior Knowledge of LOGO Language by Course. 100% 75% • Percent of Students 50% Knowing 1 X 3 0 0 25% 2% 7% 7% 9% 14% 4% 0% CPSC 101 CPSC 151 CPSC 114 CPSC 118 CPSC 210 CSED 217 Course Figure 14 Distribution of Students with Prior Knowledge of Other Languages by Course. CPSC 101 CPSC 151 CPSC 114 CPSC 118 CPSC 210 CSED 217 Course A computer language had been learned previously by 57.4% of the females (189 students) and 74.6% of the males (641 students). Figure 15 shows the percentages of each gender that had previously learned a computer language. Results Page 66 Figure 15 Percentages by Gender with Prior Knowledge of Computer Languages 57.40% 42.60% 74.60% 25.40% Females Males Prior Knowledge No Prior Knowledge Thus, males in this study had a "greater" entry level background of computer languages than females. Achievement in Introductory Computer Science Courses The average (mean) mark obtained in the computer science course by 1194 students in the sample was 68.9%. The average mark obtained in all other courses for those who took more than one course was 64.8%. (Only eight students in the sample of 1194 students were enrolled in just one course). A Student t-test for dependent means was conducted of this difference in mean scores. It was found to be significant (t=10.5, p<0.001). See Table 9. That is, students performed significantly better in their computer science course than in their other courses. This difference can be accounted for by many factors - students who worked harder in a course of interest, students who had previous experience in com-puter science, and others. Results Page 67 Table 9 Comparison of Marks in Computer Science Course with Averages in all Other Courses Number Mean Standard Correlation T 2-Tail of Cases Deviation Value Probability Mark in Surveyed Course 1186 68.9 14.9 0.534 10.50 p< 0.001 Mark in Other Courses 1186 64.8 12.7 Inferential Analyses To answer the first 14 questions stated in Chapter Three analyses of covariance were conducted with computer science course percent (MARK) as the dependent variable, the percentage obtained in all other courses (PERCENT) as the covariate, and many different factors depending upon the question, used as independent variables. An omnibus F significance level of 0.05 was used. Post-hoc comparisons were made at the 0.10 significance level. Many of these questions were analyzed using the entire sample, the introductory FORTRAN (CPSC 101 and CPSC 151) students only, and the introductory Pascal (CPSC 114 and CPSC 118) students only. The second and third analyses were done to compare differences that might occur among students enrolled in introductory FORTRAN courses with the differences among those registered in introductory Pascal courses. Multiple linear regression analysis was used with the fifteenth question. Results Page 68 Results on BASIC Language Backgrounds The first five questions deal with comparisons of achievement of a group having BASIC language knowledge with that of groups who have learned other languages, and with the order in which the BASIC was learned in relation to other languages. BASIC versus No Language How does introductory computer science achievement compare between the group of stu-dents whose first computer language was BASIC and the group who knew no languages? Hypothesis 1: There is no significant difference in computer science course achievement between the group of students who entered the course with BASIC learned as their first computer language and the group who entered without a knowledge of any computer language. A significant difference in favour of the group having had BASIC instruction was found at the 0.05 level between the two groups (F=86.20, p<0.001). See Table 10. In the first year FORTRAN (CPSC 101 and CPSC 151) courses the group of students who had learned BASIC language had a significantly greater achievement than the group who had not yet learned a language (F=58.40, p<0.001). Similarly, in the first year Pascal (CPSC 114 and CPSC 118) courses those who had learned BASIC as their first language had sig-nificantly greater achievement than those who had not yet learned a language (F=37.05, p<0.001). Thus the null hypothesis was rejected. Results Page 69 Table 10 Comparison of Results of Students Who Learned BASIC as the First Language and Students Who Knew No Languages Groups: Learned BASIC first Never learned a language Total  Number of 698 359 1057 Students Actual Mean 71.0 64.7 68.9 Adjusted Mean 71.5 64.2 F (1,1054) = 86.20, p<0.001  BASIC versus Other Languages Does achievement differ between the group with prior knowledge of BASIC and the group who had prior knowledge of computer languages other than BASIC? Hypothesis 2: There is no significant difference in achievement between the group who had learned BASIC and the group who had learned languages other than BASIC. A significant difference in favour of the group having had BASIC instruction was observed at the 0.05 level between the two groups (F=4.69, p=0.031). See Table 11. Hypothesis 2 was rejected. However, no significant differences were found between the two groups in either the FORTRAN (CPSC 101 and CPSC 151) courses (F=0.03, p=0.872) or the introductory Pascal (CPSC 114 and CPSC 118) courses (F=2.62, p=0.107). Results Page 70 Table 11 Comparison of Results of Students Who Had Learned BASIC and Students Who Had Learned Other Languages Groups: Learned BASIC Learned languages Total other than BASIC  Number of 757 78 835 Students Actual Mean 70.9 69.0 70.7 Adjusted Mean 71.5 68.4 F (1,832) = 4.69, p=0.031  BASIC Only versus BASIC and Other Does achievement differ between the group with prior knowledge of BASIC only and the group who had prior knowledge of BASIC and another language? Hypothesis 3: There is no significant difference in achievement between the group who had learned BASIC only and the group who had learned BASIC and an-other language. A significant difference in favour of the group having learned BASIC and at least one addi-tional language was found at the 0.05 level in course achievement between the two groups (F=8.57, p=0.004). See Table 12. Similarly, in the FORTRAN (CPSC 101 and CPSC 151) courses (F=6.44, p=0.012) and in the Pascal (CPSC 114 and CPSC 118) courses (F=23.57, p<0.001) the group of students who had learned BASIC and another language outperformed those who had learned BASIC language only. Hypothesis 3 was rejected. Results Page 71 Table 12 Comparison of Results of Students Who Had Learned Only BASIC and Students Who Had Learned BASIC and Another Language Groups: Learned BASIC only Learned BASIC plus other languages Total Number of Students 377 380 757 Actual Mean 69.5 72.2 70.9 Adjusted Mean 69.6 72.1 F (1,754) = 8.57, p=0.004 BASIC First versus Not BASIC First Of all students with prior knowledge of BASIC does achievement differ between the group who had learned BASIC first and the group who learned another language before learning Hypothesis 4: There is no significant difference in achievement between the group who learned BASIC as their first computer language and the group who learned another language before learning BASIC. No significant difference was obtained at the 0.05 level between the group of students who had learned BASIC first and the group of students who had learned BASIC other than as their first language (F=0.12, p=0.724). See Table 13. No significant differences were found between the two groups in either the FORTRAN (CPSC 101 and CPSC 151) courses (F=1.10, p=0.295) or the first year Pascal (CPSC 114 and CPSC 118) courses (F=0.99, p=0.32). Hypothesis 4 was not rejected. BASIC? Results Page 72 Table 13 Comparison of Results of Students Who Had Learned BASIC First and Students Who Had Learned BASIC Other than as a First Language Groups: Learned BASIC first Learned BASIC Total but not first ' Number of 698 56 754 Students Actual Mean 71.0 69.9 70.9 Adjusted Mean 70.7 70.2 F (1,751) = 0.12, p=0.724  Number of Computer Languages (0-4) Are there differences in achievement among the groups of students who have already learned 0, 1,2, 3, or 4 computer languages? Hypothesis 5: There are no significant differences in achievement among the groups of students who have previously learned 0,1,2, 3, or 4 computer languages. Significant differences were found at the 0.05 level among the groups of students who previously learned 0, 1, 2, 3, or 4 computer languages (F=25.38, p<0.001). Generally, the more languages learned, the higher the course mark. The group of students with no prior language experience had significandy lower achievement than the average of the other four groups with either one, two, three, or four previous languages learned (F=99.88, p<0.10). The average mark of those who had learned more than one language was signifi-cantly greater than that of those who did not know a language or had learned just one lan-guage (F=58.84, p<0.10). The group of students who had previously learned one lan-guage obtained significandy higher achievement than the group without the knowledge of a prior language (F=45.73, p<0.10). See Table 14. Hypothesis 5 was rejected. Results Page 73 Table 14 Comparison of Results of Students with Various Number of Computer Languages Learned Groups: No previous One previous Two Three Four Total language language languages languages languages  Number of 360 434 201 165 34 1194 Students Actual Mean 64.7 69.7 70.6 71.7 78.7 68.9 Adjusted Mean 63.9 69.8 71.6 71.0 79.1 Contrast F Coefficients: -1 1/4 1/4 1/4 1/4 99.88* -1/2 -1/2 1/3 1/3 1/3 58.84* 1 -1 0 0 0 45.73* 0 1 -1 0 0 2.99 omnibus F (4,1 188) = 25.38, p<0.001 * p<0.10, F =7.88 Results on Demographic Backgrounds The next four questions are concerned with differences in achievement among demographic backgrounds: gender, age, Faculty and majors. Gender Is there a difference in achievement by gender? Hypothesis 6: There is no significant difference in achievement by gender. A significant difference in achievement was found at the 0.05 level between males and females in the courses observed (F=15.51, p<0.001). See Table 15. Males had signifi-candy greater achievement than females. Males had significandy higher computer science course marks at the 0.05 level: Results Page 74 -in the FORTRAN (CPSC 101 and CPSC 151) courses (F=17.40, p<0.001); -in the introductory Pascal (CPSC 114 and CPSC 118) courses (F=5.53, p=0.019); -in the group of all five introductory courses (F=14.37, p<0.001). There were no significant gender differences in achievement at the 0.05 level in the second year course (CPSC 210) (F=0.02, p=0.898). Hypothesis 6 was rejected. Table 15 Comparison of Student Results by Gender Groups: Female Male Total Number of 329 859 1188 Students Actual Mean 68.1 69.1 68.9 Adjusted Mean 67.0 70.2 F (1,1185) = 15.51, p<0.001  Age Are there differences in achievement among various age groups? Hypothesis 7: There are no significant differences in achievement among age groups. Students were divided into three groupings by age: ages 17-19, ages 20-24, and ages 25 and over, from the ages as recorded on the questionnaires. Significant achievement dif-ferences were found at the 0.05 level among the age groups (F=6.40, p=0.002). The youngest group had significantly greater achievement than the oldest group (F=10.61, p<0.10), and significantly greater achievement than the average of the other two groups combined (F=12.76, p<0.10). See Table 16. Hypothesis 7 was rejected. Results Page 75 Table 16 Comparison of Student Results by Age Age Groups 17-19 20-24 25 and over Total Number of 708 378 99 1185 Students Actual Mean 69.6 68.1 66.2 68.9 Adjusted Mean 70.0 68.3 65.7 Contrast F Coefficients: -1 0 1 10.61* 1 -1 0 4.50 -1 1/2 1/2 12.76* 0 1 -1 3.61 omnibus F (2,1181) = 6.40, p=0.002 * p<0.10, F' = 4.66 Faculty Are there differences in achievement among the Faculties represented? Hypothesis 8: There are no significant differences in achievement among the Faculties represented. Significant MARK differences were observed at the 0.05 level among the Faculties repre-sented (F=15.48, p<0.001). Significant differences in achievement were observed between the Faculty of Applied Science and the average of the other faculties represented (F=37.60, p<0.10). The Faculty of Applied Science students also had greater achievement than each of: the Faculty of Arts students (F=62.14, p<0.10), the Faculty of Education students (F=47.19, p<0.10), the Faculty of Science students (F=21.50, p<0.10). Excluding the Faculty of Applied Science significant differences in achievement were found between the Faculty of Commerce and Business Administration and the average of the other faculties represented (F=14.03, p<0.10). The Faculty of Commerce and Business Results Page 76 Adrninistration students also had higher marks than the Faculty of Arts students (F=27.67, p<0.10) and the Faculty of Education students (F=28.50, p<0.10). Students in the Faculty of Science had significantly higher achievement than students in the Faculty of Arts (F=16.36, p<0.10). See Table 17. Hypothesis 8 was rejected. Table 17 Comparison of Student Results by Faculty Faculties Science Arts Commerce & Business Admini-stration Applied Science Forestry and Agriculture Edu-cation Graduate Studies/ Unclass-ified Total Number of 365 196 199 322 26 59 23 1190 Students Actual Mean 66.4 63.0 73.0 73.4 71.4 62.8 71.8 68.9 Adjusted Mean 69.5 65.2 71.7 73.9 70.3 62.1 69.3 Contrast F Coefficients: 1/6 1/6 1/6 -1 1/6 1/6 1/6 37.60* 1/5 1/5 -1 0 1/5 1/5 1/5 14.03* 1 -1 0 0 0 0 0 16.36* -1 0 1 0 0 0 0 3.84 0 0 -1 0 0 1 0 28.50* 0 1 0 -1 0 0 0 62.14* 0 -1 1 0 0 0 0 27.67* 0 0 0 -1 0 1 0 47.19* 1 0 0 -1 0 0 0 21.50* omnibus F (6,1182) = 15.48, p<0.001 * p<0.10, F' = 10.8 Major Is there a difference in achievement between students who are majoring in mathematics (enrolled in either the Faculty of Arts or the Faculty of Science) and students who are majoring in various fields within the Faculty of Arts (i.e. non-mathematics majors)? Results Page 77 Hypothesis 9: There is no significant difference in achievement between the group of stu-dents who are majoring in mathematics and the group who are majoring in various fields within the Faculty of Arts. The group of students majoring in mathematics had significandy greater achievement than the group majoring in various Departments and Schools within the Faculty of Arts (F=6.19, p=0.015). See Table 18. A complete listing of the majors indicated by these Faculty of Arts students with the number for each major, is contained in Table 19. In the introductory Pascal (CPSC 114 and CPSC 118) courses the mathematics majors signifi-cantly outperformed the non-mathematics majors (F=9.59, p=0.003). In the FORTRAN (CPSC 101 and CPSC 151) courses no statistical testing was done because of the imbal-ance in the number of students who were mathematics majors (n=l) and non-mathematics majors (n=29). A reason for this imbalance is that the CPSC 114 course or CPSC 118 course is a requirement for a Bachelor of Science degree as a mathematics major. Hypothesis 9 was rejected. Table 18 Comparison of Student Results for Selected Majors Groups: Math majors Non-math majors Total Number of Students 48 53 101 Actual Mean 66.6 61.5 63.9 Adjusted Mean 67.1 61.0 F (1,98) = 6.19, p=0.015 Results Page 78 Table 19 Number of Non-mathematics Majors in the Faculty of Arts Major Number of Students Asian Studies 1 Dietetics 2 Economics 26 English 1 International Relations 1 Languages, Linguistics 7 Law 1 Music 1 Political Science 1 Psychology 11 Sociology 1 Total 53 Results on Different Language Backgrounds The next two questions deal with achievement differences among students with different computer language backgrounds enrolled in an introductory FORTRAN or Pascal course, and achievement differences among students with different computer language back-grounds enrolled in a second year computer science course. It might be speculated that previous programming experience might have an effect on achievement in a course taught in one programming language, but have little effect on achievement in another course employing a different computer language. Different Backgrounds in First Year Is there any difference in achievement among students with different computer language backgrounds enrolled in the first year (a) FORTRAN (CPSC 101 or CPSC 151) or (b) Pascal (CPSC 114 or CPSC 118) courses? Hypothesis 10(a): There is no significant difference in achievement among students with different computer language backgrounds enrolled in an introductory FORTRAN course. Results Page 79 In order to investigate this question, students were grouped according to their previous experience with computer languages. The groups were composed of students who had learned: 1. BASIC language only; 2. BASIC language first, and at least one other language; 3. BASIC subsequent to learning at least one other language; 4. at least one language, but not BASIC language; 5. no computer languages yet. A significant achievement difference was realized at the 0.05 level among these five groups of students with different computer language backgrounds (F=17.08, p<0.001). See Table 20. Students who knew no computer languages prior to the course (Group 5) had signifi-cantly lower achievement than the average of those who had some computer language background (F=37.76, p<0.10). Students who had no prior language experience (Group 5) achieved significantly less than those who knew BASIC only (Group 1) (F=35.63, p<0.10), and significantly less than those who had learned BASIC as their first language as well as at least one more language (Group 2) (F=51.93, p<0.10). No significant differ-ence was found between students who had learned BASIC as their first language (Groups 1&2) and those who had learned a language other than BASIC as their first language (Groups 3&4) (F=0.00). Hypothesis 10(a) was rejected. Results Page 80 Table 20 Comparison of Results Based on Computer Language Background in the FORTRAN (CPSC 101 and CPSC 151) Courses Groups (1) BASIC (2) BASIC first (3) BASIC (4) At least 1 (5) No Total only and at least after at least language computer 1 other 1 other but not languages language language BASIC Number of 235 87 10 25 234 591 Students Actual Mean 73.5 75.6 74.5 74.8 67.1 71.3 Adjusted Mean 72.8 76.8 75.3 74.1 66.5 Contrast F Coefficients: 1/4 1/4 1/4 1/4 -1 37.76* -1 0 0 0 1 35.63* 1/3 1/3 1/3 -1 0 0.10 0 1 0 0 -1 51.93* 1/2 1/2 -1/2 -1/2 0 0.00 0 0 -1 0 1 5.83 omnibus F (4,585) = 17.08, p<0.001 * p<0.10, F' = 7.88 Hypothesis 10(b): There is no significant difference in achievement among students with different computer language backgrounds enrolled in an introductory Pascal course. A significant achievement difference was realized at the 0.05 level among the same five groups of students with different computer language backgrounds (F=15.53, p<0.001). See Table 21. Students who knew no computer languages prior to the course (Group 5) had significandy lower achievement: -than the average of those who had some computer language background (F=28.32, p<0.10); -than those who knew BASIC only (Group 1) (F=8.92, p<0.10); -than those who had learned BASIC as the first language and at least one other language (Group 2) (F=54.76, p<0.10); Results Page 81 -than those who had learned BASIC subsequent to learning at least one other language (Group 3) (F=15.53, p<0.10). Also, the group of students who had learned BASIC language first, and at least one other language (Group 2) had significantly higher achievement than the group of students who had learned BASIC language only (Group 1) (F=20.42, p<0.10). Again, no significant difference was found between students who had learned BASIC as their first language (Groups 1&2) and those who had learned a language other than BASIC as their first language (Groups 3&4) (F=0.03). Hypothesis 10(b) was rejected. Table 21 Comparison of Results Based on Computer Language Background in the Pascal (CPSC 114 and CPSC 118) Courses Groups (1) BASIC (2) BASIC first (3) BASIC (4) At least 1 (5) No Total only and at least after at least language computer 1 other 1 other but not languages language language BASIC Number of 130 169 19 29 104 451 Students Actual Mean 62.5 70.3 72.1 66.2 59.2 65.3 Adjusted Mean 64.3 70.7 71.4 64.3 59.6 Contrast F Coefficients: 1/4 1/4 1/4 1/4 -1 28.32* -1 0 0 0 1 8.92* 1/3 1/3 1/3 -1 0 3.44 0 1 0 0 -1 54.76* 1/2 1/2 -1/2 -1/2 0 0.03 0 0 -1 0 1 15.53* -1 1 0 0 0 20.42* omnibus F (4,445) = 15.53, p<0.001 * p<0.10, F = 7.88 Results Page 82 Different Backgrounds In Second Year Is there any difference in achievement among students with different computer language backgrounds enrolled in the second year course (CPSC 210)? Hypothesis 11: There is no significant difference in achievement among students with different computer language backgrounds enrolled in the second year course (CPSC 210). In order to investigate this question, students were grouped according to their previous experience with computer languages. The groups were composed of students who had learned: 1. BASIC language first, and at least one other language; 2. BASIC subsequent to learning at least one other language; 3. at least one language, but not BASIC language. No significant differences in achievement were found at the 0.05 level among these three student groupings of different computer language backgrounds (F=2.39, p=0.097). See Table 22. Hypothesis 11 was not rejected. Results Page 83 Table 22 Comparison of Results Based on Computer Language Background in the Second Year (CPSC 210) Course Groups (1) BASIC first (2) BASIC after (3) At least one Total and at least one at least one language but other language other language not BASIC Number of 67 14 14 95 Students Actual Mean 73.4 67.2 60.5 70.6 Adjusted Mean 70.2 67.6 63.3 Contrast F Coefficients: -1 0 1 4.60 1/2 1/2 -1 2.84 1 -1 0 0.69 0 1 -1 1.09 -1 1/2 1/2 3.75 omnibus F (2,91) =  2.39, p=0.097 * p<0.10, F' = 4.78 Results of BASIC and Pascal Backgrounds The following three questions are concerned with the achievement of students with specific background information relating to both BASIC and Pascal languages. " H o w W e l l " L a n g u a g e w a s L e a r n e d Are there any differences in achievement among the groups that respond in five different ways regarding "how well" a previous language [either (a) BASIC or (b) Pascal] was learned? Hypothesis 12(a): There are no significant differences in achievement among the groups that respond in five different ways regarding "how well" BASIC was learned. Students had responded to the question "How well do you know (BASIC, Pascal, etc.)?", by selecting one of the following five choices for each language previously learned: Results Page 84 1. vaguely familiar 2. capable of reading 3. able to write simple programs (<60 lines) 4. quite familiar 5. able to write complex programs (>200 lines) Different results were found for BASIC and Pascal languages. For those students with prior knowledge of BASIC significant differences in computer science achievement were found among the five groups at the 0.05 level (F=8.27, p<0.001). See Table 23. Those "able to write complex programs" (Group 5) had signifi-cantly greater achievement than the average of those who were not able (F=29.43, p<0.10). The group of students who had entered this course with the ability "to write complex programs" (Group 5) in BASIC had significandy higher achievement than the group who upon entering were either "vaguely familiar" (Group 1) with BASIC language (F=21.84, p<0.10), "capable of reading" (Group 2) BASIC programs (F=14.63, p<0.10), or "able to write simple programs" (Group 3) in BASIC language (F=15.87, p<0.10). Similarly, in the FORTRAN (CPSC 101 and CPSC 151) courses (F=6.77, p<0.001) and in the Pascal (CPSC 114 and CPSC 118) courses (F=8.44, p<0.001) students who were "able to write complex programs" achieved significandy higher than those who were not "able to write complex programs". Hypothesis 12(a) was rejected. Results Page 85 Table 23 Comparison of Results Based on "How Well" BASIC Language had been Previously Learned Groups (1) Vaguely (2) Capable (3) Able to write (4) Quite (5) Able to write familiar of simple familiar complex Total reading programs programs Number of 80 71 202 157 237 747 Students Actual Mean 66.4 67.6 69.7 70.9 74.2 70.8 Adjusted Mean 66.8 67.8 69.4 70.9 73.9 Contrast F Coefficients: 1/4 1/4 1/4 1/4 -1 29.43* -1 0 0 0 1 21.84* 1/3 1/3 1/3 -1 0 5.96 0 1 0 0 -1 14.63* -1 0 0 1 0 6.34 0 0 -1 0 1 15.87* omnibus F (4,741) = 8.27, p<0.001 * p<0.10, F' = 7.88 Hypothesis 12(b): There are no significant differences in achievement among the groups that respond in five different ways regarding "how well" Pascal was learned. For those students with prior knowledge of Pascal no significant differences were found at the 0.05 level among the same five groups on "how well" Pascal language had previously been learned (F=1.72, p=0.145). See Table 24. Similarly, in the FORTRAN (CPSC 101 and CPSC 151) courses (F=1.12, p=0.356) and in the Pascal (CPSC 114 and CPSC 118) courses (F=1.63, p=0.169) no significant differences were found at the 0.05 level among the five groups. Hypothesis 12(b) was not rejected. Results Page 86 Table 24 Comparison of Results Based on "How Well" Pascal Language had been Previously Learned Groups (1) Vaguely (2) Capable (3) Able to write (4) Quite (5) Able to write familiar of simple familial complex Total reading programs programs Number of 59 34 72 54 103 322 Students Actual Mean 69.0 70.8 70.6 71.4 74.8 71.8 Adjusted Mean 68.9 71.8 69.9 72.5 73.5 Contrast F Coefficients: 1/4 1/4 1/4 1/4 -1 3.25 -1 0 0 0 1 5.22 1/3 1/3 1/3 -1 0 1.37 0 1 0 0 -1 0.45 -1 0 0 1 0 2.44 0 0 -1 0 1 3.53 -1 0 1 0 0 0.23 0 0 0 -1 1 0.22 omnibus F (4,316) = 1.72, p=0.145 * p<0.10, F = 7.88 "When" Language was Learned Are there any differences in achievement among the groups that respond in three different ways regarding "when" either (a) BASIC or (b) Pascal was learned? Hypothesis 13(a): There are no significant differences in achievement among the groups that respond in three different ways regarding "when" BASIC was learned. Students had responded to the question, "When did you learn (BASIC, Pascal, etc.)?", by indicating one or more of the following three choices for each language previously learned: 1. Age 18 or more 2. Age 13 to 18 (Secondary school) 3. Age up to 13 (Elementary school) For students who selected more than one choice the youngest age group selected was used. Results Page 87 For those students with prior knowledge of BASIC significant differences in achievement were found at the 0.05 level among the various groups (F=3.16, p=0.043). See Table 25. Students who first learned BASIC at age 13 to 18 had significandy greater achievement than those who first learned BASIC at age 18 or more (F=6.31, p<0.10). Hypothesis 13(a) was rejected. Table 25 Comparison of Results Based on "When" BASIC Language had been Previously Learned Age Groups: Age 18 or More Age 13 to 18 Age up to 13 Total Number of 103 625 25 753 Students Actual Mean 67.8 71.4 70.7 70.8 Adjusted Mean 68.0 71.2 70.6 Contrast F Coefficients: -1 0 1 0.95 1 -1 0 6.31* -1 1/2 1/2 2.91 1/2 1/2 -1 0.16 omnibus F (2,749) = 3.16, p=0.043 * p<0.10, F' = 4.66 Hypothesis 13(b): There are no significant differences in achievement among the groups that respond in three different ways regarding "when" Pascal was learned. For those students with prior knowledge of Pascal significant differences at the 0.05 level were found among the age groups for "when" Pascal was previously learned (F=6.88, p=0.009). See Table 26. Students who first learned Pascal at age 13 to 18 had signifi-cantly greater achievement than those who first learned Pascal at age 18 or more. No stu-dents in this study had previously learned Pascal prior to age 13. Hypothesis 13(b) was rejected. Results Page 88 Table 26 Comparison of Results Based on "When" Pascal Language had been Previously Learned Age Group Age 18 or More Age 13 to 18 Total Number of 188 148 336 Students Actual Mean 70.6 73.5 71.9 Adjusted Mean 70.3 73.8 F (1,333) = 6.88, p=0.009  "Where" Language was Learned Are there any differences in achievement among the groups that respond in four different ways regarding "where" either (a) BASIC or (b) Pascal was learned? Hypothesis 14(a): There are no significant differences in achievement among the groups that respond in four different ways regarding "where" BASIC was learned. Students had responded to the question "Where did you learn (BASIC, Pascal, etc.)?" by selecting one or more of the following four choices for each language previously learned: 1. School, university 2. Home 3. Job 4. Other Similar results were found for both BASIC and Pascal languages. For those students with a prior BASIC language background no significant difference in achievement was observed among the various groups at the 0.05 level among the "places" where various groups of students had learned BASIC (F=1.70, p=0.147). See Table 27. Hypothesis 14(a) was not rejected. Results Page 89 Table 27 Comparison of Results Based on "Where" BASIC Language had been Previously Learned Groups Job Home School School and Home Other Total Number of 4 104 529 91 4 732 Students Actual Mean 74.7 72.9 70.4 72.5 69.3 71.0 Adjusted Mean 69.3 74.0 71.6 74.2 70.7 Contrast F Coefficients: -1 1/4 1/4 1/4 1/4 0.30 -1 0 0 0 1 0.03 1/3 1/3 -1/2 -1/2 1/3 0.30 0 -1 0 1 0 0.01 0 0 -1 1 0 3.82 1 0 -1 0 0 0.15 0 1 -1 0 0 3.68 omnibus F (4,726) = 1.70, p=0.147 * p<0.10, F' = 7.88 Hypothesis 14(b): There are no significant differences in achievement among the groups that respond in four different ways regarding "where" Pascal was learned. For those students who were previously knowledgeable in Pascal no significant differences were observed at the 0.05 level among the "places" where various groups of students had learned Pascal (F=2.38, p=0.07). See Table 28. Hypothesis 14(b) was not rejected. Results Page 90 Table 28 Comparison of Results Based on "Where" Pascal Language had been Previously Learned Groups Job Home School School and Home Total Number of 2 21 286 21 330 Students Actual Mean 80.7 76.9 71.3 73.8 71.9 Adjusted Mean 79.3 77.2 71.2 75.0 Contrast F Coefficients: -1 1/3 1/3 1/3 0.31 0 -1 0 1 0.33 0 0 -1 1 1.97 1 0 -1 0 0.89 0 1 -1 0 4.79 0 1 -1/2 -1/2 1.87 omnibus F (3,325) = 2.38, p=0.07 * p<0.10, F' = 6.33 Results of Achievement Predictors The final question is concerned with checking for a linear relationship between the com-puter science achievement and a set of predictor variables. Achievement Predictors Which factor(s) under investigation is/are the best predictor(s) of success in university introductory courses in computer science? Hypothesis 15: Success in introductory computer science courses is unrelated to background characteristics. The percentage earned in all other courses, the year of studies enrolled in, age, gender, the number of computer languages previously learned, the number of computer courses presently enrolled in, whether any computer language had been learned, whether BASIC language had been learned, the order in which BASIC was learned and how well BASIC had been learned, were used as predictor variables in a multiple linear regression equation. Results Page 91 Of these variables only the overall year percent (PERCENT) and how well BASIC had been learned (BHOWELL) were found to be significant predictors of achievement (F=306.992, p<0.0001). See Table 17 in Appendix C. A total of 34.3% of the variance in achievement was accounted for by these two variables. The prediction equation for achievement (MARK) was: MARK = 23.862 + 0.633(PERCENT) + 1.827 (BHOWELL) Hypothesis 15 was rejected. Page 92 Chapter Five Discussion This study was undertaken to determine how prior knowledge of BASIC language affects achievement in introductory computer science courses at university. It also compared achievement among groups which differed in gender, age, Faculty and major. The study assessed the differences in achievement in first year FORTRAN and Pascal courses and in a second year computer science course among groups with different computer language backgrounds. It ascertained the impact on achievement of "how well," "when" and "where" a previous computer language was learned. The study identifies factors that are the best predictors of success in introductory courses of computer science. A total of 1194 university students who responded to a questionnaire and completed an in-troductory computer science course provided the data for the study. These data were ana-lyzed using an analysis of covariance procedure for each of the independent variables in-volved. The results of the analyses reveal a relationship between prior knowledge of BASIC language and achievement in introductory university computer science courses. Summary and Conclusions The following paragraphs present a summary of the findings together with conclusions based on these findings. Limitations of the study are indicated. Finally, implications of the results are discussed and suggestions for future research are presented. BASIC Language Backgrounds The results showed that the group of students whose first computer language was BASIC achieved significandy better marks than the group who had no prior knowledge of a com-puter language. This finding suggests that: Discussion Page 93 Students who have taken BASIC do better in university intro-ductory level computer science courses than those who have no prior knowledge of a computer language. This finding is not unexpected in view of the fact that, in general, prior related-knowledge enhances further learning. However, it is in disagreement with the opinions of some computer educators referred to in Chapter Two. Other languages were not examined individually to see how previous knowledge of them compared to no prior knowledge of a computer language. These comparisons were not done because this study is primarily interested in how the learning of BASIC affects further achievement in computer science courses. For a language like Pascal to have an influence on achievement in introductory level university computer science courses it may have to be learned in more depth than BASIC language (see page 99). Relatively few of the sample in this study knew Pascal well. The group who had learned either one, two, three or four previous languages, had signifi-cantly greater achievement than the group with no prior language experience. The group who had previously learned two or more languages had significantly greater achievement than the group who had learned fewer than two languages. The group who had previously learned a computer language achieved significantly better than the group who had not learned a language. In general, the findings lead to a conclusion that: Students who have previously learned a computer language have better achievement in an introductory computer science course than those who have not learned a language. Also, students with a knowledge of at least two languages have an even higher achievement in introductory computer science. This conclusion would suggest that there are some skills which overlap languages which when learned with the first language can be transferred to the learning of additional languages. Discussion Page 94 In concurrence with this conclusion, the group who had learned BASIC and at least one other language achieved significantly better than the group who had learned BASIC lan-guage only. A significant difference in achievement was found between the group with prior knowledge of BASIC and the group with previous knowledge of another language but no prior knowledge of BASIC. This finding leads to the conclusion that: Prior knowledge of BASIC is as beneficial to achievement in university introductory computer science courses as prior knowledge of any other language identified in this study. The number of languages learned may contribute to the achievement difference. While 50% of the students who had learned BASIC had learned two or more languages, 42% of those who had learned another language had knowledge of more than one language. For students who had previously learned only one language an analysis of covariance revealed no significant difference in achievement (F=0.12, p=0.734) between those who had learned BASIC and those who had learned some other language. No significant differences in achievement were found between the group who had learned BASIC as their first language and the group who had learned BASIC other than as their first language. This finding suggests that: The order in which BASIC is learned may not be a critical factor in subsequent achievement in computer science. While the sequence in learning computer languages is likely to be educationally significant, in this study the sequence was not found to be statistically significant; this may be attributable, at least in part, to the relatively small number of students who learned BASIC after the prior learning of another language. For those questions about BASIC language backgrounds no differences were observed between the group enrolled in the FORTRAN (CPSC 101 and CPSC 151) courses and the Discussion Page 95 group enrolled in the introductory Pascal (CPSC 114 and CPSC 118) courses. That is, differences that were found to be significant among the groupings of FORTRAN students were also found to be significant among the same groupings of Pascal students. Demographic Backgrounds It was found that across all six courses males had higher computer science marks than females. In general, males have greater achievement than females in computer science courses. This finding may be due to the fact that, in general, males in this study had a "greater" entry level background of computer languages than females. It may also be due to the mark dif-ferences between courses, or between Faculty entry requirements (as mentioned below). The youngest age group (ages 17-19) had significantly better achievement than the oldest group (ages 25 and over) and had significantly better achievement than the other two age groups combined. Younger university students tend to have higher achievement than older students in introductory computer science courses. This result may be attributable to the fact that the youngest age group has only recently ac-quired their computer background while the older groups may have forgotten the computer science material they learned much earlier, or because of the ever-changing content in com-puter science, perhaps the material learned by the older groups has been drastically revised. Furthermore, the younger students may have had a greater opportunity to use computers than older students. Students from the Faculty of Applied Science had significantly greater achievement than students from all of the other Faculties within this study. This finding may be due, at least in part, to the preponderance (87.9%) of males within the Faculty of Applied Science. Discussion Page 96 Students from the Faculty of Commerce and Business Administration earned significandy higher marks in computer science courses than students from all of the other Faculties, ex-cluding the Faculty of Applied Science. It should be noted that students admitted to the Faculty of Applied Science and the Faculty of Commerce and Business Administration have higher admission requirements than do the Faculties of Arts, Education, and Science. The Faculty of Applied Science requires stu-dents entering from first year Science to have achieved at least 60% in each of Mathematics, Physics and Chemistry. The minimum standing for admission to the Faculty of Commerce and Business Administration is 60% in pre-Commerce studies. The other three Faculties mentioned have a 50% minimum entrance requirement in all courses. Inter-faculty comparisons are detailed in Chapter Four. They reveal, for example, that Faculty of Science students score significandy higher than Faculty of Arts students.Once again, the preponderance (85.2%) of males within the Faculty of Science may account, at least in part, for the superior performance over students in the Faculty of Arts (males, 51%). The results of these inter-faculty comparisons lead to the conclusion that: There are differences among Faculties in achievement in introductory computer science courses. The results of inter-faculty comparisons must be interpreted carefully since different courses and different instructors were involved. The group of students majoring in mathematics had significandy higher grades in computer science than the group majoring in other fields within the Faculty of Arts. Students majoring in mathematics outperform students taking non-mathematics majors within the Faculty of Arts in intro-ductory computer science courses. Discussion Page 97 The requirement that mathematics courses be included in the degree program of all students majoring in computer science is evidence that mathematical ability is considered to be an aptitude for subsequent learning in computer science. Different Language Backgrounds An analysis of achievement in first year FORTRAN courses, first year Pascal courses, and a second year computer science course for each of five student groups, viz., students that had learned: a. only the BASIC language b. BASIC first and at least one other language c. some other language first and BASIC later d. at least one language but not BASIC e. no language found that there were significant differences in first year FORTRAN courses: 1 between the students that had learned only the BASIC language and the students that had not learned any language, (a versus e) 2 between the students that had learned BASIC first and at least one other language and the students that had not learned any language at all, (b versus e) and 3 between the students who knew a language and those who did not know any language. (a,b,c,d versus e) The students that had not learned any language had significantly lower achievement. These findings suggest that: Students having a knowledge of any computer language per-form at a higher level in first year FORTRAN courses than others who have not learned any computer language. Similarly, in first year Pascal courses the students in each of the first-named groups significantly outperformed the corresponding second group: 1 between the students that had learned only the BASIC language and the students that had not learned any language, (a versus e) Discussion Page 98 2 between the students that had learned BASIC first and at least one other language and the students that had not learned any language, (b versus e) 3 between the students that had learned some other language first and BASIC later and the students that had not learned any language, (c versus e) 4 between the students that had learned BASIC first and at least one other language and the students that had learned only the BASIC language, (b versus a) and 5 between the students who knew a language and those who did not know any language. (a,b,c,d versus e) These findings suggest that: Students having a knowledge of any computer language per-form at a higher level in first year Pascal courses than other students who have not learned any computer language. These findings for both FORTRAN courses and Pascal courses further substantiate the earlier findings that students with a knowledge of a computer language have higher achievement than those with no previous knowledge of a computer language. In the second year computer science course no significant differences were found based on the different language backgrounds of the students, (i.e. the languages learned or the order in which they are learned). These findings suggest that: Achievement in the second year computer science course does not appear to be dependent upon the computer language backgrounds prior to entering university. In summary, prior knowledge of BASIC language enhances the achievement in first year computer science courses, generally, and in first year Pascal courses the knowledge of another computer language besides BASIC further enhances achievement. BASIC and Pascal Backgrounds For students with a prior knowledge of BASIC the group who indicated that they were "able to write complex programs" had significandy higher achievement than the groups who were "vaguely familiar", "capable of reading", or "able to write simple programs" in Discussion Page 99 BASIC language. On the other hand, for students with a prior knowledge of Pascal there were no significant achievement differences found among the groups based on "how well" Pascal language had been learned. It should be noted that these results were unchanged in separate analyses of achievement in introductory Pascal and FORTRAN courses. These findings suggest that: Students who are able to write complex BASIC programs per-form at a higher level in introductory computer science courses than others who have a limited familiarity with the BASIC lan-guage. and that: For students who had prior knowledge of Pascal language the level of achievement in introductory computer science courses did not appear to be dependent upon how well that language had been learned. These conclusions indicate that if previous knowledge of BASIC language is to be of maximum benefit to future computer science course achievement then BASIC must be well- learned. Since the level to which both BASIC and Pascal had been previously learned was stricdy a subjective measurement these conclusions must be interpreted cautiously. For the students who previously knew either BASIC or Pascal language there were signifi-cant differences in achievement among the three age groups selected to answer "when" BASIC or Pascal language had been learned. Students in the age group 13 to 18 had sig-nificantly higher achievement than those in the age group 18 or more. These findings sug-gest that: The age at which BASIC or Pascal is first learned is a critical factor and that students in introductory computer science courses who first learned the language in the 13 to 18 age range outperform others who first learned the language at an older age. This outcome may be due to these students having used the language for a longer period of time prior to university, and/or having learned other computer languages in the intervening years. Discussion Page 100 There were no significant differences in achievement among the four groups chosen to answer "where" either the BASIC or Pascal language had been learned. This finding sug-gests that: The place where either BASIC or Pascal is learned does not appear to be a critical factor for achievement in introductory computer science courses. In summary, the level to which BASIC language had been learned and the age at which both BASIC and Pascal had been learned proved to be significant factors in computer science achievement. The place at which both BASIC and Pascal had been learned and the level to which Pascal had been learned proved not to be significant factors in computer science achievement. Achievement Predictors Of the ten variables used as predictors of achievement in a multiple linear regression equa-tion only the overall year percent in all other courses taken (PERCENT) and the variable on "how well" BASIC had been learned (BHOWELL) were found to be significant predictors of achievement. However, only 34% of the variance in achievement was accounted for by these two variables. The prediction equation for achievement was: Achievement = 23.862 + 0.633 (PERCENT) + 1.827 (BHOWELL) Implications The teaching of BASIC language prior to university appears to be beneficial and should be continued. This is contrary to the belief held by some computer science educators cited in Chapter Two. The results of this study indicate clearly that a previous knowledge of BASIC language is preferable to no knowledge of a computer language when the achievement in introductory computer science courses is being compared. Furthermore, Discussion Page 101 the advantage of having previous knowledge of the BASIC language is enhanced by having the knowledge of another language. It should be noted also that students who have acquired the skill to write complex programs in BASIC outperform others who have only a limited familiarity with the language. The findings imply that the order of learning BASIC language and the place of learning BASIC do not affect subsequent achievement in computer science. This study has focused on factors of statistical significance. Attention should also be directed towards their educational significance. Clearly, there are some factors in this study that failed to stand up to the statistical test and the reader may want to consider the educa-tional implications, if any, of these aspects of achievement being considered. For example, the sequence in learning computer languages may be educationally, significant. Similarly, some factors which were statistically significant may not be of educational significance, e.g. some of the differences in results between Faculties may not be educationally significant because of other factors mentioned earlier. L i m i t a t i o n s A limitation of this study is that it is being conducted at only one university, and in only one provincial educational system and for just one school year. Furthermore, the sample selected may not best represent the population of people who have knowledge of computer languages. Responses were not received from 365 students (22.8%) of the 1602 who re-ceived a course mark. Perhaps the questionnaire should have been administered later in each term after the course change deadline date, so as to have included an even larger per-centage of students. Another limitation of the study is that students did not randomly select the courses involved. For example, CPSC 151 enrolled exclusively Faculty of Applied science stu-Discussion Page 102 dents, and 79% of the students in CSED 217 were from the Faculty of Education. For this reason, inter-faculty comparisons of achievement have been interpreted conservatively. Another potential problem is that the subjects may be assigned final grades on somewhat different bases. Thus, it is reasonable to assume that there will be some differences in the student achievement attributable to the differences in grading practices between courses or among instructors of any one course. The group of students who entered the university with no knowledge of any computer language may not have had as much interest in com-puters as that of the group who have learned at least one language already. The group of students enrolled in the second year computer science course are likely to be those who have performed well in the first year course(s), have a "keen" interest in computers, and may be planning to major in computer science. Achievement is probably influenced by the interest shown in the subject and affected by the mark earned in a prerequisite course. Suggestions for Further Research The study needs to be replicated and expanded to include students from more than just one provincial school system, more than one university and more than one school year. It could include those who program computers who have never enrolled in university com-puter science courses. It could perhaps involve students from first year computing courses and continue with students in computing courses from each year until graduation. It might also include professional programmers with many years of computer experience. (For such a study, a different measure of achievement would have to be employed). This might well provide insight as to whether the alleged "set of bad programming habits" from learning BASIC first can be overcome. There is also a pressing need to embark on similar studies to assess the impact of prior learning of other competing initial languages (such as Pascal, Comal, C, etc.) on subse-Discussion Page 103 quent achievement in computer science, and to compare the benefits with those found with BASIC. Further study is desirable to compare gender differences in achievement in computer science courses, since in this study males were found to have significantly higher achieve-ment in first year courses, but not in the second year (CPSC 210) course. This finding may be due to the relatively small number of females (11) enrolled in the course. This study can provide information that may be useful in designing subsequent research about the relationships between various factors and computer science achievement. This study may suggest covariates of achievement that might be selected for further investigation. In view of the above limitations this investigation should be considered exploratory and the findings tentative. In the opinion of the author the apparent benefits of a prior orientation with the BASIC language are sufficiently worthwhile to merit further research. Page 104 Bibliography Adams, J.M., and Haden, D.H., Introductory service courses in the Computer Science curriculum. ACM SIGCSE Bulletin. Mar. 1972,4(1), 49-52. Agee, R., Questions & answers. Computers in Education. June 1985, 2(10), p. 13. Alspaugh, C.A., A study of the relationships between student characteristics and proficiency in symbolic and algebraic computer programming (Doctoral dissertation, University of Missouri, 1970). Dissertation Abstracts  International. 1971,21, 4627-B (University Microfilms No. 71-3301,99). (Abstract) Alspaugh, C.A., Identification of some components of computer programming aptitude. Journal for Research in Mathematics Education. Mar. 1972, 3_, 89-98. American Psychological Association. Publication manual of the American Psychological  Association (2nd ed.). Washington, D.C: Author, 1974. Astin, A.W., Predicting academic performance in college. New York: The Free Press, 1971. Atchison, W.F., Computer Science preparation for secondary teachers. ACM SIGCSE  Bulletin. Feb. 1973, 5.(1), 45-47. Atherton, R., BASIC damages the brain. Computer Education. Feb. 1982, pp. 14-17. (a) Atherton, R., Structured programming with COMAL. Chichester, England: Ellis Horwood Limited, 1982. (b) Austing, R.H., and Engel, G.L., A Computer Science course program for small colleges. Communications of the ACM. Mar. 1973,16, 139-147. Austing, R.H. et al (Eds.), Curriculum 78 recommendations for the undergraduate program in Computer Science. A report of the ACM Curriculum Committee on Computer Science. Communications of the ACM. Mar. 1979, 22,147-166. Bibliography Page 105 Barker, R.J., and Unger, E.A., A predictor for success in an introductory programming class based upon abstract reasoning development. ACM SIGCSE Bulletin. Feb. 1983,15(1), 154-157. Bateman, G.R., Predicting performance in a basic computer course. Proceedings of the 5th  Annual Meeting of the American Institute for Decision Sciences. 1973, pp. 130-133. Bauer, M.A., Experiences with PASCAL in an introductory course. ACM SIGCSE  Bulletin. Feb. 1979, 11(1), 158-161. Bauer, R., Mehrens, W.A., and Vinsonhaler, J.F., Predicting performance in a computer programming course. Educational and Psychological Measurement. 1968,28. 1159-1164. Beck, J.J. Jr., The effects on attitude of anticipated Computer-Assisted Instruction in selected high school courses of study. AEDS Journal. Spring 1979,12(3). 138-145. Blaisdell, J.H., and Burroughs, A., How to tell if a programming language is OK: What's wrong with BASIC for teaching Business students how to program? ACM  SIGCSE Bulletin. Sept. 1985, 17(3), 5-8. Bork, A.M., Learning to program for the Science student. Journal of Educational Data  Processing. 1971, 8(5), 1-5. Bork, A., Learning with computer. Bedford, Massachusetts: Digital Press, 1981. Bork, A., Computers and learning. Educational Technology. Apr. 1982, 22(4), 33-34. (a) Bork, A., The fourth revolution - Computers and learning. Paper presented at the University of Oregon, Corvallis, Oregon, June 1982. (b) Bork, A., Modern approaches to learning to program. Proceedings of the 21st Annual  Convention of the AEDS. May 1983, pp. 24,27,28. (a) Bork, A., Personal communication, Dec. 5, 1983. (b) Bork, A., The computer in education in the United States: The perspective from the Educational Technology Center. Computers & Education. 1984, 8, 335-341. Bibliography Page 106 Bork, A., Computers and information technology as a learning aid. Education & Computing. Jan. 1985,1(1), 25-35. (a) Bork, A., Personal computers for education. New York: Harper & Row, Publishers, 1985. (b) Bork, A., Learning with personal computers. New York: Harper & Row, Publishers, 1987. Bork, A., Pomicter, N., Peck, M., and Veloso, S., Toward coherence in learning to program. AEDS Monitor. Sept./Oct. 1985, 24(3&4), 16-18. Braswell, J., and Wadkins, J., Pascal and the AP exam: ETS replies. Electronic Learning. Feb. 1984, 3(5), 8-10. Brookshear, J.G., The university Computer Science curriculum: Education versus training. ACM SIGCSE Bulletin. Mar. 1985, 17(1), 23-30. Buff, R.J., The prediction of academic achievement in FORTRAN language programming courses (Doctoral dissertation, New York University, 1972). Dissertation  Abstracts International. 1972, 33 ,2191-A (University Microfilms No. 72-26587,138). (Abstract) Butcher, D.F., and Muth, W.A., Predicting performance in an introductory Computer Science course. Communications of the ACM. Mar. 1985, 28, 263-268. Cafolla, R., Piagetian formal operations and other cognitive correlates of achievement in computer programming. Journal of Educational Technology Systems. 1987-88, 16, 45-55. Campbell, P.F., and McCabe, G.P., Predicting the success of freshmen in a Computer Science major. Communications of the ACM, Nov. 1984, 27, 1108-1113. Capstick, C.K., Gordon, J.D., and Salvadori, A., Predicting performance by university students in introductory computing courses. ACM SIGCSE Bulletin. Sept. 1975, 7(3), 21-29. Carney, R., Teaching BASIC. Electronic Learning. Mar. 1983, 2(6), p.8. Bibliography Page 107 Cashman, W.F., and Mein, W.J., On the need for teaching problem-solving in a Computer Science curriculum. ACM SIGCSE Bulletin. Feb. 1975, 7(1), 40-46. Chanon, R.N., An experiment with an introductory course in Computer Science. ACM  SIGCSE Bulletin. Aug. 1977,2(3), 39-42. Charmonman, S., and Ralston, A., Structured FORTRAN and the first course in Computer Science. In O. Lecarme and R. Lewis (Eds.), Computers in Education (pp. 965-970). New York: North Holland, 1975. Cheney, P., Cognitive style and student programming ability: An investigation. AEDS  Journal. Summer, 1980, 13.(4), 285-291. Cherniak, R., Introductory programming reconsidered - A user-oriented approach. ACM  SIGCSE Bulletin. Feb. 1976, 8(1), 65-68. Christensen, B.R., Beginning COMAL. Chichester, England: Ellis Horwood Limited, 1982. (a) Christensen, B.R., Programming languages for beginners and the global challenge. Computer Education. Feb. 1982. p.18. (b) Citron, J., Computer education in times of recession: Should Pascal come first? Computers & Education. 1983,7, 149-152. Cole, D.D., and Hannafin, M.J., An analysis of why students select introductory high school computer coursework. Educational Technology. Apr. 1983,23.(4), 26-29. Correnti, R.J., Predictors of success in the study of computer programming at two-year institutions of higher education. (Doctoral dissertation, Ohio University, 1969). Dissertation Abstracts International. 1970, 30, 3718A-3719A (University Microfilms No. 70-4732,116). (Abstract) Crawford, T., Solutions to the problems of teaching an introductory course in data processing. Direction. Spring - Summer 1978, 7(1), 11-15. Cronbach, L.J., Essentials of psychological testing. New York: Harper and Brothers, 1949. Bibliography Page 108 Curtis, W., A review of human factors research on programming languages and specifications. AEDS Monitor. Mar./Apr. 1983, 2i(9&10), 24-30. Dahl, O.-J., Dijkstra, E.W. and Hoare, C.A.R., Structured programming. London: Academic Press, 1972. De Blassio, J.K., and Bell, F.H., Attitudes toward computers in high school mathematics courses. International Journal of Mathematical Education in Science & Technology. 1981, 12, 47-56. Dey, S., and Mand, L.R., Effects of mathematics preparation and prior language exposure on perceived performance in introductory Computer Science courses. ACM  SIGCSE Bulletin. Feb. 1986, 18(1), 144-148. Dijkstra, E.W., Go To statement considered harmful. Communications of the ACM. Mar. 1968, i i , 147-148. (a) Dijkstra, E.W., The structure of the "THE" - Multiprogramming system. Communications of the ACM. May 1968, i i , 341-346. (b) Dijkstra, E.W., The humble programmer. Communications of the ACM. Oct. 1972, 15. 859-866. Dijkstra, E.W., How do we tell truths that might hurt? ACM SIGPLAN Notices. May 1982, i7(5), 13-15. di Persio, T., Isbister, D., and Shneiderman, B., An experiment using memorization/reconstruction as a measure of programmer ability. International  Journal of Man-Machine Studies. Oct. 1980, 13, 339-354. Dixon, V.A.. An investigation of prior sources of difficulties in learning university  computer science. Paper presented at the National Educational Computer Conference, Philadelphia, Pennsylvania, June 25, 1987. Ferguson, G.H., Statistical analysis in psychology and education (5th ed.). New York: McGraw-Hill, 1981. Fowler, G.C., and Glorfeld, L.W., Predicting aptitude in introductory computing: A classification model. AEDS Journal. Winter 1981,14(2), 96-109. Bibliography Page 109 Fudge, J.W., Predicting academic performance from biographical data (Doctoral dissertation, University of Texas at Austin, 1970). Dissertation Abstracts  International. 1971, 32 , 3785-A (University Microfilms No. 72-2336,110). (Abstract) Furugori, T., and Jalics, P., First course in Computer Science, A small survey. ACM  SIGCSE Bulletin. Feb. 1977,9(1), 119-122. Gathers, E., Screening freshmen Computer Science majors. ACM SIGCSE Bulletin. Sept. 1986,18(3), 44-48. Gibbs, N.E., An introductory Computer Science course for all majors. ACM SIGCSE  Bulletin. Aug. 1977,9(3), 34-38. Glorfeld, L.W., and Fowler, G.C., Validation of a model for predicting aptitude for introductory computing. ACM SIGCSE Bulletin. Feb 1982,14(1), 140-143. Grady, T.M., and Gawronski, J.D. (Eds), Computers in curriculum and instruction. Alexandria, Virginia: Association for Supervision and Curriculum Development, 1983. Gray, J.D., Predictability of success and achievement level of Data Processing Technology students at the two-year post-secondary level (Doctoral dissertation, Georgia State University, 1974). Dissertation Abstracts International. 1974, 35 , 2208-A (University Microfilms No. 74-23172,140). (Abstract) Greer, J., High school experience and university achievement in Computer Science. AEDS  Journal. Winter/Spring 1986, 1£(2&3), 216-225. Gries, D., What should we teach in an introductory programming course? ACM SIGCSE  Bulletin. Feb. 1974, 6(1), 81-89. Guinan, T., and Stephens, L., Factors affecting the achievement of high school students in beginning Computer Science courses. Journal of Computers in Mathematics  and Science Teaching. Fall 1988, 8(1), 61-64. Habermann, A.N., Critical comments on the programming language Pascal. Acta  Informatica. 1973, 3, 47-57. Bibliography Page 110 Hall, H.M., and Kidman, B.P., Empirical analysis of BASIC and Fortran programs. In O. Lecarme and R. Lewis (Eds.), Computers in Education (pp. 313-317). New York: North Holland, 1975. Hansen, T.P., Klassen, D.L., Anderson, R.E., and Johnson, D .C , What teachers think every high school graduate should know about computers. School Science and  Mathematics. Oct. 1981, SL 467-472. Horst, P., Differential prediction in college admissions. College Board Review. 1957, 33, 19-23. Hostetler, T.R., Predicting student success in an introductory programming course. ACM  SIGCSE Bulletin. Sept. 1983, 15(3), 40-43;49. Howerton, CP., The impact of pre-college computer exposure on student achievement in introductory computer programming courses. Computer Science Education. 1988,1(1), 73-84. Hunt, A.W., A decision-rule technique for predicting academic success. Decision  Sciences. 1977, 8, 270-286. Hunt, D., and Randhawa, B.S., Relationship between and among cognitive variables and achievement in Computational Science. Educational and Psychological  Measurement. 1973, 33, 921-928. Hyde, D.C, Gay, B.D. and Utter, D. Jr., The integration of a problem solving process in the first course. ACM SIGCSE Bulletin. Feb. 1979, 11(1), 54-59. Kernighan, B.W., and Plauger, P.J., Programming style: Examples and counterexamples. ACM Computing Surveys. Dec. 1974, 6, 303-319. Kimura, T., Reading Before composition. ACM SIGCSE Bulletin. Feb. 1979, U ( l ) , 162-166. Kirk, R.E., Experimental design: Procedures for the behavioral sciences. Belmont, California: Wadsworth, 1968. Klein, L., "GOTO" is not a four-letter word. Computers & Education. 1983, 7, 65-67. Bibliography Page 111 Knuth, D.E.. The art of computer programming (Vol.1). Reading, Massachusetts: Addison-Wesley, 1973. Konvalina, J., Stephens, L., and Wileman, S., Identifying factors influencing computer science aptitude and achievement. AEDS Journal. Winter 1983,16(2), 106-112. Konvalina, J., Wileman, S.A., and Stephens, L.J., Math proficiency: A key to success for Computer Science students. Communications of the ACM. May 1983,26,377-382. Kreitzberg, C.B., and Swanson, L., A cognitive model for structuring an introductory programming curriculum. National Computer Conference. 1974, pp. 307-311. Kurtz, B.L., Investigating the relationship between the development of abstract reasoning and performance in an introductory programming class. ACM SIGCSE  Bulletin. Feb. 1980, 12(1), 110-117. Lavin, D.E., The prediction of academic performance. New York: Russell Sage Foundation, 1965. Lecarme, O., What programming language should we use for teaching programming? In W.M. Turski (Ed.), Programming Teaching Techniques (pp. 61-67). New York: North Holland, 1973. Leeper, R.R., and Silver, J.L., Predicting success in a first programming course. ACM  SIGCSE Bulletin. Feb. 1982,14(1), 147-150. Lemos, R., Students' attitudes towards programming: The effects of structured walk-throughs. Computers & Education. 1978, 2, 301-306. Lemos, R.S., Teaching programming languages: A survey of approaches. ACM SIGCSE  Bulletin. Feb. 1979,11(1), 174-181. Lemos, R.S., A comparison of non-Business and Business student test scores in BASIC. AEDS Journal. Spring 1981, 14(3), 151-158. Lucas, H.C. Jr., and Kaplan, R.B., A structured programming experiment. The Computer  Journal. 1976,19, 136-138. Bibliography Page 112 Mazlack, L.J., Does a computer have sexual preferences? ACM SIGCSE Bulletin. Feb. 1976, 8(1), 74-78. Mazlack, L.J., Identifying potential to acquire programming skill. Communications of the ACM. Jan. 1980, 23, 14-17. McCann, T.E., Personal communication, Aug. 1, 1983. McGee, L., Polychronopoulos, G., and Wilson, C , The influence of BASIC on performance in introductory Computer Science courses using Pascal. ACM  SIGCSE Bulletin. Sept. 1987,19(3), 29-37. Molnar, A.R., The next great crisis in American education: Computer literacy. THE  Journal. July 1978, pp. 35-38. Mussio, J.J., and Wahlstrom, M.W., Predicting performance of programmer trainees in a post-high school setting. Proceedings of the Annual Computer Personnel  Research Conference. 1971, pp. 26-45. Newsted, P.R., Grade and ability predictions in an introductory programming course. ACM SIGCSE Bulletin. Jun. 1975, 7(2), 87-91. Nowaczyk, R.H., Cognitive skills needed in computer programming. Paper presented at Southeastern Psychological Association meeting, Atlanta, Georgia, Mar. 1983. Oman, P.W., Identifying student characteristics influencing success in introductory Computer Science courses. AEDS Journal. Winter/Spring 1986,19(2&3), 226-233. Papert, S., Mindstorms: Children, computers and powerful ideas. New York: Basic Books, 1980. Peck, J.E.L., Comparison of languages. In W.M. Turski (Ed.), Programming Teaching  Techniques (pp. 43-58). New York: North Holland, 1973. Bibliography Page 113 Petersen, C.G., The development and cross validation of a predictive model of achievement in Introduction to Computers at Northwest Missouri State University (Doctoral dissertation, Iowa State University, 1976). Dissertation  Abstracts International. 1977, 37 ,4168-A (University Microfilms No. 77-1033,106). (Abstract) Petersen, C.G., and Howe, T.G., Predicting academic success in Introduction to Computers. AEDS Journal. Fall 1979, 12(1), 182-191. Plog, C.E., The relationship of selected variables in predicting academic success in computer programming (Doctoral dissertation, East Texas State University, 1980). Dissertation Abstracts International. 1981, 4_L, 2903A- 2904A (University Microfilms No. 8027678,183). (Abstract) Poirot, J.L., A course description for teacher education in computer science. ACM  SIGCSE Bulletin. Feb. 1976, 8(1), 39-48. Poirot, J.L., Computer education in the secondary school: Problems and solutions. ACM  SIGCSE Bulletin. Feb. 1979,11(1), 101-104. Ralston, A., FORTRAN and the first course in Computer Science. ACM SIGCSE  Bulletin. Dec. 1971, 3(4), 24-29. Ralston, A., The first course in Computer Science needs a mathematics corequisite. Communications of the ACM. Oct. 1984, 27, 1002-1005. Ralston, A., and Shaw, M., Curriculum '78 - Is Computer Science really that unmathematical? Communications of the ACM. Feb. 1980, 23, 67-70. Ramberg, P., and Van Caster, S., A new look at an old problem: Keys to success for Computer Science students. ACM SIGCSE Bulletin. Sept. 1986,18(3), 36-39. Sail, M., A rationale for secondary Computer Science education. CUE Journal. Winter 1986, 6(2), 46-56. Sauter, V.L., Predicting computer programming skill. Computers & Education. 1986, 10, 299-302. Scheffe, H.A.. The analysis of variance. New York: J. Wiley and Sons, 1959. Bibliography Page 114 Schneider, G.M., The introductory programming course in Computer Science - Ten principles. ACM SIGCSE Bulletin. Feb. 1978, 10(1), 107-114. Schroeder, M.H., Piagetian, mathematical and spatial reasoning as predictors of success in computer programming (Doctoral dissertation, University of Northern Colorado, 1978). Dissertation Abstracts International. 1979, 39 , 4850-A(University Microfilms No. 7902855,100). (Abstract) Schulz, C.E., A survey of colleges and universities regarding entrance requirements in computer-related areas. Mathematics Teacher. Oct. 1984,77, 519-521. Self, C.C., A position on a computer literacy course. Paper presented at the University of Massachusetts-Amherst, Amherst, Massachusetts, May 20,1983. Shapiro, H.D., The results of an informal study to evaluate the effectiveness of teaching structured programming. ACM SIGCSE Bulletin. Dec. 1980,12(4), 50-56. Sharma, S., Learners' cognitive styles and psychological types as intervening variables influencing performance in computer science courses. Journal of Educational  Technology Systems. 1987,15, 391-399. Singhania, R.P., Issues in teaching the introductory course in Computers in Business curriculum. AEDS Journal. Fall 1980, 14(1), 45-51. Solntseff, N., Programming languages for introductory computing courses. ACM  SIGCSE Bulletin. Feb. 1978, 10(1), 119-124. Sorge, D.H., and Wark, L.K., Factors for success as a Computer Science major. AEDS  Journal. Summer 1984, 17(4), 36-45. Stephens, L.J., Wileman, S., and Konvalina, J., Group differences in computer science aptitude. AEDS Journal. Winter 1981, 14(2), 84-95. Stephens, L., Wileman, S., Konvalina, J., and Teodoro, E.V., Procedures for improving student placement in Computer Science. Journal of Computers in Mathematics  and Science Teaching. Spring 1985,4(3), 46-49. Sterling, T., and Pollack, S., Experience with a "universal" introductory course in Computer Science. ACM SIGCSE Bulletin. Nov. 1970, 2(3), 106-112. Bibliography Page 115 Stevens, D.J., Cognitive processes and success of students in instructional computer courses. AEDS Journal. Summer 1983, 16.(4), 228-233. Szymczuk, M., and Frerichs, D., Using standardized tests to predict achievement in an introductory high school computer course. AEDS Journal. Fall 1985,19(1). 20-27. Task Force on Curriculum for Secondary School Computer Science, Computer Science for secondary schools: Course content. Communications of the ACM. Mar, 1985, 28, 270-274. Tesler, L.G., Programming languages. Scientific American. Sept. 1984, 251(3). 70-78. Tillman, M.R., An examination of the predictive validity of several potential predictors of the work proficiency of computer programmers. Computer Personnel. Spring 1974, 5(1), 3-10. Ulloa, M., Teaching and learning computer programming: A survey of student problems, teaching methods, and automated instructional tools. ACM SIGCSE Bulletin. July 1980, 12(2), 48-64. van Dam, A., Strauss, C M . , McGowan, C , and Morse, J., A survey of introductory and advanced programming courses. ACM SIGCSE Bulletin. Feb. 1974, 6(1), 174-183. Van de Riet, R.P., Some criteria for elementary programming languages. In O. Lecarme and R. Lewis (Eds.). Computers in Education (pp. 953-963). New York: North Holland, 1975. Weinberg, G.M., The psychology of computer programming. New York: Van Nostrand Reinhold, 1971. Werth, L.H., Predicting student performance in a beginning Computer Science class. ACM  SIGCSE Bulletin. Feb. 1986,18(1), 138-143. Whipkey, K.L., and Stephens, J.T., Identifying predictors of programming skill. ACM  SIGCSE Bulletin. Dec. 1984, 16(4), 36-42. Bibliography Page 116 Wileman, S., Konvalina, J., and Stephens, L.J., Factors influencing success in beginning Computer Science courses. Journal of Educational Research. Mar./Apr. 1981, 74, 223-226. Wileman, S., Stephens, L., and Konvalina, J., The relationship between mathematical competencies and computer science aptitude and achievement. Journal of Computers in Mathematics and Science Teaching. Fall 1982,2(1), 20-21. Williams, M.H., The programming language BPL. The Computer Journal. Aug. 1982, 25. 289-301. Winer, B.J., Statistical principles in experimental design. New York: McGraw-Hill, 1971. Wirth, N., Program development by stepwise refinement. Communications of the ACM. Apr. 1971, 14,221-227. (a) Wirth, N., The programming language Pascal. Acta Informatica. 1971,1, 35-63. (b) Wirth, N., Systematic programming An introduction. Englewood Cliffs, N.J.: Prentice-Hall, 1973. Woodhouse, D., Introductory courses in computing: Aims and languages. Computers & Education. 1983, 7, 79-89. Yaney, J.P., Predicting programmer performance. Training and Development Journal. June 1970, 24(6), 21-23. Appendix A Page 117 A p p e n d i x A Letter to Instructors E R I C H A M B E R S E C O N D A R Y S C H O O L 1 9 8 5 S e p t . 12 D e a r S i r : I am w r i t i n g t o c o n f i r m t h e a r r a n g e m e n t s I made w i t h y o u i n r e g a r d s t o c o n d u c t i n g a q u e s t i o n n a i r e i n y o u r c o m p u t e r s c i e n c e c o u r s e t h i 3 m o n t h . I w i l l b e a t y o u r c l a s s o n , S e p t . a t i n a n d w o u l d a p p r e c i a t e f i v e m i n u t e s o f c l a s s t i m e f o r t h e d i s t r i b u t i o n a n d c o m p l e t i o n o f t h i s q u e s t i o n n a i r e . I w o u l d a p p r e c i a t e a c o p y o f y o u r c o u r s e o u t l i n e , a t y o u r c o n v e n i e n c e , f o r i n c l u s i o n i n a n a p p e n d i x o f t h e t h e s i s . A c o p y o f t h e r e s u l t s o f t h i s s t u d y w i l l b e made a v a i l a b l e t o y o u u p o n i t s c o m p l e t i o n . T h a n k y o u v e r y much f o r t h e a s s i s t a n c e y o u h a v e g i v e n t o t h i s p r o j e c t . I f y o u h a v e a n y q u e s t i o n s , p l e a s e d o n o t h e s i t a t e t o c a l l me. A c o p y o f t h e q u e s t i o n n a i r e s i m i l a r t o t h e f i n a l d r a f t i s e n c l o s e d . S i n c e r e l y , D a v i d N. E l l i s Appendix B Page 119 Appendix B Coding Scheme Page 120 Appendix B Coding Scheme Record 1 Record 2 Variables Columns Variables Columns Record Type ("blank") 1 Record Type ("*") 1 Student Number 2-9 Student Number 2-9 Course Count (0-25) 10-11 Course 10-12 Year Standing 12 Section 13-14 Year Percent (999.99) 13-17 Year 15 Status (l-9,blank) 18 Age 16-17 Status Date (YYMMDD) 19-24 Sex 18 Number of C.S. courses (1-6) 26 Faculty 19-20 C.S. Course Number (1) 28-30 Majors 21-22 Term (1 or 2) 32 Never learned language? 23 Mark Obtained ( /75) 34-36 Language (BASIC) 24 Course Standing 38 Order of learning 25 C.S. Course Number (2) 41-43 How well 26 Term 45 When? 27-29 Mark Obtained 47-49 Where? 30-33 Course Standing 51 Language (Pascal) 34 C.S. Course Number (3) 54-56 Order of learning 35 Term 58 How well 36 Mark Obtained 60-62 When? 37-39 Course Standing 64 Where? 40-43 C.S. Course Number (4) 67-69 Language (LOGO) 44 Term 71 Order of learning 45 Mark Obtained 73-75 How well 46 Course Standing 77 When? 47-49 C.S. Course Number (5) 80-82 Where? 50-53 Term 84 Language (Other) 54 Mark Obtained 86-88 Order of learning 55 Course Standing 90 How well 56 C.S. Course Number (6) 93-95 When? 57-59 Term 97 Where? 60-63 Mark Obtained 99-101 Course Standing 103 Record 2 Coding Information Page 121 Variable Year Sex Faculty Column 15 18 19-20 Major(s) 21-22 Value Meanina 0-9 0 for more than 9 years 0 Female 1 Male 01 Science 02 Arts 03 Commerce and Business Administration 04 Applied Science 05 Forestry 06 Education 07 Agriculture 08 Home Economics 09 Physical Education 10 Music 99 Unclassified 00 Unknown 01 Computer Science 02 Mathematics 03 Commerce 04 Engineering - all branches 05 Physics 06 Chemistry 08 Music 09 Law 10 Geology & Geophysics 11 Marine Biology 12 Comp. Science, Mathematics 13 Pharmacology, Pharmacy 14 Comp. Science, Chemistry 15 Comp. Science, Physics 16 Biochemistry 17 Comp. Science, Biology 18 Biology 19 Comp. Science, Psychology 21 Economics 22 International Relations 23 Geography 24 English 25 Mathematics, Physics 26 Asian Studies 27 Art 29 General Science 31 Marketing 32 Accounting 33 Finance 41 Computer Engineering Record 2 Page 122 Coding information  Variable Column Where did you learn? Other (specify) 33, 43, 53, 63 Value Meaning 50 Forestry 51 Harvesting 52 Resource Management 61 Recreation 62 Physical Education 63 Business Education 64 Mathematics, Science 65 Social Studies, Geography 66 Primary Education 67 Social Studies 68 Special Education 69 Oceanography 71 Agricultural Economics 72 Animal Science 73 Plant Science 74 Soil Science 75 Food Science 81 Dietetics 90 Languages, Linguistics 91 Psychology 92 Political Science 93 Sociology 94 Physiology 95 Microbiology 96 Zoology 97 Botany 98 Qualifying 99 General Usage 1 Summer course 6 2 Friends 2 3 Adult Cont Educ course 1 4 Self-taught 2 5 Practicum 1 6 Recreational 1 7 Uncle 2 8 Not specified 1 9 Curiousity 1 Other Languages 54 1 2 3 4 5 6 7 8 9 Assembler Fortran, Watfiv PL/1 Forth Machine CPL Comal COBOL DBASE 2 Other (APL, C, etc.) usage: 6, 3, many 1's Appendix C Page 123 Appendix C Statistical Tests Appendix C Page 124 Table 1 21 Aug 89 COMPUTING LANGUAGES - Preliminary Testing (Combined Data) 19:11:04 Un i v e r s i t y of B r i t i s h Columbia BORDER Order of Learning BASIC VALID CUM VALUE LABEL VALUE FREQUENCY PERCENT PERCENT PERCENT BASIC not learned 0 440 36.9 36.9 36.9 F i r s t language 1 698 58.5 58.5 95.3 Second language 2 40 3.4 3.4 98.7 Thi r d language 3 12 1.0 1.0 99.7 Fourth language 4 4 .3 .3 100.0 TOTAL 1194 100 .0 100.0 MEAN .695 STD DEV .616 MINIMUM .000 MAXIMUM 4.000 VALID CASES 1194 MISSING CASES 0 Appendix C Page 125 Table 2 21 Aug 89 COMPUTING LANGUAGES - Preliminary Testing (Combined Data) 19:20:08 U n i v e r s i t y of B r i t i s h Columbia BHOWELL How well BASIC was learned VALID CUM VALUE LABEL VALUE FREQUENCY PERCENT PERCENT PERCENT BASIC not learned 0 447 37.4 37.4 37.4 Vaguely f a m i l i a r 1 80 6.7 6.7 44.1 Capable of reading 2 71 5.9 5.9 50.1 Writes simple programs 3 202 16.9 16.9 67 .0 Quite f a m i l i a r 4 157 13.1 13.1 80.2 Writes complex programs 5 237 19.8 19.8 100.0 TOTAL 1194 100.0 100.0 MEAN 2.212 STD DEV 2.001 MINIMUM .000 MAXIMUM 5.000 VALID CASES 1194 MISSING CASES 0 Appendix C Page 126 Table 3 21 Aug 89 COMPUTING LANGUAGES - Preliminary Testing (Combined Data) 19:32:01 Un i v e r s i t y of B r i t i s h Columbia BWHEN When BASIC was learned VALID CUM VALUE LABEL VALUE FREQUENCY PERCENT PERCENT PERCENT BASIC not learned 0 441 36.9 36.9 36.9 Age 18 or more 1 103 8.6 8.6 45.6 Age 13 to 18 10 623 52 .2 52.2 97 .7 Age 13 to 18 & age > 18 11 2 .2 .2 97.9 Age up to 13 100 20 1.7 1.7 99.6 Age < 13 & age 13 to 18 110 5 .4 .4 100.0 TOTAL 1194 100.0 100.0 MEAN 7.458 STD DEV 14.670 MINIMUM .000 MAXIMUM 110.000 VALID CASES 1194 MISSING CASES 0 Appendix C Page 127 Table 4 21 Aug 89 COMPUTING LANGUAGES - Preliminary Testing (Combined Data) 19:41:56 Un i v e r s i t y of B r i t i s h Columbia BWHERE Where BASIC was learned VALID CUM VALUE LABEL VALUE FREQUENCY PERCENT PERCENT PERCENT BASIC not learned 0 448 37.5 37 .5 37.5 Summer Course 1 4 .3 .3 37.9 Friends 2 1 .1 .1 37.9 Adult Cont. Education 3 1 .1 .1 38.0 Self-taught 4 1 .1 . 1 38.1 Recreational 6 1 .1 .1 38.2 Uncle 7 1 .1 .1 38.3 C u r i o u s i t y 9 1 .1 .1 38.4 Job 10 3 .3 .3 38.6 Home 100 103 8.6 8.6 47 .2 Home & Friends 102 1 .1 .1 47 .3 Home & Job 110 1 . 1 .1 47 . 4 School, U n i v e r s i t y 1000 528 44.2 44 .2 91.6 School & Job 1010 4 .3 .3 92.0 School & Home 1100 91 7.6 7.6 99.6 School, Home & Other 1101 1 .1 .1 99.7 School, Home & Job 1110 4 .3 .3 100.0 TOTAL 1194 100.0 100.0 MEAN 542.930 STD DEV 499.308 MINIMUM 000 MAXIMUM 1110.000 VALID CASES 1194 MISSING CASES 0 Appendix C Page 128 Table 5 21 Aug 89 COMPUTING LANGUAGES - Preliminary Testing (Combined Data) 19:46:18 Un i v e r s i t y of B r i t i s h Columbia PORDER Order of Learning PASCAL VALID CUM VALUE LABEL VALUE FREQUENCY PERCENT PERCENT PERCENT Pascal not learned 0 858 71.9 71.9 71.9 F i r s t language 1 54 4.5 4.5 76.4 Second language 2 178 14.9 14.9 91.3 Th i r d language 3 85 7.1 7.1 98.4 Fourth language 4 19 1.6 1.6 100.0 TOTAL 1194 100.0 100.0 MEAN .621 STD DEV 1.074 MINIMUM .000 MAXIMUM 4.000 VALID CASES 1194 MISSING CASES 0 Appendix C Page 129 Table 6 21 Aug 8 9 COMPUTING LANGUAGES - Preliminary Testing (Combined Data) 19:47:40 Un i v e r s i t y of B r i t i s h Columbia PHOWELL How well PASCAL was learned VALID CUM VALUE LABEL VALUE FREQUENCY PERCENT PERCENT PERCENT Pascal not learned 0 872 73.0 73.0 73.0 Vaguely f a m i l i a r 1 59 4.9 4.9 78.0 Capable of reading 2 34 2.8 2.8 80.8 Writes simple programs 3 72 6.0 6.0 86.9 Quite f a m i l i a r 4 54 4.5 4.5 91.4 Writes complex programs 5 103 8.6 8.6 100.0 TOTAL 1194 100.0 100.0 MEAN .899 STD DEV 1. 667 MINIMUM .000 MAXIMUM 5.000 VALID CASES . 1194 MISSING CASES 0 Appendix C Page 130 Table 7 21 Aug 89 COMPUTING LANGUAGES - Preliminary Testing (Combined Data) 19:48:57 Un i v e r s i t y of B r i t i s h Columbia PWHEN When PASCAL was learned VALID CUM VALUE LABEL VALUE FREQUENCY PERCENT PERCENT PERCENT Pascal not learned 0 858 71.9 71. 9 71.9 Age 18 or more 1 188 15.7 15.7 87.6 Age 13 to 18 10 146 12.2 12.2 99.8 Age 13 to 18 & age > 18 11 2 .2 .2 100.0 TOTAL 1194 100.0 100.0 MEAN 1.399 STD DEV 3.262 MINIMUM .000 MAXIMUM 11.000 VALID CASES 1194 MISSING CASES 0 Appendix C Page 131 Table 8 21 Aug 89 COMPUTING LANGUAGES - Preliminary Testing (Combined Data) 19:50:37 U n i v e r s i t y of B r i t i s h Columbia PWHERE Where PASCAL was learned VALID CUM VALUE LABEL VALUE FREQUENCY PERCENT PERCENT PERCENT Pascal not learned 0 861 72.1 72.1 72.1 Not s p e c i f i e d 8 1 .1 .1 72.2 Job 10 2 .2 .2 72.4 Home 100 21 1.8 1.8 74.1 School, U n i v e r s i t y 1000 286 24.0 24.0 98.1 School & Other 1001 1 .1 .1 98.2 School & Job 1010 1 .1 .1 98.2 School & Home 1100 21 1.8 1.8 100.0 TOTAL 1194 100.0 100 .0 MEAN 262.344 STD DEV 440.477 MINIMUM .000 MAXIMUM 1100.000 VALID CASES 1194 MISSING CASES 0 Appendix C Page 132 Table 9 21 Aug 89 COMPUTING LANGUAGES - Preliminary Testing (Combined Data) 19:52:10 Un i v e r s i t y of B r i t i s h Columbia LORDER Order of Learning LOGO VALID CUM VALUE LABEL VALUE FREQUENCY PERCENT PERCENT PERCENT LOGO not learned 0 1123 94.1 94 .1 94.1 F i r s t language 1 9 .8 .8 94.8 Second language 2 31 2.6 2.6 97 .4 Thi r d language 3 19 1.6 1.6 99.0 Fourth language 4 12 1.0 1.0 100.0 TOTAL 1194 100.0 100.0 MEAN .147 STD DEV . 628 MINIMUM .000 MAXIMUM 4.000 VALID CASES 1194 MISSING CASES 0 Appendix C Page 133 Table 10 21 Aug 8 9 COMPUTING LANGUAGES - Preliminary Testing (Combined Data) 19:54:24 U n i v e r s i t y of B r i t i s h Columbia LHOWELL How well LOGO was learned VALID CUM VALUE LABEL VALUE FREQUENCY PERCENT PERCENT PERCENT LOGO not learned 0 1123 94.1 94. .1 94.1 Vaguely f a m i l i a r 1 23 1.9 1. .9 96.0 Capable of reading 2 14 1.2 1. ,2 97.2 Writes simple programs 3 26 2.2 2. .2 99.3 Quite f a m i l i a r 4 6 .5 .5 99.8 Writes complex programs 5 2 .2 .2 100.0 TOTAL 1194 100.0 100, .0 MEAN . 137 STD DEV .605 MINIMUM .000 MAXIMUM 5.000 VALID CASES 1194 MISSING CASES 0 Appendix C Page 134 Table 11 21 Aug 89 COMPUTING LANGUAGES - Preliminary Testing (Combined Data) 19:55:53 U n i v e r s i t y of B r i t i s h Columbia LWHEN When LOGO was learned VALUE LABEL LOGO not learned Age 18 or more Age 13 to 18 MEAN MAXIMUM .518 10.000 VALID CUM VALUE FREQUENCY PERCENT PERCENT PERCENT 0 1 10 TOTAL STD DEV 1124 9 61 1194 2 .203 94.1 .8 5.1 100.0 94.1 .8 5.1 100.0 MINIMUM 94.1 94.9 100.0 .000 VALID CASES 1194 MISSING CASES 0 Appendix C Page 135 Table 12 21 Aug 89 COMPUTING LANGUAGES - Preliminary Testing (Combined Data) 19:58:24 U n i v e r s i t y of B r i t i s h Columbia LWHERE Where LOGO was learned VALID CUM VALUE LABEL VALUE FREQUENCY PERCENT PERCENT PERCENT LOGO not learned 0 1127 94.4 94.4 94.4 Summer Course 1 2 .2 .2 94.6 Practicum 5 1 .1 .1 94.6 Job 10 3 .3 .3 94.9 Home 100 18 1.5 1.5 96.4 Home & Other 101 1 .1 .1 96.5 School, U n i v e r s i t y 1000 39 3.3 3.3 99.7 School & Home 1100 3 .3 .3 100 .0 TOTAL 1194 100.0 100.0 MEAN 37.050 STD DEV 185.794 MINIMUM .000 MAXIMUM 1100.000 VALID CASES 1194 MISSING CASES 0 \ Appendix C Page 136 Table 13 21 Aug 89 COMPUTING LANGUAGES - Preliminary Testing (Combined Data) 19:59:49 U n i v e r s i t y of B r i t i s h Columbia OORDER Order of Learning Other language VALUE LABEL VALID CUM VALUE FREQUENCY PERCENT PERCENT PERCENT Other not learned 0 896 75.0 75 .0 75.0 F i r s t language 1 63 5.3 5 .3 80.3 Second language 2 141 11.8 11 .8 92.1 Th i r d language 3 74 6.2 6 .2 98.3 Fourth language 4 20 1.7 1 .7 100.0 TOTAL 1194 100.0 100 .0 MEAN .542 STD DEV 1.029 MINIMUM .000 MAXIMUM 4.000 VALID CASES 1194 MISSING CASES Appendix C Page 137 Table 14 21 Aug 89 COMPUTING LANGUAGES - Preliminary Testing (Combined Data) 20:01:31 U n i v e r s i t y of B r i t i s h Columbia OHOWELL How well Other language was learned VALID CUM VALUE LABEL VALUE FREQUENCY PERCENT PERCENT PERCENT Other not learned 0 908 76.0 76.0 76.0 Vaguely f a m i l i a r 1 42 3.5 3.5 79.6 Capable of reading 2 22 1.8 1.8 81.4 Writes simple programs 3 70 5.9 5.9 87.3 Quite f a m i l i a r 4 98 8.2 8.2 95.5 Writes complex programs 5 54 4.5 4.5 100.0 TOTAL 1194 100.0 100.0 MEAN .802 STD DEV 1.562 MINIMUM .000 MAXIMUM 5.000 VALID CASES 1194 MISSING CASES 0 Appendix C Page 138 Table 15 21 Aug 89 COMPUTING LANGUAGES - Preliminary Testing (Combined Data) 20:02:57 Un i v e r s i t y of B r i t i s h Columbia OWHEN When Other language was learned VALID CUM VALUE LABEL VALUE FREQUENCY PERCENT PERCENT PERCENT Other not learned 0 897 75.1 75 .1 75.1 Age 18 or more 1 179 15.0 15 .0 90.1 Age 13 to 18 10 116 9.7 9 .7 99.8 Age 13 to 18 & age > 18 11 1 . 1 .1 99.9 Age up to 13 100 1 .1 .1 100.0 TOTAL 1194 100.0 100 .0 MEAN 1.214 STD DEV 4.109 MINIMUM .000 MAXIMUM 100.000 VALID CASES 1194 MISSING CASES 0 Appendix C Page 139 Table 16 21 Aug 89 COMPUTING LANGUAGES - Preliminary Testing (Combined Data) 20:04:08 Un i v e r s i t y of B r i t i s h Columbia OWHERE Where Other language was learned VALUE LABEL VALID CUM VALUE FREQUENCY PERCENT PERCENT PERCENT Other not learned 0 899 75 .3 75 .3 75.3 Summer Course 1 1 .1 .1 75.4 Job 10 8 .7 .7 76.0 Home 100 52 4 .4 4 .4 80.4 Home & Job 110 1 .1 .1 80.5 School, U n i v e r s i t y 1000 219 18 .3 18 .3 98.8 School & Job 1010 2 .2 .2 99.0 School & Home 1100 10 .8 .8 99.8 School, Home & Job 1110 2 .2 .2 100.0 TOTAL 1194 100 .0 100 .0 MEAN 200.696 STD DEV 397.018 MINIMUM .000 MAXIMUM 1110.000 VALID CASES 1194 MISSING CASES Appendix C Page 140 Table 17 21 Sep 89 COMPUTING LANGUAGES - Pr e l i m i n a r y T e s t i n g (Combined Data) 17:04:36 U n i v e r s i t y of B r i t i s h Columbia * * * » M U L T I P L E R E G R E S S I O N L i s t w i s e D e l e t i o n of M i s s i n g Data Mean Std Dev Label MARK 68 .953 14 .769 Mark i n Surveyed Course NOLANG 1 .233 1 .111 Number of languages p r e v i o u s l y learned NOCSCOUR 1 .362 .686 Number of Computer Courses Taken BORDER .696 .617 Order o f Learning BASIC BASIC .636 .481 Was BASIC learned? AGE 20 . 177 3 .404 Age of student i n years SEX .723 .448 Gender of student YEAR 1 .833 1 .073 U n i v e r s i t y year e n r o l l e d i n PERCENT 64 .845 12 .588 Year Percent BHOWELL 2 .219 2 .002 How well BASIC was learned NVRLRND .301 .459 Never Learned Language? N o f Cases - 1178 C o r r e l a t i o n : MARK NOLANG NOCSCOUR BORDER BASIC AGE SEX YEAR PERCENT BHOWELL NVRLRND MARK 1.000 .200 .010 . 152 . 185 -.043 .032 -.032 .531 .229 -.195 NOLANG .200 1.000 .401 .645 .691 -.106 .206 -.052 -.022 .717 -.728 NOCSCOUR .010 .401 1.000 .261 .204 .020 .122 .073 -.064 .235 -.214 BORDER .152 .645 .261 1.000 .856 - .147 .143 -.096 -.053 .701 -.741 BASIC .185 .691 .204 . 856 1.000 -.244 .127 -.194 -.058 .839 -.866 AGE -.043 -.106 .020 -.147 -.244 1 .000 .029 .479 .031 -.280 . 129 SEX .032 .206 .122 .143 .127 .029 1.000 -.058 -.118 .192 -.166 YEAR -.032 -.052 .073 -.096 -.194 .479 -.058 1.000 .078 -.206 .104 PERCENT .531 -.022 -.064 -.053 -.058 .031 -.118 .078 1.000 -.034 . 044 BHOWELL .229 .717 .235 .701 .839 -.280 .192 -.206 -.034 1.000 -.727 NVRLRND -.195 -.728 -.214 -.741 -.866 .129 -.166 . 104 .044 -.727 1.000 Equation Number 1 Dependent V a r i a b l e . * M U L T I P L E R E G R E S S MARK Mark i n Surveyed Course Beginning Block Number 1. Method: Stepwise NOLANG NOCSCOUR BORDER BASIC AGE SEX BHOWELL PERCENT NVRLRND Va r i a b l e ( a ) Entered on Step Number 1. M u l t i p l e R .53099 R Square .28195 Adjusted R Square ,28134 Regression Standard E r r o r 12.52040 Residual PERCENT Year Percent A n a l y s i s o f Variance 1 1176 Sum of Squares 72386.04905 184350.24071 Mean Square 72386.04905 156.76041 S i g n i f F -V a r i a b l e s i n the Equation V a r i a b l e s not i n the Equation V a r i a b l e B SE B Beta T Sig T V a r i a b l e Beta In P a r t i a l Min T o l e r T S i g T PERCENT .622987 .028991 .530987 21. .489 .0000 NOLANG .211303 .249302 .999531 8. .824 .0000 (Constant) 28.555214 1.915026 14. .911 .0000 NOCSCOUR .044082 .051916 .995932 1. .782 .0750 BORDER .180484 .212695 .997230 7. .462 .0000 BASIC .216732 .255338 .996644 9. .053 .0000 AGE -.059078 -.069686 .999067 -2. .395 .0168 SEX .096322 .112874 .986032 3. .894 .0001 YEAR -.073618 -.086612 .993899 -2. .980 .0029 BHOWELL .247641 .292076 .998853 10. .468 .0000 NVRLRND -.219176 -.258396 .998027 -9. . 169 .0000 Equation Number 1 Dependent V a r i a b l e V a r i a b l e ( s ) Entered on Step Number 2. « M U L T I P L E R E G R E S S I O N MARK Mark i n Surveyed Course M u l t i p l e R R Square Adjusted R Square Standard E r r o r .58584 .34320 .34208 11.97954 BHOWELL How well BASIC was learn e d A n a l y s i s o f Variance Regression Residual DF 2 1175 Sum of Squares 88112.63255 168623.65720 Mean Square 44056.31628 143.50950 S i g n i f F V a r i a b l e s i n the V a r i a b l e s not i n the Equation V a r i a b l e B SE B Beta T Sig T V a r i a b l e Beta In P a r t i a l Min T o l e r T Sig T PERCENT .632829 .027755 .539375 22. 801 .0000 NOLANG .069537 .059788 .485206 2, .052 .0404 BHOWELL 1.827271 . 174552 .247641 10. 468 .0000 NOCSCOUR -.014539 -.017410 .941777 -, ,597 .5509 (Constant) 23.862283 1. . 886345 12. 650 .0000 BORDER .013565 .011931 .508125 ,409 .6827 BASIC .029284 .019612 .294575 .672 .5017 AGE .011043 .013076 .920712 .448 .6542 SEX .050946 .061293 .950667 2, .104 .0356 YEAR -.023916 - .028800 .952447 -, ,987 .3237 NVRLRND -.082973 - .070274 .471145 -2. .414 .0159 . . . . . . * * . . . . . . . . . . . . . « . . . . Appendix D Page 141 Appendix D Personal Communications Educational Technology Columnists^ Computers and -Learnin Don't Teach BASIC or Alfred Bork .This column proposes that it is a mistake to teach BASIC. At least I claim thac it is a mistake as a first language, and perhaps it is a mistake more generally. (Actually, I know of few examples where BASIC is taught as a second language, except for self-teaching.) This anti-BASIC position will un-doubtedly be viewed as heresy by many readers, as BASIC is a very commonly taught first language at the present time. In the first section of Alfred: Bork, Professor of Physics and Pro-fessor of Information and Computer Sci-ence, is Director of the Educational Tech-nology Center at the University of Califor-nia at Irvine. The Center is working in the area of personal computers, and pro-jects are proceeding at this time in public libraries and middle schools. Dr. 8ork was one of four American speakers to the most recent World Conference on Computers in Education. He was a co-director and key-note speaker at a recent N A T O Advanced Study Institute and a consultant to the UK National Development Programme in Com-puter Aided Learning. He has served as chair of the Special Interest Group on Computer Uses in Education of .the Association for Computing Machinery and as Physics Series Editor for CONDUIT. Dr. Bork has pub-lished over 30 papers during the past three years and has written chapters in five books. He edited Computer Assisted Learning in Physics Education (Pergamon Press, 1980) and is the author of Learning with Comput-ers (Digital Press, 1981). this column, I will comment on why I believe this is an important step to take at the present time. In the second section, I will comment on some possi-ble objections to this position. Why Shouldn't We Teach BASIC? When languages such as BASIC and F O R T R A N were developed, there was little experience with programming, particularly for large, complex activi-, ties. However, this is no longer the case. Today we have much experience with writing elaborate programs. Our experience has led to a series of strategies which make it easier to write such programs with fewer errors, and easier to revise these programs. While not everyone is going to write complex programs, many people in the future will be involved in this activity. In-deed, all projections indicate that we will not have enough programmers, given our current ways of producing programmers, to meet this need. The set of ideas that has evolved for good programming practice sometimes goes under the name of "structured programming." A variety of factors are involved, which I will not attempt to review. Components of software engi-neering also play a very important role. BASIC,, because it does not lead easily to structured programming, tends to develop poor programming habits, particularly as it is almost universally taught. These programming habits are very difficult for students to overcome later. We have not done students a service if we teach them fundamental ideas and ingrained ways of working which they later must destroy, particularly if these are the first ideas that they encounter in the area. It is difficult to correct such early habits. So BASIC hurts students in the long run, in spite of any short-run- advantages (and, as will be seen, I claim that there are not really any short-range advantages either!). We' have seen many students at Irvine who have learned BASIC in high school or on their own who have considerable difficulty in the begin-ning programming courses. Our situa-tion is not unique. For example, I have heard of very similar cases at the Air Force Academy in Colorado Springs. It is not easy to overcome some ingrained habits, particularly if they have been ingrained for a long time. Another factor to be considered is the lack of standardization in existing BASICs. Although there is a BASIC, standard underway, I do not see any widespread move of existing BASICs, or even newly developed BASICs, to conform to that standard. In fact, it seems extremely unusual that a stand-ard for an existing language should depart so much from the current, implementations of the language. Finally, and perhaps the most criti-cal point, there are better languages for the student to begin with. The languages in the A L G O L family, such as Pascal, do allow a natural approach to structured programming. They too are not always taught in this fashion, unfortunately, because often the text-books are written by people with the same bad habits I referred to above! But the percentage is certainly much better; that is, more people learn to program in a satisfactory fashion with Pascal, for example, than with a lan-guage such as BASIC. The coming likely importance of Ada is also an important factor to consider. If Ada becomes as widely EDUCATIONAL TECHNOLOGY/April, 1982 V o l u m e X X I I , Number 4 33 used as seems likely, given the strong interest of the Department of Defense, it is likely to require a whole new generation of programmers to work on it. It is very hard to imagine that students brought up on BASIC are likely to become good Ada pro-grammers. Comments on Possible Objections There are a number of comments which I feel are likely to be made about the position 1 .have just taken. As I indicated, it certainly is not the common position; these views are like-ly to bring down the wrath of many BASIC users. I will consider some of the criticisms that may appear. The first argument that is often made for BASIC is that it is common, particularly on small microcomputers. This is true. Most microcomputers have some version or other of BASIC But these versions -differ widely. As already indicated, BASIC is one of the least standardized languages around, with widely varying dialects. This is particularly true when one gets be-yond the most elementary level.. But I do not feet that this is a good argument. If we always stuck with what was common in a particular time, a new and better idea would never be used. That is, we would never make any progress. Certainly, at one time F O R T R A N was extremely common, when BASIC was first being intro-duced. The same argument could have been used at one time by F O R T R A N users to persuade people not to study BASIC. Other languages, including Pas-cal, are increasingly available. The second reason one often hears for teaching 3ASIC as a first language is the belief that BASIC is easy to teach. I believe that this is an old wive's tale, a position not supported by empirical evidence. As a teacher who has taught many different pro-gramming languages, it seems to me that for almost any language a reason-able subset is relatively easy to teach to beginners. The ease depends not so much on the language but on two quite different factors. One of these is the implementation. There are clearly some advantages in teaching an imple-mentation which is based on an inter-preter, rather than an implementation that is based on a compiler. Whether one has interpreters or compilers has little to do with the nature of the language in most cases, but only the nature of the implementation. The second important factor in how rapidly beginners become ac-quainted with the language is the method in which the language is taught. Users can become familiar with the language in widely different amounts of time, depending on just how the language is introduced. Many methods of teaching languages, partic-... one often hears... that BASIC is easy to teach. I believe this is an old wive's tale, a position not supported by empir-ical evidence." ularly those based on grammatical approaches, waste large amounts of student time, independent of the lan-guage being used. But this is a separate topic that cannot be adequately ad-dressed within the present column. Perhaps the most important objec-tion is the one I have left for last. It is often argued, correctly, that one can teach BASIC in a structured fashion. This can certainly be done to some extent. But the truth is that it is almost never done! One can find very few examples of teaching BASIC in this way. Certainly, in the commonly available books which students usually use in learning BASIC, one sees almost no teaching of BASIC in a structured fashion. There are some exceptions to this, but very few. One cannot even introduce BASIC in a structured fashion with many BASICs; that is, one can achieve some aspects of structure with these BA-SICS, but not others. Many BASICs, for example, have no adequate proce-dures. The notion of teaching top-down programming without a power-ful procedure concept is, it seems to me, ridiculous. An expression that says "GOSUB 982" is no substitution for an adequate procedure. A few BASICs do have adequate procedure me-chanisms, such as some of the Digital 8ASlCs. But many do not. Another problem with many BA-SICS, which prevents them from being taught in a structured fashion, is the use of single letter or single letter plus number identifiers. Thus, one cannot give meaningful names to identifiers, and one teaches students to call identi-fiers by such names as " X 4 2 , " which have no meaning and which lead to great difficulty when one tries to revise the program at a later time. Final Conclusions I would certainly welcome any comments about these arguments. Per-haps I have missed some critical points. But it is important, particularly given the growing need for competent computer programmers in our society, to give some very careful thought to these issues. We are about to be flooded with new high school courses involving BASIC There are some alter-nates to this. There is at least one quite good high school course, "Com-puter Power," based on a variant of Pascal, available from McGraw-Hill. So perhaps there are some rays of hoDe in what appears to be a generally dark night. • Forthcoming Columns The following columns are I some of chose now in prepara- 11 tion for publication in this ' magazine: • Alfred Bork: Ronald !' Reagan's Big Mistake. * • Leslie ). Briggs A Com- ' ment on the Training of Instructional Designers. ! • James E. Eisele: Pro-gramming or Author-ing? • Albert l _ Goldberg: ',< The Eclectic Technolo- '' gist. ,< • M. David Merrill: Au- |] thoring Systems: Are ',< They Really? ; • Dean R. Spitzer: Facili-tating Training Results Back on the Job. • Bruce W. Tuckman: "This Is a Recorded ,• Message . . . " r Each Educational Technology * Columnist appears in these ' pages several times yearly. < More than a dozen columnists J are writing columns for you. * These columns cover the full J range of problems, issues, and concerns in the field of educa- j tional technology. \ 34 EDUCATIONAL TECHNOLOGY/April, 1982 THE FOURTH REVOLUTION - COMPUTERS AND LEARNING Alfred Boric "Of a l l human inventions since the beginning of mankind, the microprocessor is unique. It i s destined to play a part i n a l l areas of l i f e , without exception—to increase our capacities, to f a c i l i t a t e or eliminate tasks, to replace physical e f f o r t , to increase the p o s s i b i l i t i e s and areas of mental e f f o r t , to turn every human being into a creator, whose every idea can be applied, dissected, put together again, transmitted, changed. The theme of this paper i s that we are on the verge of a major change in the way people learn. This change, driven by the personal computer, w i l l affect a l l levels of education from e a r l i e s t childhood through adult education. I t w i l l a f f e c t most subject areas and most learners. I t w i l l a f f e c t both education and training. It w i l l be one of the few major h i s t o r i c a l changes i n the way people learn. The impact of the computer i n education w i l l not produce an incremental change, a minor aberration on the current ways of learning, but w i l l lead to e n t i r e l y d i f f e r e n t learning systems. This massive change i n education w i l l occur over the next twenty years. Schools, i f they exist at a l l , w i l l be very d i f f e r e n t at the end of that period. There w i l l be fewer teachers, and the role of the teacher w i l l be d i f f e r e n t from the role of teacher in our current educational delivery system. I use •schools" throughout this paper in the general sense to include any formal schooling a c t i v i t y , whether i t be the t h i r d grade or the university, or any other l e v e l ; for emphasis I sometimes mention particular types of schools. I hasten to say that this change w i l l not necessarily be a desirable change. Any powerful technology carries within i n i t the seeds of good and e v i l , and that applies to an educational technology. One of my major goals in making presentations of this kind i s to nudge us toward a more desirable educational future rather than a less desirable educational future. Our e f f o r t s in the next few years are par t i c u l a r l y c r i t i c a l for education. The f u l l , long-range implications of the computer i n our world of learning are seldom discussed. Indeed, people are often overwhelmed by the technology, delighted with each new toy which they receive. Yet these implications must be considered i f we are to move toward an improvement i n our entire educational system. The strategy of this paper w i l l be to f i r s t look at the ^why," then to look at the "how," and then to return to present action. Many of the issues are discussed i n more d e t a i l i n my recent book, Learning with Computers.^ 1 The U n i v e r s i t y o f Oreeon - June 1982 — — — — — WHY WILL THE COMPUTER 3EC0ME THE DOMINANT EDUCATIONAL DELIVERY SYSTEM? In making a brief case as to why the change I am suggesting w i l l take place, I f i r s t look b r i e f l y at educational factors in modern society. Then I w i l l consider aspects d i r e c t l y related to the computer. Current Status of Education F i r s t , tt dooc not take any great pffnrr to SPP that nnr. educafcicmal'"By3le.u1 ia'-curgently in tf o*rb£»r We are being to l d t h i s constantly from a l l sides. The d a i l y newspapers, the popular magazines, and recent books are f u l l of descriptions of the problems of our current educational systems. One can even measure these to some extent by declining SAT scores, declining s t e a d i l y u n t i l l a s t year, and similar results from the National Assessment tests. Independent of s t a t i s t i c s , however, the most interesting and c r i t i c a l information is the decline in f a i t h in education i n the United States. We can see this very heavily reflected among p o l i t i c i a n s at a l l levels. At one time for a p o l i t i c i a n to speak out against education was s u i c i d a l . Now we find that i t is often p o l i t i c a l l y effective. Indeed, our current president campaigned on the notion that we didn't need a Department of Education, although so far he hasn't abolished i t . But he did abolish the entire science education d i v i s i o n within the National Science Foundation, simply by cutting i t s budget ef f e c t i v e l y to zero. The p o l i t i c i a n s know that education has l i t t l e support in American society and that, indeed, i t i s p o l i t i c a l l y expedient to cut educational funds. Education has few defenders and many detractors. I do not wish to imply that these problems with education are simply a matter of public relations. Indeed education has very r e a l problems in this country and elsewhere. In the whole his t o r y of the American educational system there has seldom been a time when there was greater turmoil and where the status of teaching, in both the public schools and u n i v e r s i t i e s , has been lower than i t i s now. A l l indications point to the fact that this decline i n popular support of our educational system w i l l continue. Few positive factors other than interest i n the computer can be pointed to. Coupled with this declining appreciation of education, perhaps even a consequence, i s a factor which affects education even more di r e c t l y , the factor of increasing financial constraints. The schools do not raise enough money to run an adequate educational system i n this country today. Any adequate science or mathematics teacher can make far more money outside of the schools and universities than that individual can make within 153 the schools. A few teachers w i l l be dedicated enough to stay with the schools or to go to schools in spite of thi s . But many competent people w i l l not, and many people who are not competent to do anything else w i l l teach. These are harsh statements, ones that are not pleasant to hear, but I think they must be made. Financial constraints also show up in other important ways in education beside teacher s a l a r i e s . We have had no new major curriculum development at any lev e l in the united States for over ten years. I am referring to sizable curriculum development projects, the type which could lead to improvement i n our educational system. Indeed, since the development of the MACOS course in the early 1970*s, federal funding i n curriculum development stopped almost e n t i r e l y . I r o n i c a l l y , we were just becoming s k i l l f u l in such development when the funds vanished. What we learned i s now being used in large-scale curriculum development in other countries. Another dismal factor in American education i s the current classroom environment. Even young children frequently show l i t t l e interest in education, r e f l e c t i n g widespread parental attitudes. High school classes often seem more l i k e battle f i e l d s than educational i n s t i t u t i o n s . This i s in stark contrast to what one finds in many other countries at the present time. Hence, American education, and to a lesser extent education everywhere, i s in trouble at the moment. ^F^^^ft^S^SSSfm, aiiJ ijevr^rary^^f"'d(ytng^tyf¥iig^. Much of the pressure on education i s from the outside, and this i s the type of pressure which can lead to real change. "The teaching profession is caught in a vicious cycle, spiraling downward. Rewards are few, morale is low, the best teachers are b a i l i n g out and the supply of good recruits i s drying up." 3 When we move from this dismal picture of what is happening in education today to look at the computer situation, the picture i s enti r e l y different. The computer, the dominant technology of our age and s t i l l rapidly developing, shows great promise as a learning mode. It has been said that the computer i s a g i f t of f i r e . P i r s t , a few hardware comments. Personal computers w i l l be dominant in education. But i t is a mistake to believe that computers currently around are the ones I am talking about. We are only at the beginning stage of computer development, pa r t i c u l a r l y with regard to the personal computer. Today's Apples and even today's IBM Personal Computers, a good b i t more sophisticated than the Apple, are hardly a shadow of the types of machines that w i l l dominate learning. Central processing units are becoming cheaper and more sophisticated, and memory of a l l 3 types is rapidly dropping in p r i c e . The integrated c i r c u i t technology is only at i t s beginning, and we can expect a long steady decline in prices, increase in c a p a b i l i t i e s , and decrease in s i z e . Going along with this w i l l be increased educational c a p a b i l i t i e s , such as sound (both i n and out), much better graphics, alternate media, such as those provided by the videodisc, and a host of other rapid developments. In planning for computers i n education we must give f u l l attention to this dynamic situation rather than focusing on today's hardware. Technology is not learning. We can be too carried away with the technology and become interested i n i t to the exclusion of learning! So we should not give primary attention i n education to the new hardware developments. The r e a l interest in the computer in learning l i e s not i n i t s decreasing price and increasing c a p a b i l i t i e s , obvious to a l l , but rather to i t s effectiveness as a learning device. How does one demonstrate this effectiveness? In education the t r a d i t i o n a l mode of experiment has seldom proved to be satisfactory. Neither the f i n a n c i a l resources nor the number of subjects are adequate in most existing educational research. The d i f f i c u l t i e s have to do with the many variables which cannot be controlled, so different from the experimental situations that were typical of the physical sciences. Few large-scale experiments have proceeded with the computer, and these were often flawed. Further, our s k i l l s i n developing materials have advanced, and many of the studies are based on minimal early material. We can find l i s t s of research projects that supposedly do or don't demonstrate that the computer i s good in learning, but I am singularly unimpressed with most of these studies when I examine them closely. So the use of adequate comparison studies in demonstrating that computers are useful in education i s seldom p r a c t i c a l . A l l i s not l o s t , however, in demonstrating effectiveness for users. One important way to do t h i s , very convincing i n many situations, i s to look at some examples of what i s possible and to point out the features of those examples which lead to the computer becoming generally very effective i n learning. I t is this approach we w i l l follow here. Another approach i s through peer evaluation, the examination of materials by pedagogical experts in the area involved. Educational Technology Center Projects I w i l l describe i n this section three projects i n computer based learning from the Educational Technology Center. The f i r s t used a timesharing system; the others, more recent, were developed d i r e c t l y on personal compters. The f i r s t project i s a beginning quarter of a college based 4 physics course for science-engineering majors. The key computer materials are the on-line tests, taken at a computer display. Other computer learning materials are also available. The tests have in them a large amount of learning material. As soon as a student is in d i f f i c u l t y , he or she is given aid which is -s p e c i f i c a l l y related to the d i f f i c u l t y . Each test is unique. Passing is at the competency l e v e l ; students either demonstrate that they know the material or are asked to study further and then take another variant of the test. In 10 weeks we give about 15,000 individual tests to 400 students. The computer keeps the f u l l class records. 4 The National Science Foundation provided support. The second project i s concerned with s c i e n t i f i c l i t e r a c y . I t hopes to acquaint students with some fundamental notions about science: What i_s a s c i e n t i f i c theory or model? How is such a theory discovered? How do we use i t to make predictions? What determines i f i t is a good theory or a bad theory? The material, currently six two-hour units, i s designed for a general.audience, with i n i t i a l testing done extensively i n the public l i b r a r y . The materials have also been tested in junior high schools, high schools, community colleges, and u n i v e r s i t i e s . Support was from the Fund for the Improvement of Postsecondary Education.5 The third project aims at helping students become formal operational in the Piaget sense. The primary l e v e l i s junior high school. The format for these units i s simi l a r to that for the science l i t e r a c y materials. The project i s supported by the National Science Foundation.^ Computer Advantages Given a brief view of several a c t i v i t i e s involving the computer in learning, we can now say why the computer i s such a powerful learning device. At least two factors are c r i t i c a l in considering the effectiveness of the computer in aiding learning, the'"'TnT'er-a^EitwB^ and *;hfl 1 i *-y to^^d-£gT3Wfti,i ?.a fchp 1 paf»ainq ftyrp&fi*aasa»»tefiT«fe^^ One of the major problems in education, p a r t i c u l a r l y education which must deal with very large numbers of students, is the f a c t that we have lost one of the most valuable components in e a r l i e r education, the p o s s i b i l i t y of having learners who are always playing an active role i n the learning process. In c l a s s i c a l Greece, with the Socratic approach to learning, two or three students worked closely with Socrates, answering Socrates' questions and therefore behaving as active learners. The process was highly labor intensive. As we had more and more people to educate i t became less and less possible to behave i n this way. We cannot afford or produce enough master teachers to base our educational system on the Socratic approach. But we can develop good computer based learning material in which the 5 student Is always active. The computer may enable us to get back to a much more humanistic, a much more friend l y , educational system by making a l l of our learners participants rather than the spectators they frequently are in our present book- and lecture-learning environments. The second advantage offered by the computer i s ind i v i d u a l i z a t i o n of the process of learning. Everyone says that students are d i f f e r e n t , that each student i s unique, that each student learns i n d i f f e r e n t ways. But most of our standard learning procedures, such as the lecture, are very weak i n allowing for these individual differences. They t y p i c a l l y treat most students i n the same way. For example, i f a student i n a particular point i n a course lecture i s lacking some important background information, that student is swept along i n our t r a d i t i o n a l courses with everyone else in the c l a s s . The missing information i s hard to acquire under those circumstances. The rational procedure would be to allow the student needing special help to stop the major flow of learning at that point and to go back and pick up the background information. But most of our present structures for learning have no adequate provisions for such a p o s s i b i l i t y . The actual needs vary between what can be learned i n a few minutes and what can be learned i n a whole course. With the computer the situation i s e n t i r e l y d i f f e r e n t . Each student can move at a pace best for that student. Each student w i l l be responding frequently to questions. (We have found i n our recent programs that a student responds about every f i f t e e n seconds). So the computer, with curriculum material prepared by excellent teachers, can determine what the student understands or does not understand at a given point. Remedial aid can be given where appropriate, simply as part of the flow of the material with no break from the student point of view. Indeed, the student, using well-prepared computer based learning material, does not have the impression that any "special" treatment i s taking place, so no psychological stigma is attached to such aid. With the individualization possible with computers, one can hope to achieve the goal of mastery learning, where everyone learns a l l material ess e n t i a l l y perfectly. So much for "why" computers are going to become the dominant educational delivery system. The two factors mentioned, the unpleasant situation i n education today and the usefulness of the computer as a way of learning p a r t i c u l a r l y i n dealing with large numbers of students, suggest to me that the computer w i l l move rapidly forward i n education. But we s t i l l must look at the other side of the question, the "how" of the development. That i s , how do we move from our present situation, where computers are l i t t l e used i n learning, to a situation i n which they are the dominant delivery system? This i s the subject of the next section. HOW WILL WE MOVE TO MUCH GREATER COMPUTER USE? 6 157 Let me f i r s t recapitulate e a r l i e r information. The period ahead in education, for at least ten years and probably longer, is l i k e l y to be one of tremendous turmoil and s t r i f e . We are just beginning to see the outlines of that s t r i f e at the present time. The s t r i f e w i l l be increased greatly in education as we begin to move toward such ideas as voucher systems and more detailed accountability. The t r a d i t i o n a l methods of preserving the status quo in education, or allowing only small incremental changes to take place, such as the power of the administrators and the unions, w i l l have r e l a t i v e l y l i t t l e effect; much of the turmoil in schools w i l l be imposed from the general community. Often changes w i l l be generated by f i n a n c i a l decisions which lead to less money to the schools. The challenge w i l l be the most serious one that has been seen in a very long time in the educational system. The following comment by Peter Drucker gives a view of the situ a t i o n from outside academia: "In the next ten or f i f t e e n years we w i l l almost c e r t a i n l y see strong pressures to make schools responsible for thinking through what kind of learning methods are appropriate for each c h i l d . We w i l l almost certainly see tremendous pressure, from parents and students a l i k e , for result-focused education and for accountability in meeting objectives set for individual students. The continuing professional education of highly educated raid-career adults w i l l become a third t i e r in addition to undergraduate and professional or graduate work. Above a l l , attention w i l l s h i f t back to schools and education as the central ca p i t a l investment and infrastructure of a 'knowledge society'." 7 Thus, we w i l l have a society more and more unhappy with the current educational system, a society groping for new ways to handle education. Few "solutions" to the problem w i l l be apparent. Home Computers During the same period of time computers, p a r t i c u l a r l y personal computers, w i l l be decreasing in cost, increasing in c a p a b i l i t i e s , or (more l i k e l y ) some mixture of these two trends. The changes w i l l often be d r a s t i c . While the term one hears in the computer industry, TTTTCTTfiT I'nflrrivrTiif^r i s intentionally something of an exaggeration, i t does r e f l e c t what i s happening in many areas of computer technology. One aspect of the rapid development of personal computers that w i l l be extremely important for the future of education w i l l be the increasing presence of the computer in homes. Homes w i l l represent the largest possible market for personal computers, since in no other situation can one speak of millions of units. There are approximately eighty m i l l i o n American homes; so the number of computers which can be sold for home use, provided the ordinary person can be convinced that the computer i s valuable to 7 own, is enormous. The home w i l l be education too, since the commercial be very great. In a sense, education i s never " f i r s t " with computers. For many years we piggybacked on essentially a business or s c i e n t i f i c technology in computers with education only a poor follower. The new situation w i l l be similar, but with the home market the dominant one. To s e l l computers for the home, i t w i l l be necessary that they do something. The average home owner i s not going to buy a computer on the grounds that they are currently being sold to homes, primarily for hobbyists. The home user of equipment buys an appliance, a device such as a refrigerator or stove that accomplishes some task or tasks. They don't buy a gadget that they can put together in various ways to accomplish d i f f e r e n t types of tasksl The size of the home market w i l l depend on the s k i l l of vendors in convincing people that the computer in the home w i l l be useful to the average person. Some estimates have suggested sixty million computers i n homes i n ten years. I do not wish to imply that a single appliance-like use of the computer w i l l drive the home market. On the contrary, a variety of such uses are l i k e l y to be important. Home word processing, for example, w i l l be an extrenely important use. Home fi n a n c i a l systems, complete enough to keep a l l the f i n a n c i a l records and write the income tax when asked to, and to aid in home fi n a n c i a l decisions, w i l l also be of importance. Personal record keeping systems, including class notes, l i s t s , and similar uses, are also l i k e l y to be of major use in the home. F i n a l l y , educational material w i l l be one of the types of material that without question w i l l drive the home market. The size of this market w i l l depend on the quality and quantity of such appliance-l i k e programs. Thus, we w i l l find learning material based on the computer being developed for home computers, in some cases almost independently of whether i t w i l l also be usable in elementary and secondary schools, university, or other learning environments. Schools w i l l use the material developed primarily for education in the home even though i t may not be i d e a l l y suited. It may be that this material w i l l often have more careful thought put into i t than some of the earli e r products developed pa r t i c u l a r l y for the school environment, simply because the potential market i s so much larger and users more discriminating. Schools are already desperately searching for computer based learning material and are finding that l i t t l e good material i s available. The people who are using the new learning materials in the home w i l l be coming to cur schools and universities. ThioyiiMiill aHCTwynatey^^ i f the educational institutions wish to survive, they w i l l provide i t . 152 the driving force for pressures for home sales w i l l 8 I S3 I an taking what may seem to many of you, given the nature of the audience, to be a very market-oriented point of view. But we aust be r e a l i s t i c in trying to plot the future. We must understand that the most fundamental issues that w i l l determine the future are these marketing issues, not the academic issues which may be at the forefront of our own minds. Companies When we look at the school market, we see interesting commercial pressures. The dominant s e l l e r s of educational materials to schools today are the commercial textbook publishers. Yet commercial textbook publishing i s a s t a t i c domain at almost a l l levels of publishing. That i s , i t i s d i f f i c u l t for a company to make much progress there, i n the sense of increasing p r o f i t s . Education i t s e l f i s getting declining amounts of money. There w i l l be declining numbers of students for many years. The competition between companies is f i e r c e . To end up with a much larger share of that market at the present time, considered purely as a textbook market, i s extremely d i f f i c u l t . So i t i s not surprising that many of the most i n f l u e n t i a l textbook publishers are now beginning to devote sizable amounts of e f f o r t , attention, and money to computer based learning. They see this as a new market, where i t i s not at a l l clear at present who w i l l become dominant. Thus, a minor textbook publisher could see the p o s s i b i l i t y of becoming a major computer based learning publisher, or a major publisher could see that computer based materials would very much increase revenues. Or a new company could see this as a particular opportunity for advancement, allowing them to leap over the established companies. A l l these situations are happening now. The l i s t of textbook publishers putting sizable resources into computer based learning i s a distinguished one. I t includes such names as John Wiley, Harper & Row, Scott Foresman, Science Research Associates, McGraw-Hill, Random House, Encyclopedia Brittanica, and many others. The type of involvement i s d i f f e r e n t i n different companies—this i s , after a l l , a new market, one that i s poorly understood by everyone. The degree of involvement also d i f f e r s from company to company and i s l i k e l y to d i f f e r i n time. In addition to these established companies, new companies, often particularly devoted to either educational software or to personal computer software more generally, are coming into existence. Sizable amounts of venture c a p i t a l are available for such companies. These companies, old and new, w i l l be s e l l i n g their wares, and so more and more school d i s t r i c t s and universities w i l l be able to e a s i l y acquire computer based learning materials. Both old and new companies w i l l have people actively s o l i c i t i n g school business. The older textbook companies may want to t i e i n the computer material with their existing textbooks, but 9 the newer companies w i l l have no need for t h i s , and so may be open to more adventuresome a c t i v i t i e s . Some of the companies w i l l be s e l l i n g to a combination of the home and school market. In general the materials developed for the home market w i l l be available in the school market too. Schools Given the turmoil and f i n a n c i a l r e s t r a i n t s i n the schools, the commercial pressures, the pressures created by the home market, and the increasing effectiveness of the computer as a learning device, more and more schools w i l l turn to computers for del i v e r y of learning material. Indeed, we can already spot this happening, although in a minor way. One interesting sign i s the fa c t that many schools, p a r t i c u l a r l y small schools, no longer have adequately prepared teachers to teach many of the important courses in the curriculum. Thus i f we look at high school courses such as trigonometry, advanced mathematics, and science courses, r u r a l schools i n the United States presently are often not providing these c a p a b i l i t i e s , at least not i n a way that i s competitive with the better large urban schools. Computers w i l l be a mechanism for equalizing opportunity for students by providing computer based learning courses in these declining areas, courses that otherwise would not be available. Hopefully, these courses w i l l be developed by the best individuals from a l l over the country. We may see a decreased role of the formal school and the formal university in our educational system. Much education w i l l be able to take place in the home i n a f l e x i b l e fashion. At the univ e r s i t y l e v e l we already see one outstanding example of a development of this kind, The Open University in the United Kingdom, but s t i l l with r e l a t i v e l y l i t t l e use of computers. The Open University has demonstrated that good curriculum material i n home environments can be effective as a learning mode and economical as compared with the standard cost of education. Voucher systems, i f they are enacted, w i l l make home learning much more l i k e l y . I do not wish to imply that a l l education w i l l move to the home. Indeed, a view of the educational system such as that shown i n George Leonard's book, Education and Ecstasy, suggests that the so c i o l o g i c a l components, the factors associated with l i v i n g with other people and li v i n g with oneself, w i l l s t i l l probably best take place i n small group environments within schools. But many of the knowledge-based components of learning may move to the home. Types of Usage We have discussed very l i t t l e about the way computers w i l l be used within the school system. Something needs to be said about t h i s , i f only to counteract some of the current propaganda. I wish to go on record as stating that the computer w i l l be used in a very wide variety of ways within our educational system. The notion that some "right" way exists to use the computer, and that other modes of computer usage are somehow wrong, is one that has been promulgated, I am afraid, by a number of individuals and groups i n recent years. Indeed, often staged debates at meetings comparing types of usage have been held, with the implication that there are right and wrong ways to use the computer in education. Books have been organized in such a way that i t sounds as though there were a competition for different types of computer usage. These debates, often on philosophical grounds, have made a t a c i t assumption that a right way to use the computer exis t s , i f only that way could be discovered. Mostly the authors have had a naive b e l i e f in their "right" way, and then set out to try to esta b l i s h a case for their b e l i e f s . The p r i n c i p a l problem with this type of reasoning i s that i t often does not proceed from in s t r u c t i o n a l bases, nor does i t proceed from empirical bases, experimental studies. That i s , the issues that dominate are often technological issues, the nature of the computer hardware and what can be done with the computer hardware. These writers are trying to carve some unique niche for the computer among other learning media. These taehnologically-based and media-based arguments for a single type of computer usage are, I believe, e n t i r e l y misleading. The decisions as to how to use computers—the modes of computer usage, the areas—should be made en t i r e l y on pedagogical grounds, the questions of what aids learners rather than on these philosophical, media, or technological grounds. Whenever decisions are made on pedagogical grounds, i t w i l l be found that a wide variety of computer uses w i l l be employed, uses which are often adapted to the individual situation being considered. There is no single "right" way to use computers, but rather a great variety of ways. I w i l l give a brief c l a s s i f i c a t i o n of the various ways the computer can be used. This l i s t i s not exhaustive nor does i t show fine d e t a i l . But i t may be useful to at least consider the range. Computer Literacy. Computer l i t e r a c y i s i l l - d e f i n e d and so much debated. I t i s recognized that at a l l l e v e l s of education, sta r t i n g perhaps as early as eight or nine years old and continuing through the school system, university, and adult education, that individuals in our society need to understand the various ways the computer i s going to be used i n that society; they need to understand the positive and negative consequences of those ways. Few f u l l - s c a l e computer l i t e r a c y courses exist. Indeed, what often passes as computer l i t e r a c y i s vague history or learning to program i n a simplified way, to be discussed i n a moment. So this i s s t i l l very much an open area for computer uses. Specialized courses are needed for each group addressed; thus, computer l i t e r a c y for teachers i s a pressing national issue. 1<*Z A l l these courses need to consider such important future uses as word processing, personal financial and record keeping systems, and educational material. Learning to Program. Learning to program i s already a rapidly increasing a c t i v i t y in our universities and schools. It represents in grade six through twelve the most common usage of computers at the present time. Unfortunately, where i t happens at this l e v e l i t i s often a disaster, harming more than helping the student. The major problem is the way programming is taught. A whole group of people i s being taught a set of techniques which are no longer adequate to the programming art today. These techniques were common in the early days of computing, but they are inadequate according to today's standards. Many of the people learning to program in junior high school and high school cannot overcome the i n i t i a l bad habits which have often been i n s t i l l e d in them when they come to the universities. Many u n i v e r s i t i e s are now reporting this phenomenon. The main c u l p r i t i s BASIC. It is not that BASIC has to be taught in a way that is a n t i t h e t i c a l to everything we know about programming today. But i t almost inevitably _is_ taught i n such a 2&Sd&£S£23ttt&t^^ Indeed, fashion. the analogy i s close in that junk food tends to destroy the body's desire for better types of food. But the analogy is weak in one regard: BASIC i s the i n i t i a l language of the vast majority of these people. It i s as i f you started feeding junk food to babies one day old and didn't give them anything else u n t i l they were six] If I could leave you with one raessace, perhaos the roost message pressing The following recent comment by a disntinguished computer sc i e n t i s t , ?fiftg3SraEffip^^, i s relevant: JenLs "LrfJ What programming languages should we teach? There are a number of p o s s i b i l i t i e s for junior high and high schools. "tJiigjGj^is certainly one interesting p o s s i b i l i t y , although I must confess that some features of Logo are different from those recommended in the best modern programming practices. Logo, however, i s introduced in a problem solving environment, and that i s very much to i t s advantage. Often i t s main intent i s presented not to teach programming but to teach more general problem solving c a p a b i l i t i e s 11.3 or some s p e c i f i c area of mathematics. But i t s general problem solving effectiveness has yet to be demonstrated in our mass school environments with ordinary teachers. Another good p o s s i b i l i t y is^gggj^or ,^^§f^^^^^^SnHt l^ . The material developed at the University of Tennessee and sold by McGraw-Hill under the name of "Computer Power" is an excellent example of an approach of this kind. If one looks for p r i n t material that is usable at the high school and perhaps even at the junior high school l e v e l at the present moment, the "Computer Power" material looks to me to be e a s i l y one of the best p o s s i b i l i t i e s . Another approach is to develop some interesting c a p a b i l i t y based on a ' a g ^ ^ S ^ i i ^ i^pgjgg^^^^^K^cagff For example, the recent Karel, The Robot from Wiley follows such an approach. Turtle geometry, in Logo, is the best known example. Learning Within Subject Areas. Undoubtedly the largest use of the computer in schools at a i l levels w i l l eventually be not the categories just discussed but rather the use of the computer as an aid in learning mathematics, in learning to read, i n learning to write, in learning calculus, and i n a l l the other tasks associated with the learning process. One person may work alone at a display or several may work together. When one looks at these learning tasks in d e t a i l , again one"finds a great variety of computer use, ranging from t u t o r i a l material, to intuition building, to testing, to aids in management of the class for the student (feedback on what is needed and how to go about getting i t ) , and the teacher. The three projects presented earlier show something of the range of p o s s i b i l i t i e s . Unfortunately, much of the material now available of this type is very primitive. We are, however, rapidly learning to develop better material to aid learning. PRODUCTION PROCESS If we are to move to meet this new future, where the computer w i l l be the dominant educational delivery system, a c r i t i c a l aspect w i l l be the generation of e f f e c t i v e learning material. We need new courses and entire new c u r r i c u l a , spanning the entire educational system. Hence, the development we are talking about i s a n o n t r i v i a l process. It i s the degree of success of the development process that w i l l t e l l whether we w i l l improve or hurt education. We must convince the l i k e l y distributors that i t i s important to develop quality materials, not the junk t y p i c a l l y available today. The development of curriculum material i n any f i e l d and with any medium and at any le v e l i s a d i f f i c u l t process. It cannot be done by amateurs who are doing i t simply as a spare time a c t i v i t y . Many new observers in this f i e l d , looking at the problems quickly, tend to underrate these problems of developing e f f e c t i v e learning material. Hence, some of the solutions which have been proposed are solutions which are simply not adequate to the problems. Some of these solutions assume only small incremental changes in the curriculum structure and do not understand the magnitude of the development neceszry. We cannot discuss f u l l y i n this paper a l l the aspects of the production process. The Educational Technology Center has extensive l i t e r a t u r e available concerning these issues for those interested. . Several c r i t i c a l points concerning products should be made to give the reader a reasonable o v e r a l l viewpoint. The production system i s a complex system, one that should involve many types of people with many diffe r e n t s k i l l s . I f one looks at the production of any educational material, one sees that that i s the case. We can learn much by examining effective curriculum production systems, such as that currently i n use i n The Open University, that used in producing the major curriculum e f f o r t s i n the United States more than ten years ago, and that involved i n such areas as the development of textbooks. What we need to r e s i s t i s the notion that one person, perhaps a teacher in his or her spare time, w i l l do i t a l l . I do not believe that any sizable amount of good curriculum material w i l l be produced by this method. Furthermore, I do not believe that the devices which are being urged for these teachers, such as simple-minded authoring systems based on toy languages (Pilot) w i l l be effective. Nor do I think that languages such as Tutor w i l l be effective, because they do not meet the reasonable c r i t e r i a associated with modern programming languages. Most of these languages are old i n their design, and few of them understand the nature of structured programming. A serious professional approach i s needed i f we are to maintain the quality of the computer based learning materials produced. We can see a number of stages needed in such a professional approach, l i s t e d below. a) Preplanning b) Establishing goals, objectives, and rough outlines c) Specifying the materials pedagogically d) Reviewing and revising this s p e c i f i c a t i o n e) Designing the s p a t i a l and temporal appearance of the material f) Designing the code g) Coding h) Testing in-house 1) Revising j) F i e l d testing k) Revising The l a s t two stages may be repeated twice. In the entire process the educational issues, as opposed to the technical issues, should be dominant. The best teachers and ins t r u c t i o n a l designers should be involved i n stages c and d to assure the quality of the product. This paper has presented an overview of some of the problems associated with reforming an entire educational system during the next twenty years. Many details are either not mentioned or treated very hastil y . But I hope I have given enough d e t a i l s to convince you of the main directions that need to be taken. As teachers, most of you are undoubtedly interested in what you should do now to work toward a more eff e c t i v e future for education. F i r s t , you must decide whether you would l i k e to be involved in the type of curriculum development I suggested w i l l be necessary. If you do want to be involved, you must take a long-range view of how to prepare for this a c t i v i t y . I would not advise you to buy an Apple and start to use i t l Nor, as you might suspect, would I advise you to take courses in BASIC. But i t would be desirable to take a variety of courses, i f they are accessible to you or to study on your own, in certain areas. Here are some suggestions. The f i r s t three refer to areas of learning, either through formal courses or through informal methods. 1. Learning theory. Good curriculum development cannot be developed without some appreciation of how people learn, even though there is no single coherent theory there. Courses in learning theory may help, based on the research l i t e r a t u r e concerning learning. 2. Curriculum development. The question of how to develop good curriculum material i s one that deserves serious study. Some un i v e r s i t i e s provide such courses. Some textbooks exist. Many of the issues are independent of computers, referring to developing with any learning media. 3. Modern programming languages. You might want to become acquainted with modern programming languages, such as ^ ^s^^» and *3§t> Again, you must be careful here. I t i s possible to meet these languages either in an old fashioned environment or in one that stresses structured programming. You want the second p o s s i b i l i t y . Look at the textbook. If i t doesn't introduce procedures u n t i l a third of the way or even further along, don't take the course. This isn't the only factor, but i t i s a good way of distinguishing reasonable from unreasonable courses. Avoid the "CAI" languages—they are inadequate, not suitable for serious material. Look at the authoring approaches based on PRESENT STEPS 15 modern structured languages. 4. Listen to students. In your own teaching, begin to move away from the lecture mode presentation into a more Socratic mode. A c r i t i c a l factor i s listening to what students say and watching what they do. This means that when you ask questions, you have to wait for answers! It also means working more in d i v i d u a l l y with students i n groups of two to four. I t i s only by this procedure that you w i l l begin to build up the insights you need for how students ac t u a l l y behave when they are learning. People whose primary mode of interaction with students i s through the lecture mode or through textbooks are seldom the best choices for preparing computer based learning material. The development of computer based learning material w i l l need vast numbers of experienced teachers, teachers who have been li s t e n i n g to their students and who understand student learning problems. 5. Personal computers. Begin to use a variety of personal computers, with particular emphasis on the new generation of 16 b i t machines. Read the journals that t e l l you about new equipment. Watch for voice input, better graphics, and f u l l multimedia c a p a b i l i t i e s . 6. C r i t i c a l attitude. Look at a good b i t of computer based learning material, trying to develop a c r i t i c a l attitude toward i t . Don't be overwhelmed simply because i t i s interactive or because the computer i s involved. Keep your mind on the learning issues and learn to develop some s e n s i t i v i t y as to what existing material helps learning and what doesn't. Most existing material is poor. Pind out why. Read the journals that specialize in c r i t i c a l reviews. 7. Work with others. The development of good computer based learning material i s best done in a group. Work with others in discussing goals, strategy, and the d e t a i l s of design. 8. Future orientation. Concentrate on the long-range s i t u a t i o n , not today or tomorrow. Decisions which are "good" from a short-range point of view may be undesirable in the long range to both you and to the future of our entire educational system. So keep the long-range point of view strongly i n mind. 9. Visions. Begin to think about what type of future educational system would be both desirable and possible. If you want to influence the future, you must have visions. •Developing quality computer-assisted instruction demands forethought; those of you who are unfortunately caught up in expedient movements in education need to take a closer, more courageous look at the nature of the hope on Pandora's chip. You're dealing with as powerful a tool as the gods have ever given us."^ lfc7 References: 1. Servan-Schreiber, Jean-Jacques, The World Challenge. New York: Simon and Schuster from The Mitsubishi Report. 2. Staky h. r t i e w ^ M ' wi th Cw...pul&j»a. D l l l B i i i a , Masii'aihusefciai* 3. Boyer, E., quoted i n "Report on Educational Research," February 3, 1982. 4. Bork, A., "Computer-Based Instruction in Physics." Physics  Today, 34, 9, (September 1981). 5. Bork, A., Kurtz, B., Franklin, S., Von Blum, R., Trowbridge, D., "Science Literacy in the Public Library." Paper, Association of Educational Data Systems, Orlando, February 1982. Von Blum. R., "Computers in Informal Learning: A Case Study," November 1980. Arons, A., Bork, A., Collea, F., Pranklin, S., and Kurtz, B., •Science Literacy in the Public Library - Batteries and Bulbs." Paper, Proceedings of the National Educational Computing Conference, Denton, Texas, June 1981. 6. Trowbridge, D. and Bork, A., "A Computer Based Dialog for Developing Mathematical Reasoning of Young Adolescents." Paper, Proceedings of the National Educational Computing Conference, Denton, Texas, June 1981. Trowbridge, D. and Bork, A., "Computer 3ased Learning Modules for Early Adolescence." Paper, World Conference on Computers in Education 1981, Lausanne, July 1981. 7. Drucker, Peter F., The Changing World of the Executive. New York: Times Books, 1982. 9. Quote from Burns, H., "Pandora's Chip: Concerns about Quality CAI," Pipeline, F a i l 1981. 11 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0097772/manifest

Comment

Related Items