Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Effects of students’ background vs. school conditions on primary schools students’ academic achievement… Chinvarakorn, Vasana 1993

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

Item Metadata


831-ubc_1994-0021.pdf [ 2.45MB ]
JSON: 831-1.0087273.json
JSON-LD: 831-1.0087273-ld.json
RDF/XML (Pretty): 831-1.0087273-rdf.xml
RDF/JSON: 831-1.0087273-rdf.json
Turtle: 831-1.0087273-turtle.txt
N-Triples: 831-1.0087273-rdf-ntriples.txt
Original Record: 831-1.0087273-source.json
Full Text

Full Text

EFFECTS OF STUDENTS’ BACKGROUND VS. SCHOOL CONDITIONS ONPRIMARY SCHOOLS STUDENTS’ ACADEMIC ACHIEVEMENT IN THAILANDbyVASANA CHINVARAKORNBA. (1St class, Hons.), Chulalongkorn University, 1988A THESIS SUBMITTED IN PARTIAL FULFILLMENT OFTHE REQUIREMENTS FOR THE DEGREE OFMASTER OF ARTSinTHE FACULTY OF GRADUATE STUDIES(Department of Anthropology and Sociology)We accept this thesis as conforming,1t6e required standardThE UNIVERSITY OF BRITISH COLUMBIANovember 1993@ Vasana Chinvarakorn, 1993In presenting this thesis in partial fulfilment of the requirements for an advanceddegree at the University of British Columbia, I agree that the Library shall make itfreely available for reference and study. I further agree that permission for extensivecopying of this thesis for scholarly purposes may be granted by the head of mydepartment or by his or her representatives. It is understood that copying orpublication of this thesis for financial gain shall not be allowed without my writtenpermission.__________________________Department of Ci Sci’oloy”The University of British ColumbiaVancouver, CanadaDate Nvv,er iisDE-6 (2/88)IIAbstractThis study attempted to investigate whether or not the results derived fromresearch conducted in industrialized countries, about school resources andschool conditions having very little effect on students’ academic achievementwhen students’ background is controlled for, apply to a developing country suchas Thailand. Data on Grade-6 students from the BRIDGES Project in 1987-1988were analyzed; however, the aggregated nature of data only allowed aninvestigation at the school level. The analyses show that both sets of variablesrepresenting students’ background and school conditions have a significantimpact on students’ achievement scores. However, the national-level results donot agree with results from any single region. The study concluded that specificsocial and economic conditions in each locality seem to have a significant impacton how students’ background and school conditions affect students’ academicachievement; therefore, one should not assume that results from research inindustrialized countries will necessarily apply to a developing country, or thatnational-level results will apply to regions within the country.IIITable of ContentsAbstract iiTable of Contents iiiList of tables viAcknowledgement viiChapter 1 Introduction 1Equality of Opportunity 2Background of Thailand: regional variations 4Primary education in Thailand 8Historical background of Thai primary education 8Chapter 2 Literature review and theoretical background 161. The influence of socioeconomic background onstudents’ achievementResults from Western industrial countries 16Results from research in developing countries 18Results of research in Thailand 222. The influence of school and teacher characteristics onstudents’ achievementResults from research in industrialized countries 25Results of research in developing countries 28Results of research in Thailand 33Chapter 3 Research Methodology 37Sample Population 37Data collection 38ivResearch instruments 39Research Variables 41Units of Analysis 47Analytical procedures used in this study 50Treatment of missing values 51Chapter 4 Results and Discussion 534.1. The quality of data 534.2. Regional variations in achievement scores 534.3. Regional variations in students’ background andschool conditions 554.4. Preliminary Analysis 584.5. Comparisons of the relative effects of students’background and school conditions on students’achievement in Thailand: results from multipleregression analyses 614.5.1 The influence of students’ background onachievement scores: 63Aggregate students’ socioeconomic status 63Aggregate non-material measures ofstudents’ background 684.5.2 The influence of school-related variables onstudents’ academic achievement 714.5.3 The effect of ‘overlapping’ variables 754.6. Regional differences in variable effects 77V4.7. Comparisons of the relative effects of aggregatestudents’ background and school conditions onacademic- versus non-academic subjects 78Chapter 5 Conclusion 81Equality of educational opportunity in Thailand 85Limitations of data and analyses 87Suggestions for future research 88Conclusions 89Bibliography 91Appendix A: List of educational regions, provinces, and districtsin the study 104Appendix B : Outlier cases 106Appendix C: Correlation matrices of variables selected in the study 108viList of tablesTable 3.1 Variables selected for this study 41-42Table 4.1 Average standardized achievement scores in fivedifferent subjects for the five regions in Thailand 54Table 4.2 Average values for students’ background and schoolcondition variables in five different regions of Thailand 56Table 4.3 Distribution of achievement scores by provincial socioeconomic categories (using GPP per capita as a proxymeasure for SES) 58Table 4.4 Distribution of achievement scores by school size (asa proxy measure of school resources) 59Table 4.5 Distribution of achievement scores by number of textbooks available 60Table 4.6 Rank orders of independent variables for each subject,nationally and within each region 62Table 4.7 The effect (B coeflicients) of students’ backgroundvariables on achievement scores in five subjects forprimary school students, in Thailand, 1987 64Table 4.8 The effect (B coefficients) of school-related variableson achievement of primary school students in Thailand,1987 72Table 4.9 The effect (B coefficients) of ‘overlapping’ variables onachievement of primary school students in Thailand,1987 76VIIAcknedgementI would like to thank the following people and organization for providing mewith support and assistance in order for the research to be completed. Dr. NeilGuppy, my senior research supervisor, for his patience, guidance, andencouragement throughout the project. The other two members of my researchcommittee, Dr. Tissa Fernando and Dr. George Gray, have also providedvaluable comments to improve the thesis draft.A sincere thank goes to the Canadian International Development Agency(CIDA) for providing me a scholarship to do graduate studies at UBC.I would like to thank Dr. Panom Pongpaibool and Dr. Chinnapat Bhumirat,of the National Education Commission (NEC) of Thailand, who allowed meaccess to some of the data set from Project BRIDGES, on which the analyses inthe present study are based. Dr. Bhumirat also took some time to clarify part ofthe data set which has facilitated the research.Personal thanks to my family for constant encouragement, and inparticular, my sister, Aporn (then) Chinvarakorn, for sending me some of theThai literature.Finally, Brian Chisholm has consistently given me moral and intellectualsupport. Over the years of my stay in Vancouver, he has taught me a fewimportant things. In particular, without his help and that of his family, it isunforeseeable how the present study would ever have been completed.1Chapter 1IntroductionIn modern society, one measure of social change is the extent to whichindividual success or failure rests on ability and effort, rather than on social originor parental socioeconomic status. Greater stress on equality of educationalopportunity is one of the most common signals of this shift from an ascriptive(who you are) to achievement (what you can do) principle.Research in the sociology of education, of which the majority is done in theUnited States, has consistently found that schooling has relatively less effect onstudents’ academic achievement than does social background. This implies thatschools may not be able to moderate the influence of students’ socioeconomicbackgrounds, and thus not reduce the gaps in academic achievement amongstudents from different social origins.Findings from research in developing countries, on the other hand, tend to bemore equivocal. Some research even suggests that schools may exert a greaterinfluence on students than does their socioeconomic background (Heynemanand Loxley, 1983). Thus this line of research points to schools as a potentialagent for bringing about more effective equality of opportunity, and recommendsthat an increase in educational expenditure together with a redistribution ofschool resources (particularly those that are found to have significant effects onstudents’ achievement) should be carried out if the desire is to reduce theinequalities in achievement among students from various backgrounds.The purpose of this research is to evaluate the relative effects ofsocioeconomic background and school quality on the scholastic achievement ofprimary school students in Thailand. It is not known whether primary schools in adeveloping country like Thailand will have independent effects on students’academic achievement relative to that of social background. Moreover, the2effects of social origins and schools may vary among geographic regions andcommunities with different degrees of economic development as suggested byprevious Thai research (Setapanich, 1982). To understand these effects, the roleof primary schooling in promoting equality of educational opportunity will beassessed by comparing the degree to which the resources of Thai primaryschools impact on the achievement of students with the effect of the students’socioeconomic backgrounds on achievement.To facilitate the comparison, this research will attempt to answer the followingquestions: 1) what effect does school resources have on students’ academicachievement?, 2) what effect does social background have on students’achievement?, 3) do the effects of social origin or school vary among differentgeographic regions?, and 4) most importantly, which effect is stronger, the schooleffect or the effect of socioeconomic context? Answers to these questions willallow a comparison of the results of Thai research with those from othercountries, especially America. These questions will be discussed in greaterdetail in the following chapters.Equality of OpportunityThe notion of equality of opportunity was originally used to mean that eachindividual should have an equal chance to succeed or to fulfil their potential,regardless of their social origins (Turner, 1986). According to this initial conceptof equality of opportunity, the society was responsible for providing an equal‘opportunity’ (in such areas as schooling, health care, elections, free markets) toevery person regardless of that person’s sex, age, racial and ethnic origin, orsocioeconomic background. However, the achievements or failures of eachindividual were attributed to that person’s own talent or ability. The originalconcept of equality of opportunity assumed that the ‘differences’ between3individuals were natural, inevitable, and not transferable from generation togeneration (Green, 1988).More recently, the idea of equality of opportunity has been closely related toschool-based education. Schools were perceived as relatively autonomousinstitutions where each individual was admitted and recruited on a universalisticcriterion (i.e., a person’s performance determined his/her success at school). Byattending school, each person has an equal chance to be equipped with‘knowledge,’ to move up the social ladder. Therefore, schools are regarded as animportant agent of social mobility.The concept of equality of opportunity has been criticized, partly due to thepersistent differences in achievement between different socioeconomic groups(Hum, 1985: 111 -112). Individual talent and ability are argued to result morefrom the differences in status origins than from innate abilities (see Coleman,1990). Equal access or ‘opportunity’ is then perceived as inadequate, sincehigher-status children may grow up in a more supportive environment, providingthem with an advantage in ‘cultural capital’ over lower-status people. Poorchildren may have to drop out of school in order to help their family earn money.Poor parents may not encourage their children to study diligently, nor socializethem in a way that makes them ready for life in school. A completely free andcompetitive society is thus no longer considered an ‘equal’ society. Some arguethat society is expected to provide not only equal formal ‘opportunity’, but alsoequal ‘conditions’ for every individual through social welfare and education, aswell as positive discrimination in favour of the ‘disadvantaged’ groups (cf. Green,1988:3).Under this latter view, schools are perceived not only as an agent of socialmobility but also as an agent of social equality. Schools are expected to be ableto effectively bridge the differences in socialization patterns among different4social groups, in order to bring about an equal level of achievement among allstudents (Coleman, 1968). Various educational programmes have been set up inattempts to provide equal ‘conditions,’ especially for lower-status students, suchas the Head Start Program in the U.S. which attempted to prepare minoritystudents for Grade 1, or the programmes to integrate students from differentracial backgrounds with a belief that high-achieving students will help or have apositive influence over low-achieving students (Hum, 1985: 112, 129-130).Changes in the conception of equality of educational opportunity haveresulted from debates on what actually constitutes an equal opportunity. Theoriginal idea of equal opportunity, that emphasized an equal access to schoolresources, has been shown to be inadequate since children from the lower socialgroups appear to have already lagged behind children from higher social groupsby the time they start school. Attempts to prepare lower-status children for life inschool or to provide extra help and resources when they are in school result fromthe belief that every child, regardless of his/her social origin, should have anequal opportunity to compete. The concept of equality of opportunity that hasevolved thus emphasizes the effectiveness of schools in providing equal‘conditions’ for each child. In other words, the effect of school comes to beevaluated against the influence of the child’s socioeconomic background.Background of Thailand: regional variationsTo examine whether students’ background and school conditions havedifferential impacts on achievement in Thailand, two pertinent aspects of thiscountry should be considered, namely, 1) how does Thailand compare with othercountries in levels of economic development, and 2) how do different regions inthe country compare with one another.5On the basis of national average per capita income (US $ 1,454, in 1990),Thailand is considered to be a lower-middle income country (World Bank,1 990a). The country’s economy, which has shifted from agricultural-based, tomore industrial and service-based, has grown rapidly and somewhat consistentlysince the late 1960’s (Fry, 1992: 93; World Bank, 1 990b: 7, table). To some,Thailand thus appears to be approaching the rank of Southeast Asia’s newlyindustrialized countries (NICs) (see Fry, 1992: 84-85). However, its dramaticeconomic growth is mostly limited to Bangkok Metropolitan areas, while themajority of the population continues to reside in the rural areas (Economist, 1991:36; Fry, 1992: 85). The agricultural sector remains Thailand’s largest employer,taking in about 65 per cent of the total labour force in 1987 (World Bank, 1990b:11, table). Similarly, Thailand remains largely a rural country, with the exceptionof Bangkok, the largest city, and a few other cities in the regions outside BangkokMetropolis (Knodel et al., 1987: 28-30). In addition, its rapid economic growthhas concealed large regional differences in levels of economic development (seeTDRI, 1987: 42), which are somewhat associated with ethno-linguistic differencesamong populations in each region (see Kaplan, 1980: 61).Thailand appears to be relatively homogeneous, i.e., the majority ofpopulation are Theravada Buddhists (about 95 per cent), and speak a languageof the Tai family (about 85 per cent or more) (Tuchrello, 1989: 69). However,differences in social, economic, and political conditions do exist between regionsand ethno-linguistic groups (Tuchrello, 1989: 69; Cohen, 1992; Keyes, 1987).The country can be divided into four ‘natural’ regions (i.e., the Central, North,Northeast, and South) on the basis of landforms and drainage (Kaplan, 1980:55). Although this regional classification does not entail any administrativesignificance, it appears to coincide more or less with the differences in ethnolinguistic background of population, basic resources, and level of social and6economic development (Kaplan, 1980: 55; see Cohen, 1991: 12; cf. Moore,1974: 5). In addition, Bangkok Metropolis, which is in the Central region, shouldbe considered a fifth region, due to its distinct social, political, and economicconditions (see Knodel et al., 1987: 30; Tuchrello, 1989: 62).Generally, those who speak the Standard Central Thai dialect, concentratedin Bangkok and some central provinces, have had advantages over people in therest of the country. Economically, Bangkok and to a lesser extent, the Centralregion, have consistently enjoyed much higher average income than those in theother regions, due largely to the concentration of commercialized agriculture, andindustrial- and service-based industries there (see Fry, 1992: 87, table;Chowdhury, 1989: 53). Linguistically, the Standard Central Thai dialect, which isdifficult to learn for speakers of other dialects, entails high status and prestige tothe speakers, since it is the sole official language required in schools andgovernment offices (Kaplan, 1980: 63; see Diller, 1991: 99-100; Tuchrello, 1989:70-71). Politically, Bangkok has been the centre of power and decision makingthat has influence on the rest of the country (Moore, 1974: 4).The North has the second lowest per capita income of the country (Knodel etal., 1987: 47, table; see also Fry, 1992: 87, table). Increasingly, there has been ashift from subsistence-based agriculture to a more commercialized one, whichhighlights the problems of landlessness and concentration of land ownership (seeGanjanapan, 1989; Turton, 1989), observed earlier in the Central region(Chiengkul, 1983). Northern residents have their own dialect, called Lanna Thaior Kham Muang (see Keyes, 1987: 6), which has been neglected until recently(see Tuchrello, 1989: 72; Diller, 1991: 115). The hilltribes people in thenorthernmost parts of the region also have their own distinct languages, andhave been engaged in shifting cultivation of various crops, including opium (seeTuchrello, 1989: 75-77).7The Northeast (or Isan) is the poorest region of the country (see Fry, 1992:87, table TDRI, 1987: 42, table; see also Economist, 1991: 36) With most areasbeing dry and arid, the majority of the northeastern population (about 31 per centof the country’s total population) continues to rely on subsistence agriculture(Moore, 1974: 4-5). A large number of the northeasterners have consistentlymigrated to work in Bangkok, a few cities inside and outside the region, or toforeign countries (see Economist, 1991: 36; TDRI, 1987: 43). Generally, thesenortheasterners speak a variety of the Lao language (Cohen, 1992; Diller, 1991),and to a certain degree, have developed regional consciousness on the basis ofthis shared linguistic and cultural background (see Tuchrello, 1989: 81 -82;Keyes, 1987; Cohen, 1992).The South generally ranks third in per capita income (see Fry, 1992: 87, table;TDRI, 1987: 42, table). Its population is engaged in a variety of economicactivities, such as mining, lishing, and rubber production. Southerners have theirown dialect (Pak Tai), which also belongs to the Tai language-family with anumber of Malay loan-words (Diller, 1991: 95; Kaplan, 1980: 63). Moreover, alarge number of Muslims are concentrated in the four border provinces, andspeak Malay language (e.g., Yawi). In the past, some of these Muslims, who aregenerally of lower socioeconomic status than other ethnic groups in the region,were involved in separatist movements against the central government; however,the government has recently introduced various conciliatory measures thataccept the distinct religion and ethno-linguistic characteristics of the southernMuslims (see Pitsuwan, 1985; Satha-Anand, 1987: 3; Tuchrello, 1989: 82-83).The success of these government-run programmes remains to be seen.In sum, the differences in levels of economic development closely associatedwith ethno-linguistic backgrounds of the Thai population appear to coincide withgeographic boundaries of the country. The ‘regional’ differences have8contributed to ethnic and I or regional consciousness (Kaplan, 1981: 61), whichhas occasionally led to separatist movements in the past, such as in theNortheast and the South (Cohen, 1992: 12), and may possibly again in the future(see Keyes, 1987: 14, 201). It is not known whether and how the differencesamong regions in Thailand will contribute to the impact of government-controlledprimary schooling, as a major instrument by the Central government to unifydifferent ethnic groups and regions, on students’ academic achievement.Primary education in ThailandPrimary school can be considered the most important level of formaleducation in Thailand for the following reasons: 1) it has the largest enrolmentand accordingly requires the largest proportion of the national budget foreducation, 2) it is the only formal education that the majority of Thai peoplereceive, as compulsory education exists only at this level, and 3) it is the base forhigher education, or on-the-job training (NEC, 1977). In fact, it is estimated thatabout 80 per cent of workers, in all sectors of the country, have only primaryeducation (Myers and Sussangkarn, 1991: 49, table 12).Historical background of Thai primary educationPrior to the introduction of modern education, education in most parts ofThailand was in the monastic tradition of Theravada Buddhism. The wais(temple-monasteries) were the centres of learning, wherein Buddhist monks livedand taught village boys how to read and recite religious texts (Keyes, 1991: 90-91). The relationships between the monks and villagers were considered sacred,whereby the latter usually paid a high degree of reverence and obedience to theirteachers (Hanks, 1958, 9). In addition to literacy in religious texts, villagers couldalso learn various crafts and other indigenous forms of knowledge such as9midwifery, preparing corpses, traditional medicines, astrology, legends andmyths, poetry or customary laws, through an apprenticeship with the monks orother learned villagers (Hanks, 1958: 9-10; Keyes, 1991: 91 -92).The expansion of Western colonial powers (especially Great Britain andFrance) into Southeast Asia during the nineteenth century posed significantthreats to the political sovereignty and stability of Siam (as Thailand was knownthen). Thus the central government of the country at that time sought tostrengthen its control over various parts of the country, to improve its militaryorganization, and to reform its administrative systems (Sudaprasert et aL, 1980:203). Several Western values and ideas were also adopted in programmes tomodernize the country, one of which is the development of public education.Modern (i.e., secular) education was perceived as one of the more effectivetools to help modernize and unify the country wherein a variety of linguistic andethnic groups reside (Na Thalang, 1970; see Myers and Sussangkarn, 1991).The growing administrative bureaucracy also demanded a large number ofliterate government employees (Wyatt, 1969; Samudavanija, 1987: 222). Inaddition, an expansion of education seemed also to bring about an increase inequality of educational opportunity for every Thai citizen, as intended by KingRama V (Chulalongkorn, 1868-1910) who laid the foundation for moderneducation in the country,so that everyone - be it my children or the children of my poorest subject,being children of a noble lord or children of a slave - shall receive an equalopportunity in education. I hereby declare that education in this country isour first priority and that full development in education must soon takeplace. (Kasemsak, 1974, cited in Chantavanich et al., 1990:15)Primary education is the first level of the formal schooling system that wasintroduced nation-wide. A Royal Proclamation, inviting Thai people to send theirchildren to school, was issued in 1885, the same year as the establishment of thefirst government school outside the Royal Palace (Jumsai, 1951: 21; Watson,101980: 255). The promulgation of the Primary Education Act in 1921 requiredevery child to attend school, and this law has been implemented in mostcommunities throughout the kingdom since the 1 930s (Keyes, 1991: 89). TheBuddhist Sangha (monks) community was actively involved in the earlydevelopment of secular schooling through providing their temple buildings forschools and some monks as teachers (Jumsai, 1951: 15, 45, table), while thecentral government worked on building new schools and training lay teachers.However, the initial steps in the expansion of state-controlled schools were slow,in part due to a shortage of trained teachers (Jumsai, 1951; Wyatt, 1969;Watson, 1980; Grandstaff, 1986). Massive construction of (primary) schools inalmost every village did not begin until the 1960s and 1 970s followingrecommendations of the Karachi Plan of 1960 (Watson, 1980: 57-59, 218).Since 1959, the Thai government has dramatically increased its nationaleducation budget, and in particular, allocated over half of the budget to theprimary level (Watson, 1980: 152-153; Chantavanich and Fry, 1985: 5238).However, at least until the Educational Reform (see below) school resources(e.g. budget per student, teachers’ qualifications) were not distributed equallyamong different provinces and types of schools (Sudaprasert et a!., 1980: 243-246, 252, 257-259; Leonor, 1982; NEC, 1974).Modern education differs from traditional education in many respects. The lawnow requires every child to attend school for a certain period of time, unless it isextremely difficult to do so. The curriculum has also been made uniformthroughout the country and covers more subjects, such as arithmetic, generalscience, history, geography, and (boy) scouting, with the content of somesubjects being biased toward conditions in Bangkok (the largest city and capitalof the country). The Central Thai dialect is used as the sole language ofinstruction. Monks were replaced by lay teachers who are government11employees; however, the respectful relationships between students and teacherscontinue to prevail (see Gurevich, 1972). A teacher is generally perceived bystudents as a venerable person whose behaviour has to be almost perfect(Suvannathat, 1978, cited in Avalos and Haddad, 1981: 7). The centralgovernment has control over the administration, finance, curriculum, andpersonnel of all schools except the private ones. In addition, the implementationof formal education allowed women to attend schools for the first time, which hasresulted in a dramatic increase in the enrolment of female students in primaryschools (Jumsai, 1951: 41 -42; Watson, 1980: 107).Responses to modern education are varied. Many parents perceive modernschooling as a means of social mobility for their children (Hanks, 1958; Saradattaand Savannathat, 1973) allowing them to move out of farming into governmentservice and other sectors. On the other hand, some parents see moderneducation as lacking relevance to the conditions in their localities and of theiroccupation (Hanks, 1958; cf. Wyatt, 1975: 146). A resistance to state-sponsorededucation is especially evident among the Southern people whose Muslimbackground differs sharply from the prevailing Buddhist culture in the rest of thecountry (see Dulyakasem, 1991). However, this resistance to modern educationmay gradually fade away, considering the increasing scarcity of lands and theconcentration of land in the hands of a small group of people, which makes itincreasingly difficult for people to survive solely by farming. Educationalqualifications may become very important for obtaining positions in the modernjob markets and they can be obtained only through schooling (see Chantavanichetal., 1990:149).Equally important is the finding that the nation-wide implementation of primaryeducation has been followed by disparities in the achievement levels of studentsfrom different geographic regions and types of schools. These educational12disparities have been characterized as the differences between the core, usuallyreferred to as Bangkok and the neighbouring provinces in the Central region, andthe periphery or the rest of the country (Sudaprasert et aL, 1980: 206).In standardized achievement tests of the Thai language and arithmetic carriedout in 1973, Grade-3 students in Bangkok obtained average scores twice as highas those in the Northeast. The scores of students in the Central region, theNorth, and the South were somewhere in between (NEC, 1974: 6, 8, tables;NEC, 1977: 25, table 1). Similar patterns in the distribution of achievementscores persisted in another nation-wide test carried out in 1980 (Chantavanich etal., 1990: 29, table 6).In 1967-1977, under the old curriculum, wherein primary education lasted forseven years and was separated into lower and upper primary levels (Grade 1-4,and Grade 5-7), only about 30-58 per cent of Grade-4 students went on to Grade5 (Chantavanich eta!., 1990: 2; Leonor, 1982: 106-108), with high disparities inenrolment between provinces in the Central region and those in the other regions(Sudaprasert et a!., 1980: 222-229; Fry, 1983: 209). This may be due in part tothe previous use of selective examinations for students who wished to continueto the upper primary level (Grade 5) (Sudaprasert et a!., 1980: 222), and ashortage of upper primary schools and teachers especially in the rural areas(Watson, 1980: 161). In addition, there were wide regional disparities in theproportion of students who completed their primary education. During 1972-1975, in the Central region, which includes Bangkok, about half of the studentswho had started in Grade 1 reached Grade 7, whereas the proportion was about10 per cent for students in the frontier provinces of the Northern and Southernregions of the country (Leonor, 1982:108, 113-114).Prior to the Educational Reform in the late 1 970s, there were dual systems forthe administration of primary schools in the urban and rural areas. Rural schools,13which made up about 90 per cent of all primary schools in the country, wereunder the authority of the Ministry of Interior (MCI). Schools in the urban areasconsisted of three major categories, namely private schools, municipal schools,and schools under the authority of the Ministry of Education (MOE). The differentschool systems appeared to cater to students from different social backgrounds.In the urban areas, many children from elite and middle class backgroundsattended private and MOE schools, while children from low-income families wentto municipal schools. In the rural areas, MCI or provincial schools served themajority of children, who came from a farming background (Sudaprasert et aL,1980: 251 -252). In general, provincial schools ranked the lowest in the quantityand quality of school resources available to their students (NEC, 1974: 12-15).Likewise, the levels of students’ scholastic achievement, and of repetition anddropout rates, in provincial schools were inferior to those of the other types ofschools (Leonor, 1982: 117, Table 7; Sudaprasert eta!., 1980: 251 -254).This problem of educational disparities has been of concern to policy makers.Following the student led political movement in 1973, the Thai governmentattempted to redress these problems, especially Ath regard to the equity, quality,and relevance of primary education. An Education Reform Committee was setup in 1974, followed by the initiation of the National Education Scheme in 1977.There were attempts to allocate basic educational resources more equitablyacross provinces (Fry, 1983: 206; Ketudat, 1984: 526-527; cf. Sudaprasert et a!.,1980: 257-258). Nation-wide studies on Equality of opportunity in primaryschools (NEC, 1974; Leonor, 1982: 116) and The factors affecting scholasticachievement (NEC, 1977; Leonor, 1982: 116-120) called for large-scale changesin primary education in four major areas: 1) the unification of diverse educationalorganizations under the MOE, and the decentralization of educationaladministration to the local provinces, 2) establishment of new curriculum goals14leading toward an appreciation of the relation between education, life, andsociety, 3) improvements in quality and relevance of education, and 4) specialemphasis on non-cognitive learning and moral values (Chantavanich and Fry,1985: 5237). The new curriculum included four learning areas: basic skills(literacy, arithmetic), life experience, habit formation, and work orientation. Thetwo levels of primary education were also merged, and the duration of this levelof schooling was reduced to six years.As part of the Educational Reform programme, the administration of provincialschools was returned to the Ministry of Education in October 1980, under theOffice of the National Primary Education Commission (ONPEC), which isresponsible for policy making and planning. However, actual day-to-dayoperations are controlled by the Provincial Primary Education Commissions,which have flexible control over placement of teachers, selection of schooltextbooks and materials, and location of schools (Chantavanich and Fry, 1985:5237). Each school has its own educational committee which is in charge of theschool’s non-academic management and operations, and which consists of bothleaders and other residents of the community where the school is located(Chantavanich etaL, 1990:150-151).Despite the government’s attempts to reduce the differences in educationalachievement and attainment among primary school students, the disparities incompletion rate and achievement levels appear to continue. Almost all studentsin Bangkok complete the sixth grade, the terminal year of primary education,while only 56 per cent of students in the South do (Myers and Sussangkarn,1991: 24, citing World Bank, 1990). Moreover, from 70 to 90 per cent of studentsin Bangkok continue to the secondary level while less than 10 per cent of primaryschool students in rural areas continue their studies beyond the primary level(Komin, 1989: 47, Fry, 1983: 207). From 1985 to 1988, the scores of Grade 615students in the mathematics and Thai language tests (World Bank, 1990; seealso NEC, 1990a and 1990b) showed a similar pattern of disparity betweenBangkok and the rest of the country, with the South and the Northeast regionsbeing particularly low.Modern education has been introduced in Thailand partly in response to thepenetration of Western colonial powers into Southeast Asia during the nineteenthcentury. Education has been perceived as an important tool to develop thecountry, to unify various ethnic and linguistic groups of people. However, existingdisparities in educational achievement among different geographic regions led toattempts by the Thai central government to reform the educational systems, butwithout much success. The present study will investigate possible factors thatcontribute to differences in the average level of achievement of primary schoolstudents, as measured by the standardized tests in 1987. Previous studies offactors that contribute to differences in the academic achievement of Thaiprimary school students will be discussed in the next chapter.16Chapter 2Literature review and theoretical backgroundEducational research has consistently found a large gap in the learningachievement between students from different socioeconomic backgrounds. Thishas spurred interest in highlighting the factors that affect academic achievement.Two major sets of variables that have been identified are those relating to familyconditions and those relating to school conditions. However, the general resultsof research in Western industrial countries appear to contradict those of researchin developing countries: in industrially advanced countries family socioeconomicbackground is much more influential on students’ achievement than are schoolresources, while the opposite seems to hold for the developing countries.Differences in national educational policies as well as in the social and economicconditions between industrialized and developing countries raise concerns aboutthe generalizability of research findings across the two groups of countries.1. The influence of socioeconomic background on students’ achievementResults from Western industrial countriesResearch in industrialized countries has generally found that socioeconomicbackground, at both individual and aggregate levels of analysis, has a greaterimpact on students’ academic achievement than do school and teachercharacteristics. In Equality of Educational Opportunity one of the most importantsurveys on this issue, Coleman et aL (1966) show that student’s background(especially socioeconomic status) appears to have a significant and persistentinfluence on the achievement of students at both elementary and secondarylevels of schooling. Subsequent large-scale survey research done in the UnitedStates, where the majority of educational research has been conducted, and inother industrialized countries tends to confirm the results of Coleman et al. s17report (Jencks at a!., 1972; Sewell and Hauser, 1976; Rutter at at, 1979;Anderson, 1967; Armor, 1972; Hanushek, 1972; Cummings, 1977).The measures of student’s background that have been used in large-scalesurvey research include: educational and occupational status of parents, familyincome, family size, and number of possessions in the home (see Bridge at al.,1979: 21 3-228). Bridge et at (1979: 215) noted that most of the large-scalesurvey research is more interested in estimating the effects of school and teachercharacteristics; therefore, the selection of measures for family socioeconomicbackgrounds is largely to control for the differences in students’ backgroundsbefore they enter school, and thus may not necessarily correspond to the actualdifferences between families. However, the research findings about the strongimpact of students’ socioeconomic background on their academic achievementsuggest that these conventional measures of socioeconomic status in industrialsocieties do adequately reflect the differences between families in the practicesof child rearing, which in turn influences the children’s academic achievement(Bridge etat, 1979).Status attainment researchers, who studied male high school students inWisconsin over more than a decade, found that students who had low SESorigins tended to exhibit lower ‘ability,’ as measured by lQ tests administered atabout age fourteen, and also lower aspirations for educational achievement thanstudents from higher status backgrounds (Sewell and Hauser, 1976). Whendividing the index of socioeconomic status into quartiles, the researchers foundthat the highest quartile had a 4-to-i advantage over the lowest in reachingcollege graduation, and a 9-to-i advantage in attaining graduate or professionaleducation.The influence of student’s background is felt mainly in the socializationprocesses (see Kerckhoff in Richardson, 1986). Students from a high status18family and/or more developed community may be encouraged to develop highaspirations for education. Middle-class children may grow up in an environmentwhich prepares them for socialization in school, through the use of moresophisticated language and social manners compatible to what school expects(see Bernstein, 1973, 1974, 1976, cited in Hum, 1985: 192-193; Bourdieu andPasseron, 1991).In sum, research in industrial countries has consistently observed a strongand significant influence of socioeconomic background on students’ academicachievement. Parents’ educational and occupational status appears toadequately indicate patterns of child socialization which results in differences inacademic achievement.Results from research in developing countriesDifferences between Western industrial societies and the developingcountries in terms of national educational policies, the structure of socialstratification, as well as in the roles and meanings of “family” cast doubt on thegeneral applicability of research findings about the influence of students’socioeconomic background on their academic achievement. The findings ofresearch in some developing countries are generally more equivocal thanfindings of research in Western industrial societies about the persistent influenceof students’ socioeconomic background. Some even suggest that familysocioeconomic background may have a weaker impact on students’ achievementthan do school resources (Heyneman and Loxley, 1983).Burstein et a!. (1980) argued that different national policies on howeducational resources are distributed and managed in each community may leadto differential effects of students’ family! community background on theireducational achievement at the school level. In their comparisons of the effects19of students’ socioeconomic backgrounds, aggregated at the school level in theUnited States, England, and Sweden, they found that the socioeconomic statusof a community had a greater impact on students’ achievement at the schoollevel in the United States and England than in Sweden (although they did not finddifferences in the SES effect on student achievement within schools). Burstein etal. (1980) argued that local communities in the United States and England hadgreater control over the style of programs offered, the curriculum, and the abilityof the school to seek better facilities and personnel, and thus students’achievement was more dependent on the community’s socioeconomic status.On the other hand, the pattern of resource allocation in Sweden was morecentralized (which is similar to many developing countries), and thus the level ofcommunity socioeconomic status may not have as much influence over students’achievement. The differences in academic achievement between schools inSweden were much lower than in the U.S., and England.Foster (1977) suggested that in many non-Western societies (modern)educational and occupational status may not relate to differences in child rearingpractices that influence the child’s academic achievement to the same extent asthey do in Western countries. The levels of parental education and occupation indeveloping countries may not affect the value attributed to the children’seducation (Fagerlind and Munck, 1981, cited in Heyneman and Loxley, 1983:1182, footnote), nor the self-concept of children. Thus the conventionalmeasures of students’ socioeconomic background (e.g. the parents’ educationand occupation) being used in research in industrial societies may not be valid foridentifying the differences between families in many Third-World countries.Heyneman (1979: 177, table) found no relationships between students’socioeconomic background and their self-attitudes, which were identified inAmerican research as contributing to students’ learning performance (Coleman et20aL, 1966; see Heyneman 1979:176-177). Ugandan students from low socialorigins did not seem to have lower opinions about themselves than those fromhigher socioeconomic backgrounds (cf. Lockheed et aL, 1989: 245-246).Heyneman (1976, 1979; see also references in Heyneman, 1982: 135-136)also found that the correlation between achievement of primary school studentsand their father’s education in Uganda was very weak. He obtained similarresults with other indicators of students’ socioeconomic backgrounds, similar tothe ones used in the Coleman report, (i.e., mother’s education, number ofmodern possessions in the home, father’s occupation, and the summary measureof the four SES variables). (Heyneman and Loxley [1983: 1176-1177] claimedthat the lower degree of variance in students’ SES backgrounds in developingcountries did not contribute to the weak effects of students’ SES. For example,they could not find any particular patterns of relationships between mother’seducational attainment and student’s scores in developing countries). In anotherstudy, Heyneman (1977) compared the average scores of primary schoolstudents from communities in Uganda with different levels of development. Hefound that students from more backward communities did better than those fromurban backgrounds. Heyneman pointed to a smaller proportion of the school-agepopulation in backward regions who attended schools, and postulated that theselection processes there may be more competitive and relatively restricted tomore talented students. Elsewhere, Heyneman (1979: 177-178) asserted thatthere was no relationship between SES and the performance of students in bothbackward and more developed communities of Uganda.Some research suggests that other measures may be more relevant to thesocial and economic conditions in the Third-World countries. Dialect, castemembership, amount of land holding, type of residence, access to electricity,nutritional status, and parent’s demands for their children’s labour, have been21proposed as measures that better indicate the differences in social and economicbackgrounds among the population in the less industrial societies than doparents’ educational and occupational status (Schiefelbein, 1979: 138).Research in India, Malawi, and other developing countries examined some ofthese variables and found that they affected students’ academic achievement(Lockheed et a!., 1989: 242, 250-253). The effects of some of these identifiedvariables will be tested in the present study.In addition, the concept of ‘family’ may differ between Western societies andnon-Western ones. While the term family in Western societies usually refers tofather, and/or mother, and child(ren), in many cases of rural societies familycovers more generations of members and/or kin. Theisen et aL (1983)suggested that poor families in Third World countries may be able to rely onresources provided by their kin, and thus their own socioeconomic status may notreflect their ability and support for children’s education as would be the case inWestern societies. (However, the data available to me do not allow me toexplore the issue.)Very few studies looked at the possible differences in effect of students’socioeconomic background between regions and areas of residence (urban vs.rural), and few use non-urban samples (Lockheed et a!., 1989: 241). Forexample, the Thai sample in Heyneman and Loxley (1983) was limited to urbanschools in Bangkok and nearby areas, and it was thus not possible to explore thevariations among regions.The influence of students’ background appears to vary by school subject (e.g.reading vs. science). Students’ background in Uganda was found to have a weakbut statistically significant impact on their achievement in English, while it did notseem to have any significant effect on other subjects (such as mathematics,general knowledge in history, geography, and general science) (Heyneman,221976: 47). SES was shown to have less impact than school variables on Greekstudents’ achievement in physics (Kostakis 1987).Evidence from research in some developing countries about the weakerimpact of socioeconomic background may be due to 1) the centralized pattern ofschool resource allocation, 2) different bases of social stratification other thansocioeconomic status, and 3) different meanings of ‘family.’ The effects ofsocioeconomic background may, however, vary among geographic regions andschool subjects.Results of research in ThailandThe results of studies on the influence of students’ socioeconomicbackground on students’ achievement in Thailand are rather mixed. The impactof socioeconomic background on students’ academic achievement remainslargely uncertain. Jso, the effect of socioeconomic background seems to varyamong different geographic regions, largely due to the differences in social,economic, and cultural backgrounds in each region.A re-analysis of data from a nation-wide survey (Setapanich, 1982: 127, table)found that student socioeconomic background (measured by father’s educationand occupation as well as the degree of exposure to newspapers, television, andmagazines) appeared to have less impact on Grade-3 student achievement thanschool and teacher related variables, in all regions. On the other hand, inanother survey in 1979-1980 (Chantavanich et aL, 1990) found that theirmeasure of students’ SES (the regularity of children having pocket money to taketo school) had a stronger effect on Grade-3 students’ test scores in Thai andarithmetic than did other school variables. The analysis of Chantavanich et al.was based on aggregate-level data and at the national level only.23In a longitudinal study on the effects of various measures of student’sbackground on mathematics achievement of lower secondary school students inThailand, Lockheed et a!. (1989: 244, 246, 248, table) discovered that mother’seducation and father’s occupation (conventional measures of socioeconomicbackground) contributed little to students’ achievement, On the other hand, themotivational variables, such as students educational expectations, perceivedparental support, and attitudes toward one’s ability and usefulness ofmathematics for future occupation continued to exert a moderate and significantimpact on the gains in students’ achievement scores. Since father’s occupationand mother’s education did not seem to be strongly related to the motivationalvariables, the researchers contended that these motivational variables may beconsidered additional family background variables (Lockheed et a!., 1989: 246).It is possible that student’s socioeconomic background may not have as strongan impact on achievement of secondary school students since these studentsmay be a select group. (Only 30 per cent of Thailand’s school-age populationcontinue beyond the primary level. In the rural areas, about 10 per cent ofpnmary school graduates do so [Komin, 1989].) It was not possible to estimatethe effect of parental encouragement on achievement of students at the primarylevel since no analysis has been done on this issue (NEC, 1977).It is not clear whether the effect of students’ background in Thailand ishomogeneous among different regions or not. NEC (1977: 37, table) observedthat contrary to expectations, family SES had a lower impact on studentachievement in Bangkok than in the other regions of the country. In addition,Setapanich (1982: 128-129) found that in the Southern region (where the religion,language, as well as the ethnic origins of the majority of the population aremarkedly different from the rest of the country), SES seemed to have a strongerimpact on student achievement in Thai language than in mathematics. She24postulated that the Muslim resistance to the Thai central government (and thus togovernment-run schools) in the South may be higher among lower SES groups(e.g. fishermen, farmers) than among higher SES groups (mostly governmentofficials) who foresee greater benefit of education. On the other hand,Raudenbush etal. (1991: 264) found that the effects of SES and age onachievement in ma language and mathematics were quite homogeneous acrossschools. They also found that the average SES in the school was significantlypositively correlated with maths and Thai language achievement in both urbanand rural areas (Raudenbush etaL, 1991).A possibility that the measures of family SES may not be valid in all regions ofthe country poses a major difficulty to previous Thai studies in estimating theinfluence of family socioeconomic background on students’ achievement.Setapanich (1982: 133, table) discovered that the mean scores of students fromeach paternal occupational category varied among geographic regions. Shepostulated that the status of government officers may be relatively lower inBangkok, but higher in the Northeast where the majority of the population are inthe agricultural sector. In a preliminary survey in both urban and rural areas ofThailand (Larpthananon and Wongkiattirat, 1992), the observed SES indicators(education, occupation, and income) did not seem to represent the differencesamong rural populations to the same degree as they did in the urban areas (i.e.,when using those indicators, the distribution of population in the urban areas wasnormal, while it was highly skewed in the rural areas). There is also a wide rangeof differences in socioeconomic status within each occupational category. Forexample, the category ‘farmer’ may include large landowners, tenants, oragricultural labourers (Setapanich, 1982: 59, 79, 132).It is not clear how well socioeconomic status predicts students’ achievementin different regions of Thailand, due to specific social, economic, and cultural25conditions in each regions. A number of questions thus arise regarding theeffects of family socioeconomic background on students’ achievement inThailand:1. Does aggregate students’ SES predict average achievement scores?2. How well do other measures of students’ background (e.g., dialect,absenteeism rate) predict the average achievement scores?3. Do the effects of SES (or other measures of social origin) varyamong regions (e.g. Bangkok vs. the South, the Northeast)?4. Do the effects of SES (or other measures of family backgrounds) vary byschool subjects: Thai vs. mathematics, life experience, work experience,and character development?2. The influence of school and teacher characteristics on students’achievementResults from research in industrialized countriesGeneral findings from American research show that the differences in schooland teacher characteristics between schools are quite small; moreover, thesesmall differences do not seem to exert a significant influence on students’academic achievement, when their family backgrounds are controlled (Colemanet aL, 1966). Research in other industrial societies tends to support the results ofAmerican studies (Peaker, 1971).Large-scale survey research has examined the effects of both materials(school equipment, textbooks) and personnel (mainly teachers’ qualifications,experience, and teaching methods) (Bridge et a!., 1979: 235-283; Fuller, 1986:appendix a). In order to assess the impact of certain school variables across alarge number of schools in large-scale survey research, most of these measureshave to be general and thus are rather crude.26Most large-scale survey research on the effects of schools in America usuallyrelies on school average characteristics and thus may underestimate the effect ofschool-related variables on students’ performance. Heyns (1986: 311-312)pointed out that most research that uses the between-school variance inachievement scores to determine the effect of a particular school variable, suchas the Coleman report, will tend to underestimate the effect of schoolcharacteristics since it assumes that all students in the school are similarlyaffected by the school variables. On the other hand, individual-levelcharacteristics, such as socioeconomic background, are bound to have greater“effects” on achievement than school-level variables, since only 15 to 25 per centof the total variation in achievement lies between schools, while 75 to 85 per centis within any one school (Heyns, 1986).School and teacher quality appear to have only a small impact on students’achievement. Most research thus generally supports the notion that animprovement in provision of certain ‘desirable’ school resources, such as thenumber of books available, the student-teacher ratios, and the qualifications ofteachers, will not necessarily improve the performance of students. Jencks et aL(1972: 109) argued that equalizing the quality of elementary schools wouldreduce the disparities in achievement scores by only 3 per cent or less, whileequalizing the quality of high schools would reduce the disparities by only 1 percent or less.Although large-scale survey research in the United States tends to find onlyminor effects of school factors on students’ achievement, the results are not clearand conclusive. School effects seem to vary among different ethnic groups,areas of residence, and level of schooling.School resources appear to have relatively greater effect on the achievementof ethnic minority students, who tend to come from low socioeconomic27backgrounds (Coleman et al., 1966). Integrating high-achieving students with thelow-achieving ones seems to have a more positive effect on black students thanon white students (Thornton and Eckland, 1980: 247-250; Meyer, 1970: 59-70).The Coleman report (1966: 305) also found differences between urban andrural areas in the effect of the social composition of the student body, measuredhere by the turnover rate of students in schools. In the Southern U.S., students’achievement in rural schools was high where the turnover rate was high,whereas in the North, the achievement of students in urban schools was lowwhere the turnover rate was high. This implies that the school effect may varyregionally.Finally, elementary schools seem to exert relatively more influence onstudents’ performance than high schools, as reported in Jencks et aL (1972).Research on the effects of high school tracking, with controls of students’ pastperformance, show a very small effect of tracking on students’ achievementscores (Jencks eta!., 1972; Heyns, 1974; Alexander and Cook, 1982). On theother hand, data obtained by observing classroom interactions in elementaryschools show the effect of ability grouping assignment on the level of readingskills of students (Eder, 1981; McDermott, 1977). Although this line of researchsuffers from a small number of cases and difficulty in controlling the initialcharacteristics of students, it reveals the cumulative effect of teachers’expectations on students’ performance. Rist (1970) and Mackler (1969) foundthat once the initial ability group assignments had been made, they were likely tobe permanent with little or no mobility between groups even in higher grades (seealso Eder, 1981; McDermott, 1977).Research in industrialized countries has generally observed a small effect ofschool and teacher quality. However, the school effects may vary among28students from different socioeconomic backgrounds, regions, and level ofschooling.Results of research in developing countriesDespite a relatively large quantity of research on school effects in America(Jamison et aL, 1981; Heyneman, 1986) which leads to a pessimistic view aboutschools’ ability to moderate the influence of students’ family backgrounds (seeSimmons and Alexander, 1979), some findings from studies recently conductedin developing countries question the general results of American research aboutthe small effects of school-related variables. Comparing the effects of school andteacher characteristics across 29 high- and low-income countries, Heynemanand Loxley (1983) found that the effects of school-related variables appeared tocorrelate negatively with the level of economic development of the country (seealso Fuller, 1986; Fuller and Heyneman, 1989). Assumptions about researchmethods, differences in the distribution patterns and availability of schoolresources, as well as the roles of modern (often Western-styled) schooling in thedeveloping countries raise the possibilities that the effects of school and teacherattributes on students’ achievement there may be stronger than would be thecase in the Western industrialized countries.An assumption that family or school-related variables should have similareffects on students’ achievement in i[ societies may lead to an underestimationof some potential variables. Heyneman and Loxley (1982) re-analysed the datafrom the survey by the International Association for the Evaluation of EducationalAchievement (lEA) in 18 high- and low-income countries, and estimated theeffects of school-related variables for each country separately. They found asubstantial increase in the effect of school and teacher variables among the lowincome countries (cf. Passow et al., 1976).29An assumption about linear relationships between school-related variablesand students’ achievement may also underestimate the effect of school andteacher characteristics. Most of the school resources in developing countries arestill below the standard considered minimum in industrialized countries (Jamisonet aL, 1981; Fuller and Heyneman, 1989). Some school variables may havecertain threshold effects beyond which their variance does not contribute to anysubstantial increase in students’ achievement (Mosteller and Moynihan, 1972;Schiefelbein, 1979:139-140; cf. Bridge eta!., 1979: 22-25), which may result inthe small effects of school resources found in American research. Schiefelbein(1979: 138) suggested that while American research such as Jencks et aL (1972)may assume that every student in American schools had at least a minimumnumber of textbooks and thus was more interested in their ‘quality,’ as measuredby the prices of the books, the availability of textbooks itself may be important forthe academic achievement of students in less industrial societies.There seems to be little difference in the availability of educational resourcesbetween different schools and for students from different socioeconomicbackgrounds in the U.S. (see Coleman etaL, 1966), which may partly explain thelow degree of school effects. On the other hand, school resources in thedeveloping countries seems to be more unevenly distributed (Schiefelbein, 1980:137-139; Inkeles, 1979: 401; Heyneman and Loxley, 198), and thus they may bemore effective in influencing student achievement.It is not known whether the distribution of school resources in developingcountries is more biased toward students from higher SES backgrounds than inindustrialized countries or not. Heyneman and Loxley (1 983b) found that thecorrelation of GNP per capita and the school quality/SES coefficients was notstatistically significant, which means that there seem to be no significantdifferences between high- and low-income countries with regard to students’30access to school resources. On the other hand, the skewed distribution of schoolresources may still pose a serious problem of multicollinearity between two majorsets of independent variables, if the school and teacher attributes are highlycorrelated with students’ socioeconomic status (see Coleman, 1975). Students’SES may correlate with the amount of educational resources in the schools thatthey attend. For example, Heyneman (1975: 56-57, 60, table) found that inUganda the location of schools (whether they are near urban centres or not) andthe average socioeconomic status of the school was highly correlated with theamount of textbooks received by each school (cf. Niles, 1981). It may not bepossible to resolve the problem of multicollinearity (Pedhazur, 1982), and thusthe interpretations of school effects will have to be extremely cautious. However,in light of the recent claims by governments in Third-World countries onredistribution of educational resources as part of the programmes to reducedisparities in educational outcomes (for example, see the Educational Reformprogramme in Thailand mentioned in the first chapter), it is important toinvestigate first, whether the allocation patterns of school resources havebecome more equal or not, and second, what effects they have on students’achievement.Although it is suggested by American research that in order to properlyassess the effect of school and teacher characteristics on student achievement,these variables should be matched to each student in a school, Heyneman andJamison (1980: 212) argued that using aggregated data for school variables maybe appropriate for the educational system in Uganda (and thus to otherdeveloping countries with similar systems), because, pedagogy, curriculum,teachers and other specialist personnel, school equipment and materials werecontrolled by the central government and the differences that occurred werebetween schools, rather than within them. Also, in such cases, schools are31usually the basic unit for the central government to plan and decide on how toallocate educational resources.Formal schooling, which many developing countries adopted from theWestern societies, may be foreign to local traditions, and thus family orcommunity may not be able to assist students in homework (Saha, 1983). Manyparents in rural communities do not question what is taught in schools becausethey feel that they are not qualified to criticize teachers’ knowledge or thecurriculum that is planned from the central government (see Chantavanich et aL,1990: 164). (Lareau [1993] noted similar situations with respect to lower-classparents’ participation in school activities in the United States.) In such cases, thedifferences in family socioeconomic backgrounds may not be as influential as thecharacteristics of schools and teachers in predicting the academic achievementof children.General findings from large-scale survey research in developing countriesindicate stronger impact of school and teacher attributes than would be found inAmerican research. Also, some experimental studies on the influence ofavailability of textbooks on students’ achievement in developing countries (e.g.,Nicaragua, Philippines) have found a small but significant impact for this school-related variable (Jamison et al., 1981; Heyneman et al., 1984). Reviews ofstudies on the influence of teacher characteristics on student achievement indeveloping countries (Husen et aL, 1978; Avalos and Haddad, 1981; Saha, 1983)showed that teacher characteristics (e.g. sex, qualifications, experience,personality) may be more influential than would be the case in Western industrialsocieties.Results from research in developing countries on the influences of specificschool and teacher variables are still few in numbers and in many cases,inconsistent (Simmons and Alexander, 1980: 77-95; Schiefelbein and Simmons,321979; cf. Heyneman, 1980: 150-155). Some research yields similar findings tothose found in industrialized countries, about the small effect of the school-related factors, such as class size (NEC, 1977; Heyneman, 1980: 155).However, other research has detected greater effects of other school and teacherattributes, although the evidence appears inconsistent. For example, Carnoy(1971, cited in Heyneman, 1980: 151) found a positive relationship betweenteacher’s experience and student’s achievement at the primary level in PuertoRico, while Heyneman (1976) found no significant relationship between the twovariables for Ugandan students. In addition, a teacher’s cognitive ability wasfound to have moderate but significant impact on students’ learning in Uganda(Heyneman, 1980: 152), but not in Thailand (Fuller and Chantavanich, 1976).The availability of books seems to be more consistently associated with thehigher achievement of students in developing countries (Husen et aL, 1978,Heyneman, 1980: 153-154).The effects of school-related variables seem to vary among students ofdifferent socioeconomic status. For example, the relationship between theavailability of textbooks and higher achievement appears to be stronger amongstudents from low socioeconomic backgrounds in Thailand and Chile (NEC,1977; Schiefelbein and Farrell, 1977, cited in Heyneman, 1980: 153). Thisfinding is reminiscent of the results obtained by the Coleman report about theimpact of availability of science laboratories on achievement of ethnic minoritystudents in the United States (Coleman etaL, 1966: 22).Findings from research in some developing countries raise a possibility thatan equal distribution of school resources may be able to reduce disparities inachievement of students from different socioeconomic backgrounds, owing tothree main reasons, 1) school resources are still scarce, 2) a high degree ofvariance in resource availability between schools, and 3) the curriculum of33schools may be too recent for the family or community to prepare the children for.However, a larger number of case studies are needed to estimate the effect ofschool and teacher attributes in different condWons.Results of research in ThailandAs elsewhere, research on the relative effects of school-related variables inThailand has obtained somewhat mixed results. The effects of schools andteachers’ characteristics on students’ achievement appear to vary amonggeographic regions, subjects, and types of residence.The overall impression about the effects of school and teacher attributes inThailand appears rather pessimistic: two nation-wide studies conducted in 1973and 1979-1980 (NEC, 1977: 63-65; Chantavanich eta!., 1990: 88-89, 180)indicated that despite the existing disparities in educational resources betweenschools, programmes to equalize these resources would not significantly reducethe differences in achievement among students from different socioeconomicbackgrounds. On the other hand, Leonor (1982: 120-121, table) using a differentstatistical method (canonical correlation) to reanalyse data from the 1973 survey,found a slight increase in the effects of teachers’ scores in teaching methods. Inanother reanalysis of the 1973 survey, Setapanich (1982: 127, table) found thatschool-related variables had a relatively stronger impact than students’ SES onstudent achievement in Thai language and mathematics, in all regions. Most ofthese studies also suffer from a small sample size and an under-representationof certain socioeconomic groups (Setapanich’s data came from a sample of twostudents from each school, and students from low-educated parents in mostregions were under-represented, see Setapanich, 1982: 53, 54), or limited to thenational level only (Chantavanich et aL, 1990). Moreover, all of these studieswere conducted prior to changes in curriculum content and administrative34systems, as well as to implementation of programmes to redistribute schoolresources during the Educational Reform (see Chapter 1).The impact of specific school variables remains ambiguous and inconsistent.Thai research on the effects of availability of school textbooks did not yieldconsistent results for this school variable (as did research conducted indeveloping countries). Lockheed etal. (1986) found that teachers’ regularity inusing textbooks had a small but significant impact on mathematics achievementof lower secondary school students. On the other hand, NEC (1977) did not findthat availability of textbooks had a strong and statistically significant impact onprimary school student achievement, which may be partly due to a small degreeof variance of this variable.The roles of modern (Western-styled) schooling in a rural village mayindirectly influence students’ academic achievement; however, very few studieshave examined the relationships between a school and the community where it islocated, and how this may affect the average achievement of students in thatcommunity. Chantavanich et a!. (1990: 53, table) found that the attitudes ofcommunity residents toward school seemed to directly exert a small butsignificant effect on students’ achievement. An anthropological study of fourvillages in the Central region and the Northeast found that, regardless of school’simpact on students’ achievement in academic subjects, schools seemed to havea significant impact on students’ social manners (Chantavanich etai, 1990: 155).A study of the role of teachers in a Northeastern village (Gurevich,1 972: 227-230)noted a lack of participation in school activities among most villagers. However,the study was limited to one village and its main purpose was to investigate theleadership roles of teachers, and thus it could not explore how school-communityrelations influence students’ achievement.35Some previous studies found that the effects of school and teacher qualityseemed to vary among different regions, groups of students, and types ofresidence (rural vs. urban). Setapanich (1982: 156-158) found that in Bangkokthe impact of school-related variables was greater for achievement inmathematics than in language; while in the other regions, school-relatedvariables had greater influence on student achievement in language than inmathematics. Setapanich postulated that schools may help to improve the verbalachievement of students whose mother tongue is not Central Thai.Raudenbush et a!. (1991: 268) observed that the effect of pre-primary schoolattendance on student achievement in Thai language was more pronounced forhigh- than for low-SES students, and for the rural areas only. They contendedthat the pre-primary schools that high-SES students in the rural areas attendedmay be more effective in teaching the Central Thai dialect, and thus preparingthese students for primary schooling. No data were available on the quality ofpre-primary schools, however. On the other hand, low-SES students in the ruralareas who do not use the Central Thai dialect in their daily life and did not attendpre-primary school may have greater difficulty in learning the Central Thai dialectat school (Raudenbush et aL, 1991). A lack of pre-schooling effect onachievement in mathematics found in their study may be because learningmathematics do not require as much verbal communication as learning theCentral Thai dialect.Lockheed and Longford (1989) asserted that the effects of lower secondaryschools on student achievement in mathematics were much more uniform thanprevious research in developing countries would have suggested. However, theirassertion may not apply to the primary level of schooling because: 1) thebackgrounds of primary school students may be more varied than those ofsecondary school students due to the very low rate of continuation to the36secondary level (about 30 per cent for the national average, and about 10 percent for the rural areas), and 2) the administrative agencies in charge of primaryschools are much more varied, which may result in a larger range of quality.The impact of school and teacher attributes on primary school students inThailand is highly related to differences in socioeconomic and linguisticbackgrounds among regions and areas of residence of the country. A number ofquestions arise with regard to the influence of school and teacher attributes onstudents’ achievement in Thailand:1. How well do the school and teacher characteristics predict students’scores?2. Do the effects vary among regions?3. Do the effects vary among different school subjects?Research on factors affecting academic achievement in the developingcountries has raised questions about the generality of findings from research inthe industrialized countries. The present research is in part an effort to test theapplicability of models derived from American studies to situations in Third-Worldcountries such as Thailand.37Chapter 3Research MethodologyThe data for this study were collected by the National Education Commissionof Thailand (NEC), in cooperation with other government agencies, as part of amulti-purpose assessment of primary school quality in 1987-1988 (ProjectBRIDGES - Basic Research and Implementation in DevelopinG EducationSystems) (NEC,1 990).Sample PopulationThe population under consideration in this study consists of all primaryschools, their staff, and their students, situated in 72 provinces or 13 educationalregions throughout Thailand. In 1987, there were 34,098 primary schools underthe authority of different administrations, mainly the Office of the National PrimaryEducation Commission (ONPEC), the Private Education Office, some municipalgovernments, and the Bangkok Metropolitan authority (NEC, 1 992b).Schools were selected for participation in Project BRIDGES by a multi-stagestratified random sampling method. The first step involved the random selectionof 25 per cent of the provinces in each educational region (with a minimum of oneprovince in each region). Seventeen provinces and the Bangkok Metropolitanarea (as the thirteenth educational region) were chosen in this way. Next, ineach selected province, 20 per cent of all districts were randomly selected. Forthe Bangkok metropolitan area, 25 per cent of the 13 sub-administrative districtswere selected. (See Appendix A for the names of educational regions, provincesand districts where the samples were selected.) Within each administrativedistrict, a simple random sample of schools was selected, whereby about 20 percent of each type of school (or at least one school of each type) in that districtwas chosen. From each school, data were collected from the school principal, all38Grade-six teachers and one teacher from each of the five other grades (chosenthrough simple random sampling), and all sixth-grade students who participatedin the national standardized tests in 1987 (organized by the Ministry of Educationand the ONPEC). Twenty per cent of the parents of those students in the samplewere also selected by random sampling to participate in the survey.Data collectionA pilot study was conducted between 24-25 December, 1987, in order to testand improve the questionnaires before putting them into use in 1988. The pilotstudy was done at three schools in Mg Thong province (in the Central region),through 4 different questionnaires administered to 3 school principals, 36teachers, 110 students, and 24 parents of students. All four types ofquestionnaires were then redesigned based on the data collected in the pilotstudy.Data collection for the main study took place between 8 February and 15March, 1988. Representatives from the four educational administrativeauthorities, namely the ONPEC, the Private Education Office, the municipaleducational offices, and the Bangkok Metropolitan educational office, wereinvited to attend a meeting where the objectives and procedures of the researchwere presented. In the Bangkok metropolitan area, the researchers from theNEC were in charge of the distribution and collection of questionnaires from thesample schools. Government officers in charge of education at the provincialand district levels implemented the survey in the rest of the country. The data forthe present study were drawn from 415 schools, 415 principals, 3,808 teachers,and 9,768 students.39Research instrumentsIn the BRIDGES Project, four types of questionnaires were administered tofour groups of people, namely school principals, teachers, students, and parents.(For this study, the information collected from the parents is not available. SeeNEC [1992] for more details.) Standardized achievement tests in live subjectblocks were conducted with Grade-6 students in 1987.The questionnaires for school principals asked about their personalbackground, such as sex, age, marital status, religion, educational qualifications,experience as school principal in general and in the particular sample school,attendance at academic-related training programmes in the past three years, aswell as their working conditions, such as the proportion of time school principalsspent per week on teaching and administration, in-service training for schoolpersonnel, activities in human resource development, provision of services forthe students, and activities within the local community. The questionnaires forschool principals also asked about general conditions of the school andcommunity, such as the size of school, location of the school, school equipmentand facilities, teaching facilities, and the access to community infrastructure.The questionnaires for teachers were about their personal background,including age, income, educational qualifications, in-service training over the pastthree years, as well as about their working conditions, including the proportion oftime teachers spent per week on preparation of lessons, counselling, andremedial lessons, about the availability and utility of teaching facilities, in-servicetraining programmes held inside or outside school, and allocation of time tovarious activities per teaching hour.The questionnaires for Grade-6 students covered their backgroundcharacteristics, such as dialect, education and occupation of parents, number ofyears students attended pre-primary schools, record of repetitions in previous40grade(s), previous grade average (Grade 5), record of absence from class(es) forthe term when the questionnaire was administered, doing homework after school,assistance with homework from other family member(s) or friend(s), parentalencouragement (in reading books), students’ assisting with parents’ work, amountand regularity of students having pocket money to take to school, sufficiency offood, report of sufficiency of textbooks and exercise books at home.The achievement tests were conducted by the ONPEC and the Ministry ofEducation (MOE) with Grade-6 students in 1987, in five subject areas, namely,1. the Thai language, including listening, reading comprehension, grammar(language usage), and writing (spelling),2. mathematics, including concepts, problem solving, and applications ofmathematic skills,3. life experience, including understanding health and how to achieve goodhealth; understanding the importance of the nation, religion, and monarchy;understanding the existing pattern of democracy in the country(constitutional monarchy); understanding the economy and different kindsof occupations; understanding science and technology, and (basic)scientific skills,4. work experience, including basic skills for work,5. character development, including self-discipline, how to live and work withother people, motivation for work, certain desirable habits such asdiligence, frugality, honesty, tolerance, loyalty to the nation-religionmonarchy, and a sense of appreciation for the national art and culture.The achievement tests have been tested for their reliability (Cronbach’s alphais 0.89, indicating that the tests results are consistent with each other) (Personalcommunication with Bhumirat, 1992).41The data analyzed in this study were supplied by Dr. Chinnapat Bhumirat, ofthe National Educational Commission of Thailand (with the permission of theSecretary General of the NEC). A selection of variables aggregated to the schoollevel was provided in the form of an SPSS/PC file. Because this data set wascompiled for another study (NEC, 1990a, 1990b, 1992), some of the relevantvariables collected are not available in the data set.Research Variables:The research questions first focus on how the two major sets of variables,namely, students’ background, aggregated at the school level, and schoolconditions, may influence average students’ achievement, and second, on howthe effects of these two sets of variables may vary among regions with differenteconomic and cultural conditions. To answer those questions, a number ofvariables (Table 3.1) have been selected from the available data to representaggregate students’ backgrounds, school conditions, and students’ achievement.Table 3.1 Variables selected for this studyIndependent VariablesVariables related to Students’ Backgrounds:Aggregate students’ SESProportion of students speaking Central ThaiAverage absenteeism rateProportion of students receiving assistance with homeworkVariables related to School-Conditions:School size (number of classrooms)Percentage of teachers with B.A. degrees, in 1987Average student-teacher ratios, in 1987Number of available teaching aid categoriesProportion of teachers’ time spent teaching (% per week)Proportion of teachers’ time spent checking homework (% per week)Overlapping variables:Average number of textbooks per studentAverage number of years students spent in pre-primary schools42Table 3.1 (cont’d) Variables selected for this studyDependent VariablesAverage Students’ Achievement:Thai languageMathematicsLife experience (general knowledge)Work experienceCharacter developmentThese variables are described and discussed below.1. Variables representing aggregate students’ backgroundPrevious research in industrialized and developing countries has obtainedopposing results about the effect of students’ background on achievement (seeBridge et a!., 1979: 213-227; Heyneman and Loxley, 1983). In addition, studiesin developing countries (Lockheed eta!., 1989; Moock and Leslie, 1986; Hess eta!., 1980) have suggested that non-material variables may better reflect students’backgrounds and that they operate independent of SES. The following variables(Table 3.1) are selected to represent both the material and non-material aspectsof aggregate students’ backgrounds, that have been observed, or suggested byprevious research, to contribute to students’ achievement.a. Aggregate students’ SESUse of students’ socioeconomic status (SES) as an independent variable isproblematic. The BRIDGES project data on aggregate students’ SES, compiledfrom information on parents education and occupation, and the amount of pocketmoney the students took to school, are questionable because they showed thatthe average SES of the Northeast was higher than that of the BangkokMetropolis, which according to the extant evidence (Ikemoto, 1991: 60, table;Knodel et a!., 1987: 47, table) is unlikely. Use of other variables in the BRIDGES43database as a proxy measure of aggregate students’ SES is likely too subjectiveto be reliable; for example, asking the students about their nutritional status;whether they have had enough food or not. Therefore, in this study aggregatestudents’ SES is represented by the only available alternative data, the grossprovincial product (GPP) per capita in 1987, as reported to the NationalEconomic and Social Development Board (NESDB) of Thailand (NSO, 1991:364-367, table). Although the NESDB data on GPP are collected at theprovincial level - not the school level, and thus do not show the variations withineach province, this measure of aggregate students’ SES appears to be a moreobjective, and reliable measure than the BRIDGES SES data, and will suffice.For the Bangkok Metropolitan region, average household income for the threemajor zones (i.e., core, suburbs, and fringe areas), obtained from the Nationalcensus (NSO, 1986), will be used because the single GPP datum for Bangkokdoes not allow calculation of variance, which is essential for the regressionanalyses.Since both GPP per capita and average household income are being used asproxy measures for the aggregate socioeconomic status of the students’, thesevariables will be referred to by the single name “aggregate students’ SES” in thefollowing discussions. It should be noted that GPP per capita in fact reflects thelevel of economic development in a province rather than the average level ofSES of students in a school, therefore, “aggregate students’ SES” can only beused for comparisons at the province or region levels and will not provideinformation about individual schools or students within the province.b. Proportion of students speaking Central ThaiThe proportion of students in a school who speak the Central Thai dialect athome reflects the compatibility between language used in school (i.e., Central44Thai dialect) and at home. Schools with a large proportion of students whospeak Central Thai may have advantages in that students in those schools canlearn subject material and communicate with teachers more easily, thus resultingin higher achievement levels.c. Average absenteeism rateThe average absenteeism rate may reflect the overall degree of poverty ofstudents in a school, lack of students’ motivation for studies, or even schoolpolicies on students’ attendance. Hence, average absenteeism rates (measuredat the school level) may show a correlation with averaged achievement scores forschools.d. Proportion of students receiving assistance with homeworkStudents in a school who receive assistance from their family members orfriends with school work may benefit from improved understanding and academicachievement. The proportion of students in a school who report that they receiveassistance from their family members or friends with school work may thus becorrelated with averaged achievement scores for their schools.Other measures of aggregate students’ background such as housingconditions, students’ opinions about their ability, or reports of how studentsactually spend time, may better predict their academic achievement. However,these measures are not available in the data set. In addition, personalcharacteristics of the students, such as sex, age, or tested ability, will not beexamined, since this study compares the influence of students’ contextualbackgrounds with the influence of school resources.452. Variables representing school conditionsWhile research conducted in industrialized countries is generally pessimisticabout the effectiveness of school resources in enhancing students’ achievement(Coleman et aL, 1966; Jencks et aL, 1972; Bridge et al., 1979; Peaker, 1971),more recent studies in developing countries suggest that certain material aspectsof schools (i.e., school conditions and personnel), which are largely controlled bythe central government, may be more influential than originally expected(Heyneman, 1986; Fuller, 1987; Fuller and Heyneman, 1989).a. School sizeSchool size is measured by the number of classrooms. Research indeveloping countries has indicated that school size has a positive effect onstudents’ achievement scores (Fuller, 1985: 26; Chantavanich et aL, 1990: 53,table). This may be because larger schools are better equipped and thus providea better opportunity to learn.School size may also correlate with or act as a proxy for other schoolvariables relating to the administration and management of school resources,and the provision of a learning conducive atmosphere, that are not directlymeasurable or present in the data set.b. Percentage of teachers with B.A. degrees, in 1987The educational qualifications of the teachers may reveal their levels ofknowledge and skill, which in turn may affect the students’ achievement.c. Average student-teacher ratios, in 1987Average student-teacher ratios may reflect the teachers’ work load and abilityto pay attention to each student, and thus influence the students’ achievement.46d. Number of available teaching aid categoriesTeaching aids (e.g., teachers’ instruction manuals, textbooks andsupplementary reading materials, chalk, paper, posters, maps, scientificinstruments, kitchen utensils, carpentry tools, agricultural tools, etc.) may assistteachers’ work, and thus affect students’ achievement.e. Proportion of teachers’ time spent teaching (% per week)The proportion of time per week that teachers allocate to teaching reflects theamount of contact students have with their teachers.f. Proportion of teachers’ time spent checking homework (% per week)A similar variable to number 5, the proportion of time per week that teachersallocate to checking homework may reflect the amount of attention that teacherspay to students’ work.Additional school related variables such as the teachers’ subjectspecialization, teaching styles, degrees of motivation and responsibility, cognitiveability, and the students’ peer group conditions, may have a significant impact onthe students’ achievement. The present data set does not provide information onthese variables.3. ‘Overlapping’ VariablesThe average number of textbooks and average number of years that studentsspent in pre-primary schools (both aggregate school level measures), can beconsidered as components either of the aggregate students’ backgrounds or ofschool resources, and are thus called overlapping variables. On the one hand,most studies consider them school-related variables (Jamison et aL, 1981;47Heyneman etaL, 1984; Raudenbush etaL, 1991), largely because educationaladministrators can reduce inequalities in school resources by providing freetextbooks and I or pre-primary school programmes to students from low incomefamilies, or who live in less developed communities. On the other hand, thesevariables may reflect the aggregate students’ backgrounds, as some parents maybe more able than others to provide students with preschooling or additionaltextbooks and exercise books. This dichotomy must be considered in discussionof these variables.4. Variables representing average students’ achievementIn this study average student achievement (hereafter called students’achievement) is represented by five dependent variables. Previous researchsuggests that the influence of aggregate students’ background and schoolresources may vary among different school subjects (Coleman,1 975; Kostakis,1987). Students’ achievement is represented by the average standardizedscores that Grade-6 students in each school obtained on tests in five differentsubject areas, specifically: Thai language, Mathematics, Life experience (orgeneral knowledge), Work experience, and Character development.Units of AnalysisPrevious research has differed in the units of analysis used to representstudent achievement, students’ background, and school characteristics. Theydepend on the researchers’ objectives and interests, as well as on the availabilityof data. Some researchers have used aggregate data for both dependent andindependent variables (Armor, 1972; Hanushek, 1972; Bidwell and Kasarda,1975; Heyneman, 1977). The larger proportion of researchers have usedindividual-level data to represent student achievement and background, while48data on schools are aggregated and averaged or are about the overallcharacteristics of school itself (Coleman et aL, 1966; Heyneman and Loxley,1983a and 1983b; see Heyns, 1986: 311-312). This imbalance in units ofanalysis has been mentioned as a factor contributing to the small effect ofschool-related variables relative to student background found in previousresearch (Coleman,1975: 377; Heyneman and Loxley, 1983b: 1172; Heyns,1986) (see Chapter 2).In this study, only aggregate data (averages, proportions, standardizedscores) are available for students’ achievement, background, and schoolcharacteristics. The dependent variables, achievement scores, are averagedfrom the individual scores of students in each school and then standardized.Independent variables related to aggregate students’ background are averaged(e.g., average absenteeism rate), are only proportions of students in the schoolwho fall into certain categories (e.g., speaking Central Thai, receiving assistancewith homework), or are averaged across the province where the school is located(e.g., the average income of residents estimated as gross provincial product percapita). School-related variables represent resources available to all students inthe schools; they do not account for differential access among the students tothose resources. In short, these aggregate measures do not show thecharacteristics of each individual student in a school.Using aggregate data as in the present study means that while the resultspertain to the students as a group (school in this case), they do not necessarilyprovide information about individual students. Robinson (1950) has describedthe problem of ‘ecological fallacy’ which is an error that occurs when using ‘grouplevel’ data to make inferences about relationships between variables at theindividual level (see also Goodman, 1959; Duncan et aL, 1961; Blalock, 1964: 97;Dogan and Rokkan, 1969; Selltiz eta!., 1976: 439-440; Kidder and Judd, 1986:49318-319, 371-372; Lam and Quattrochi, 1992: 89-90). For example, a highposWve correlation between the percentage of immigrants and the percentage ofilliterate population in an area does not mean that immigrants are illiterate, i.e.,the correlation between the two variables at the indMdual level may be low oreven negative. In addition, aggregate data obscure the variations amongindividuals within the schools when information about the distribution of individualattributes within the school is not available, i.e., when we do not know how widelydispersed a particular sample is from the central tendency or averages. Finally,the correlations between variables obtained from analyses using aggregate datatend to be stronger than those obtained from individual-level data (Robinson,1950:356; Bridge etaL, 1979:90-92; see Hannan, 1971: 489-490).Due to the aggregate nature of the BRIDGES data, the research questionsoutlined in Chapter 2 must be modified, as follows:1. How does aggregate students’ SES affect the average standardizedscores of students in the school?2. How do the aggregate non-material characteristics of students in a schoolaffect the average scores of students in that school?3. How do school characteristics affect the average scores of students in thatschool?4. Do the effects of either aggregate students’ background or schoolcharacteristics on average scores vary by region?5. Do the effects of either aggregate students’ background or schoolcharacteristics on average scores differ from subject to subject?There are two points that need to be emphasized before discussing theresults in the next chapter. First, the relationships between variables in this studywill pertain to the school level (and in the case of aggregate students’ SES, to the50provincial level) only, because relationships between variables tend to be specificto a particular scale or unit of analysis (Lam and Quat[rochi, 1992: 89-90; Harvey,1968: 71 -72), and so far it has not been possible to predict a priori which sets ofrelationships will be influenced by the change in units of analysis of variables(Fotheringham and Rogerson, 1993: 6; Blalock, 1964, cited in Hannan, 1971:491). Second, it should be noted that the differences in levels of aggregationbetween aggregate students’ SES and other variables make it impossible toreliably compare the effects of aggregate students’ SES with other independentvariables.In spite of the limitations inherent in using aggregate data, the results of thisstudy will still be useful in providing information on the relative impact ofaggregate students’ background and school conditions on average achievementscores of students, at the school level. Although Armor (1972) has observedsimilarities in the relative strengths of the effects of students’ background versusschool conditions on achievement, whether the units of analysis are individual oraggregate; the lack of individual-level data in the present study does not allowexamination of such similarities. However, the results of aggregate-levelanalyses may be useful for policy planners interested in increasing the overallachievement levels of students since the school is generally considered the basicunit for (re)allocation of resources in most government-run redistributionprogrammes.Analytical procedures used in this studyThe procedures used in this study were the analysis of variance (ANOVA) inorder to determine whether there are in fact regional differences in achievementto be explained, and multiple regression analyses in order to estimate and51compare the relative effects of the selected independent variables on students’achievement.First, the students’ achievement scores in five subjects was subjected toANOVA, to see if they vary significantly across regions or not. Regionalvariations in aggregate students’ background and school conditions were alsoexamined. Second, a preliminary ANOVA was done to identify the presence ofeffects on achievement from the selected independent variables. This was doneusing one representative variable from each of the three independent variablegroups, namely, aggregate students’ SES, school size, and average number oftextbooks available to each student in the school. These three variables havebeen suggested by previous research to affect students’ achievement (Colemanetal., 1966; Fuller, 1985; Heyneman, 1980).Since some of these independent variables may correlate with each other,leading to under- or over-estimation of other variables effects, ordinary leastsquares (OLS) regression analysis was conducted, whereby all variables areentered simultaneously into the regression equation. Standardized regressioncoefficients were used to determine the relative effects of aggregate students’background and school conditions on students’ achievement. The analyses weredone at both the national and regional levels, and for each of the five differentsubjects.Treatment of missing valuesIn this study, listwise treatment of missing values was used, whereby caseswith missing values for any of the variables were excluded from the analysis.Except for two school-related variables, teachers’ qualifications and studentteacher ratios, the number of missing values for each of the variables included inthe analysis is lower than 5 per cent. However, the missing values are52distributed randomly across the cases; therefore, of the 415 primary schoolssampled in the BRIDGES survey project, complete data for the analyses wereobtained from 342 cases (or 82 per cent of the total sample).53Chapter 4Results and Discussion4.1 The quality of dataAveraged students’ achievement scores in three academic subjects (Thai,mathematics, life experience) and two non-academic subjects (work experienceand character development), which are the dependent variables in this study,were first tested for reliability. Cronbach’s Alpha, which measures the reliabilityand consistency (see Bohrnstedt, 1970: 89-91) of the achievement scores was.911 (on a scale of 0 to 1), indicating that the values obtained from theachievement scores are highly consistent with one another. In other words, inschools where students’ averaged scores are high in one subject, they areusually high in the other subjects too. However, achievement scores in characterdevelopment have slightly lower correlations with scores obtained in the othersubjects.Due to the aggregate nature of the available data, the following discussionpertains to variations in achievement only at the school level, and thus cannotassess how aggregate students’ background and school conditions affectacademic achievement at the individual level.4.2 Regional variations in achievement scoresStudents’ achievement scores are the outcomes of various factors, principallyaggregate students’ background and school conditions. There has been adebate between research conducted in Western industrial countries and researchin developing countries on which of the two sets of factors, students’ backgroundor school conditions, has greater impact on students’ achievement (see Colemanet a!., 1966; Jencks et a!., 1972; Heyneman and Loxley, 1983). Rather thancomparing countries I begin by examining within country regional variation. The54variations in students’ achievement scores across the five regions of Thailand willbe discussed first.Table 4.1 shows the distribution of average achievement scores among thefive regions of Thailand. The differences in achievement scores betweenregions, as indicated by the F-ratios, are large and statistically significant (p <.05). Regional variation in achievement is higher for the three academic subjectsthan for the two non-academic subjects, as reflected in the higher F-ratios for theformer.Table 4.1: Average standardized achievement scores* in five different subjectsfor the five regions in Thailand +Subject Region F-ratioCentral North Northeast South BangkokThai -.00 .08 -.32 -.38 .38 20.69(.541) (.542) (.516) (.728) (.500) (<.000)Mathematics .03 .02 -.30 -.35 .37 13.76(.621) (.689) (.571) (.551) (.737) (<.000)Life experience -.03 -.07 -.25 -.29 .32 10.95(.479) (.637) (.584) (.606) (.555) (<.000)Work experience .00 -.03 -.07 -.30 -.01 3.28(.525) (.536) (.598) (480) (.374) (.012)Character de- -.04 -.08 -.10 -.10 .13 2.46velopment (.406) (.599) (.593) (.368) (.398) (.045)* Achievement scores are standardized, with a mean of 0 and standard deviation of 1.Standard deviationssignificance levels for the F-ratio+ The average scores reported for the Central and Northeast regions have excluded two outlierprovinces, Sing Bun and Loei, respectively, see Appendix B.The rank ordering of achievement scores is generally similar for academicand non-academic subjects. The differences in students’ achievement betweenBangkok and the rest of the country observed in previous research (NEC, 1977;Sudaprasert et aL, 1982; Chantavanich et a!., 1990) continue to be evident.Schools in the Bangkok Metropolitan region have the highest average scores,55except in work experience. Schools in the South and the Northeast have thelowest scores. Schools in the Central and North regions rank in the middle.4.3 Regional variations in aggregate students’ background and schoolconditionsRegional variation in achievement outcomes may be related to differences inthe social context of regions. The five regions in Thailand differ in bothaggregate students’ background and availability of school resources, and thesemay contribute to the variation in students’ achievement.As indicated by the F-ratios and their levels of statistical significance, theregional differences in aggregate students’ background and availability of schoolresource variables are large, except for one school condition variable, theproportion of time teachers spend teaching (Table 4.2).The rank ordering of aggregate students’ background variables is similar tothat of achievement scores, with the Bangkok Metropolitan region having higheraverage socioeconomic status (SES), and a large proportion of studentsspeaking Central Thai and receiving assistance with homework. It is followed bythe Central region. However, the average absenteeism rates of students in bothBangkok and Central regions are higher than in the other regions. In general, theaverage SES of Southern students is relatively high, but a very small proportionof them speak the Central Thai dialect, which is the language of instruction atschool. The Northeast has the lowest SES level and proportion of students whospeak Central Thai. The North has the lowest proportion of students whoreported receiving assistance with homework than the other regions, butotherwise ranks in the middle for other aspects of aggregate students’background.56Table 4.2: Average values for aggregate students background and schoolcondition variables in different regions of Thailand.Variable Region F-ratioCentral North Northeast South BangkokAggregate students’ background:GPP /capita in 1987 21,040 15,076 6,717 21,280 82,905 2,225.12(Baht) (8,581)* (2,797) (260) (9,786) ( ) (<.000)Students’ dialect .95 .35 .03 .19 .97 197.26(.167) (.432) (.071) (.288) (.031) (<.000)Average absenteeism .81 .63 .63 .72 .91 11.39rate (.292) (.328) (.344) (.351) (.198) (<.000)Students receMng .51 .46 .47 .49 .57 2.76homework help (.226) (.250) (.277) (.255) (.148) (.028)School conditions:School size 12.18 11.85 10.15 13.76 22.24 13.69(# of classrooms) (1 0.648) (1 0.392) (4.289) (1 2.181) (1 4.821) (<.000)Teachers with 60.10 57.45 44.33 51.90 57.14 6.30B.A. (%) (20.971) (22.072) (18.826) (21.046) (29.057) (<.000)Student-teacher 19.21 20.57 21.81 21.51 24.05 5.48ratios (5.793) (6.389) (6.135) (4.778) (5063) (<.000)Teaching aids 24.68 22.02 21.75 23.14 23.70 9.19(3.671) (3.825) (2.714) (2.584) (4.337) (<.000)Time teaching 48.65 50.56 50.74 49.41 49.08 1.15(% per week) (8.987) (7.021) (7.281) (8.213) (6.522) (.333)Time checking 12.54 10.41 9.28 10.93 14.80 23.50homework (%) (4.963) (3.017) (2.875) (3.476) (3.624) (<.000)Overlapping Variables:Textbooks 4.75 4.52 4.08 4.24 5.41 21.53(.940) (1.043) (1.057) (.721) (.661) (<.000)Pre-schooling .90 .73 .21 .77 .95 26.07(.556) (.563) (.407) (.536) (.560) (<.000)+ For a description of variables see Chapter 3.* Standard deviationssignificance levels of F-ratiosThere are also differences in school conditions among the five regions ofThailand. Schools in the Central region have a higher proportion of teachers withB.A. degrees, low student-teacher ratios, and more teaching aid categories thanschools in the other regions. Bangkok schools are larger than those in the otherregions; however, the student-teacher ratios are also the highest of the country.57In addition, there is a greater degree of variation in the proportion of teacherswith B.A. degrees among Bangkok schools. Northeastern schools are on averageof smaller size, with the lowest proportion of teachers with B.A., and fewerteaching aids available. Schools in the South and the North generally rank in themiddle. Teachers in the Bangkok Metropolitan and Central regions spend alarger proportion of time per week on checking homework than teachers in theother regions, while northeastern and northern teachers spend a slightly largerproportion of time on teaching in the classroom.The five regions also differ in availability of textbooks and pre-schooling, withstudents in the Bangkok Metropolitan and the Central regions having on averagemore textbooks available and having attended pre-schools for longer periods oftime than those in the other regions. The Northeast has the lowest averagenumber of textbooks per student, and a lower level of pre-school participation. Itis not known how much these regional variations in availability of textbooks andpre-schooling result from differences in aggregate students’ backgrounds or fromthe government’s policies on allocation of these resources. For this reason theyare considered “overlapping variables.”Clearly, regional differences are present in both the dependent variables(achievement scores) and independent variables. Because of these differences,the analysis of effects at the national level, while it may provide for comparisonswith other studies, is unlikely to provide much useful information on how schooland family conditions affect achievement scores within Thailand. From thestudents’ perspectives schools are a local phenomenon, thus the effects ofindependent variables in the students’ achievement will be felt at that level.584.4 Preliminary AnalysisA preliminary analysis was conducted with the national-level data in order tosee how achievement scores are distributed among different groups of students,and to identify the presence of effects on achievement from the selectedindependent variables. Students were grouped according to their socioeconomicbackground, the average number of textbooks available, and the size of schoolsthat they attend. The F-ratios and their level of significance show that thesethree variables all influence students’ achievement scores (Tables 4.3, 4.4 and4.5). Generally, the variations in achievement scores on the basis of SES groupsare smaller than the differences among students who attend schools of differentsizes and possess different numbers of textbooks. The differences inachievement are larger for academic than for non-academic subjects.Table 4.3: Distribution of achievement scores by provincial socioeconomiccategories (using GPP per capita as a proxy measure for SES).Subject SES Category of Provinces F ratioLow Middle HighThai -.26 .01 .17 15.43(.557)* (.783) (.568) (<.OOO)Math -.26 .06 .15 12.56(.566) (.832) (.702) (<.000)Life Experience -.27 .03 .15 12.36(.565) (.910) (.559) (<.000)Work Experience -.09 .09 -.02 2.36(.591) (.959) (.451) (.096)Character Develop- -.15 .08 .05 6.08ment (.531) (.780) (.396) (.003)* Standard deviationsA Levels of statistical significance for the F-ratiosThe general pattern in Table 4.3 is consistent with the hypothesis that as theSES of an area rises, the aggregate achievement scores of students in that area59will rise. On the three academic achievement measures this pattern isconsistent. For instance, the lowest SES group scores -.26 in Thai languageachievement, relative to scores at .01 and .17 respectively, for the middle andupper SES categories. This pattern is not, however, consistent for the non-academic subjects, and for work experience there is no systematic differencebetween SES groups.Table 4.4: Distribution of achievement scores by school size (as a proxymeasure of school resources)School SizeSubject Small Medium Large F ratiosThai -.25 -.13 .34 31.26(.659)* (.622) (.596) (<.000)Math -.11 -.18 .28 14.75(.672) (.701) (.758) (<.000)life Experience -.23 -.13 .33 23.09(.696) (.727) (.642) (<.000)Work Experience -.07 .01 .09 1.84(.750) (.7968) (.5608) (.160)Character Development -.04 -.13 .16 7.48(.580) (.718) (.475) (.001)* Standard deAationsLevels of statistical significance for the F-ratiosSchool size is considered a proxy measure of availability of school resources,which are hypothesized by research conducted in developing countries to have apositive impact on students’ achievement (see Fuller, 1985: 26, table). Theconsistent correlations between students’ achievement scores and school size asshown in Table 4.4 support this claim. For example, students in small schoolshave lower average scores in Thai language (-.25) than those from the mediumand large schools (-.13 and .34 respectively). The F-ratios also indicate that60school size has a stronger effect on academic subjects than on the non-academic ones.The availability of textbooks has a consistent influence on students’achievement scores for both academic and non-academic subjects (Table 4.5).In the group where the average number of textbooks per student is low, theaverage achievement scores are also low (-.26 for Thai language, in comparisonto -.08 and .27 in the other two groups.) This agrees with findings from researchin developing countries, that the quantity of textbooks has a more consistenteffect on students’ academic achievement than found in American research(Jamison et a!., 1981; Lockheed et al., 1986; Hanushek, 1986).These three variables may be correlated with one another and with othervariables associated with aggregate students’ background and school conditions.Moreover, it is not yet clear whether aggregate students’ background has more orless impact on academic achievement than school resources.Table 4.5: Distribution of achievement scores by number of textbooks availableNumber of Texts per StudentSubject Few Some Many F ratiosThai -.26 -.08 .27 24.96(.626)* (.603) (.683) (<.000)Math -.21 .08 .24 14.54(.618) (.675) (.820) (<.000)Life Experience -.26 -.12 .29 24.89(.627) (.617) (.793) (<.000)Work Experience -.08 -.14 .19 8.44(.629) (.522) (.879) (<.000)Character Development -.16 -.10 .25 19.83(.510) (.599) (.624) (<.000)* Standard deviationsA Levels of statistical significance for the F-ratios614.5 Comparisons of the relative effects of aggregate students’ backgroundand school conditions on students’ achievement in Thailand: results frommultiple regression analysesWhile research conducted in industrialized countries found that students’ SEShas a stronger impact on achievement scores than do school resources(Coleman et a!., 1966; Jencks et a!., 1972), recent research in developingcountries has obtained opposite results (Heyneman and Loxley, 1983; see Fuller,1985). However, due to the aggregate nature of data in this study, it is notpossible to directly test these propositions. The effects of aggregate students’background and school conditions observed here will apply to studentscollectively, but not necessarily to individual students. Multiple regressionanalyses were thus conducted to examine which of these claims apply inThailand.Table 4.6 lists the three strongest variables affecting achievement for eachsubject in each region, as indicated by standardized coefficients. For the nationallevel, all of the significant variables have been listed. Variables that havestatistically significant effects on students’ achievement scores are printed in boldletters. The criteria of statistical significance for this study is .05 or lower.The aggregate nature of data may have obscured the degree of collinearitybetween students’ background and school resources, which has been observedin previous research (Coleman, 1975). In this study, the variables selected donot appear strongly correlated with each other (see Appendix C: Correlationmatrices). Moreover, the effect of each variable is assessed when all othervariables in the model are controlled. Therefore, the regression coefficientsobtained reflect the independent effects of the selected variables.62Table 4.6: Rank orders of independent variables for each subject, nationally andwithin each region. (Bold entries are statistically significant.)Thai Math Ufe Work CharacterExperience Experience DevelopmentNationalDialect Dialect Dialect Dialect AbsenteeismSchool Size School Size Absenteeism GPP I Capita No. of TextsAbsenteeism Absenteeism School Size Absenteeism Pre-schoollngNo. of Texts Time Teaching No. of Texts Teaching Aids DialectTeachers with BA Time Teaching Time TeachingHelp in homeworkCentralGPP / Capita GPP I Capita GPP I Capita GPP I Capita No. of TextsSchool size Textbooks Pre-schoding Marking Homework GPP I CapitaAbsenteeism Pre-schoding School Size Time Teaching Pre-schoolingSchool Size School Size Help in homework Help in homework GPP I CapitaP re-schooling Students! Teacher School Size Absenteeism Teaching AidsGpp I Capita Help in homework Students [reacher Pre-schooling AbsenteeismNortheastSchool Size Time Teaching Time Teaching GPP I Capita Time TeachingNo. of Texts GPP I Capita GPP I Capita Markg Homework No. of TextsGPP I Capita Absenteeism School Size Time Teaching Mark’g HomeworkSouthSchool Size School Size Absenteeism Absenteeism AbsenteeismHelp in homework Absenteeism School Size School Size School SizeGPP I Capita Dialect Pre-schooling Dialect Pre-schoolingBanakokTeachers with BA School Size School Size School Size School SizeAbsenteeism Pre-schooling Absenteeism Dialect Pre-schoolingPre-schooling Marking Homework Dialect Help in homework Help in homeworkGeneral results from the national-level analyses show that the aggregate non-material measures of students’ background have slightly greater impact onachievement scores than school conditions, while aggregate students’ SES doesnot have a strong effect except on work experience scores. School conditionshave a greater effect on the three academic subjects than on the non-academicones. However, the national-level results do not apply to any single region,which may be due to regional differences in social, economic, and culturalconditions that influence how the two sets of independent variables affectstudents’ achievement scores. Again, it should be emphasized that the effects63of students background and school conditions variables observed here onlypertains to the school level (and for aggregate students’ SES, to the provinciallevel) (see Chapter 3).Based on the frequency of occurrences of significant effects in each region,i.e., the bold-faced entries, it appears that for both academic and non-academicsubjects, both the aggregate students’ background and the school-related groupsof variables are equitably effective. In other words, it is not possible todetermine that one group of variables is always more important than the other.Exceptions are the Central region, where the aggregate students’ SES has amuch stronger effect than all the other variables, and the Northeast whereschool -related variables are more influential.4.5.1 The influence of aggregate students’ background on achievementscores:Aggregate students’ socioeconomic status (SES)The first research question is whether the aggregate students’ SES(measured by GPP per capita and for Bangkok Metropolitan region, averagehousehold income) affects average achievement scores or not. The variations insize of beta coefficients (and level of significance) shown in Table 4.7 indicatethat this variable does not have a consistent effect across all regions of Thailand.The effect of aggregate students’ SES on achievement scores reported in Table4.3 becomes much weaker once other independent variables are controlled inthe multiple regression analyses.It should be emphasized that the measure of aggregate students’ SES used inthis study was collected at the provincial level, and does not show variation withinthe populations of the provinces, while other variables were collected at the64school level. This may have contributed to the lower effect of aggregatestudents’ SES compared to the other independent variables.Table 4.7: The effect (B coefficients) of aggregate students’ background variableson achievement scores in five subjects for primary school students, inThailand, 1987Variable SubjectThai Math Life Work CharacterExperience Experience DevelopmentNationallyGPP per capita (1987) .02 .06 -.02 -.19 -.07Dialect .28* .24* .24* P33* .15*Absenteeism rate ...19 -.20k -.21 * .19*Assistance with homework .07 .05 .13* .10 .06Central RegionGPP per capita (1987) .46* •55* .51* •44* .28Dialect -.19 .12 -.05 -.05 .15Absenteeism rate -.25 -.15 -.10 .10 -.20Assistance with homework -.01 .08 .11 .03 .22Northern RegionGPP per capita (1987) -.18 -.15 -.12 -.12 -.21Dialect .10 -.06 .06 .16 -.07Absenteeism rate -.12 -.09 -.15 -.17 -.13Assistance with homework -.01 .15 .29* .25 .09Northeastern RegionGPP per capita (1987) .19 •34* .25* 42* .15Dialect -.12 .12 .11 .07 .09Absenteeism rate -.15 -.18 -.16 -.08 -.21Assistance with homework -.02 -.06 .04 .10 -.08Southern RegionGPP per capita (1987) .18 -.03 .15 .07 .15Dialect .18 .18 .21 .35 .17Absenteeism rate -.12 39* ..38* .59*Assistance with homework .22 -.10 .13 -.02 -.05Bangkok Metropolitan RegionAverage household income (1986) -.22 -.14 -.12 -.17 .05Dialect -.10 -.04 -.20 -.19 -.02Absenteeism rate -.30 -.08 34* .02 .04Assistance with homework .03 -.10 .11 .19 -.27* statistically significant at the .05 level.At the national level, aggregate students’ SES does not have any significanteffect on achievement scores, except in work experience, where the effect isnegative. In the North and the South, aggregate students’ SES is among thethree strongest variables; however, its effect is not statistically significant. It has65a weak and negative effect in the Bangkok Metropolitan region. Aggregatestudents’ SES only has a significant effect in the Central and Northeasternregions. For the Central region, the beta coefficients of this variable aregenerally higher than all the other variables; moreover, its effect on mathematicsand life experience scores is about twice as great as the next strongest variables.In the Northeast, aggregate students’ SES has a significant effect onmathematics, life experience, and work experience. Its effect on work experienceis stronger than that of all the other variables.The strong influence of aggregate students’ SES in the Central region may bedue to the higher degree of variation (standard deviation) in SES compared toother aggregate students’ background variables. Moreover, the majority ofCentral region students speak the Central Thai dialect, so dialect is not likely adeterminant factor for students’ achievement there.Previous researchers (Heyneman and Loxley, 1983; Heyneman and Fuller,1989) have suggested that differences in student SES will have a greater impacton achievement within more industrially advanced countries than withindeveloping countries. Although they used individual-level data to represent SES,as opposed to school level data in this study, their suggestion appears consistentwith intra-regional results for Thailand, where the Central region, a region that isrelatively higher in both aggregate students’ SES and school resource levels thanare the other regions of the country, exhibits a greater effect of aggregatestudents’ SES on average achievement scores among its provinces thanobserved in the other regions.Results from the South raise some doubts about aggregate students’ SEShaving a strong effect in the richer regions. The average SES in the South issimilar to that of the Central region (Table 4.2); however, the GPP per capita ofone of the provinces representing the South is inflated as a result of industrial66activity. This province has low average achievement scores, suggesting that, atleast in this province, aggregate students’ SES does not have a significant effecton achievement. The degree of urbanization in the Central and the Southregions does not seem to be a factor since there is little difference in urbanizationbetween the two regions (Knodel et al., 1987: 47, table).The weak effect of aggregate students’ SES in the South may relate to ethnicconflicts between certain groups of the Muslim population and the centralgovernment. Previous research in the South observed resistance to government-controlled schools among the Muslim population (Dulyakasem, 1990; SathaAnand, 1987; see also Setapanich, 1982: 128-129). In the past Muslim parentstended to send their children to religious schools (Pondok schools) or abroad, tothe Middle East, rather than to government-run schools (Pitsuwan,1985: 196).Therefore, students’ ethnic I religious background may have greater influencethan their SES. At the same time, the Muslim populations tend to beeconomically worse off than other ethnic groups in that region (Satha-Anand,1987; Pitsuwan, 1985), and thus aggregate students’ SES may exert a joint effectwith this ethnic I religious factor. However, the lack of data on ethnicity (e.g., theproportion of students in a school belonging to certain ethnic groups) makes itimpossible to test whether aggregate students’ SES is correlated with their ethnicI religious background or not, and whether these factors may jointly affectachievement.The Northeast situation where aggregate students’ SES has a positivesignificant effect on achievement scores appears to be anomalous. The averageSES in the Northeast is much lower than that of the other regions, and has a lowstandard deviation, indicating that there is not much variation in SES among thenortheastern provinces (Table 4.2). One possible reason that aggregatestudents’ SES has a strong impact is because very few northeastern children67speak the Central Thai dialect, and thus dialect is not likely to be an importantintra-regional factor here. However, the effect of aggregate students’ SES is stillweaker than that of school conditions for all subjects except work experience.It is not possible to make a reliable comparison of the effect of aggregatestudents’ SES between the Bangkok Metropolis and the rest of the countrybecause data on aggregate students’ SES were drawn from different samplestrata in Bangkok than they were elsewhere. The measure of aggregatestudents’ SES for the Bangkok Metropolis is the average household income,collected from three major zones - or strata, while the measure of aggregatestudents’ SES in the other regions is gross provincial product per capita.The weak impact of aggregate students’ SES on average achievement scoresin the Bangkok Metropolitan region may mean either 1) that aggregate students’SES has a weaker effect than other variables in Bangkok, or 2) that the measureof aggregate students’ SES is not adequate to capture the actual variations inSES within the Bangkok population. Setapanich (1982) also observed the loweffect of SES in Bangkok, however, she postulated that it may be due to theinadequacy of the SES measure (father’s occupation) used in that study.In sum, the effect of aggregate students’ SES on achievement scores inThailand appears weak and not consistent across regions. The effect ofaggregate students’ SES seems to relate to the level of economic development,and cultural conditions in each specific region. However, the weak impact ofaggregate students’ SES may also be because this variable was collected at theprovincial level, thus obscuring the variations among residents within theprovinces.68Aggregate non-material measures of students’ backgroundThe second research question is, do the aggregate non-materialcharacteristics of students’ in a school affect the average achievement scores ofstudents in that school? Although previous research has suggested that the non-material measures of students’ background may have a greater impact onachievement than does students’ SES (Schiefelbein, 1979; Lockheed et al.,1989), a proper comparison of effects cannot be done, due to the imbalance inlevels of aggregation of the data.As shown by the very high beta coefficients of the three aggregate non-material measures of students’ background in the national analyses (Table 4.7),these aggregate non-material measures of students’ background do have a verystrong impact. In particular, the proportion of students speaking Central Thai in aschool has the strongest effect on all subjects, except character development forwhich the effect is small but significant. The average absenteeism rate also hasa significant and equivalent effect on all subjects. The proportion of studentsreceiving assistance with homework has a low but significant effect on students’scores in life experience.The influence of the proportion of students speaking Central Thai on averageachievement scores is stronger in the national-level analyses than in the within-region analyses, perhaps because the variations in dialects between regions aregreater than within regions. Since all school subjects are taught in Central Thai,schools that have a high proportion of students speaking this dialect will haveadvantages over schools where the majority of students speak other dialects.The proportion of students speaking Central Thai may also be an indirectmeasure of differences among ethnic groups within regions. In the South, it hasa positively strong effect on mathematics and work experience, while its effect ismuch weaker in the other regions. The relatively high variation of this variable69among southern schools may thus reflect differences among ethnic groups in thatregion. Although the southern population generally does not speak the CentralThai dialect, the differences in languages appear sharpest between the Muslimsin the four border provinces who speak Yawl (a Malay language), and otherethnic groups.The average absenteeism rate was found to have a significant negative effecton average achievement scores in the South and Bangkok Metropolitan regions.The average absenteeism rate may reflect the overall degree of poverty ofstudents in a school, lack of students’ motivation for studies, or even schoolpolicies on students’ attendance. It should be emphasized that the aggregatenature of the data on this variable means that the significant effect of averageabsenteeism rate observed in this study may not hold for individuals. It ispossible that at the individual level, students who are more often absent fromclassrooms may turn out to do better in the tests. In spite of this, it is still usefulfor policy planners interested in improving the average achievement level of aschool to pay attention to the overall attendance rates of students and find waysto improve it.In the South, the average absenteeism rate has the strongest effect of allvariables on achievement scores, while in the other regions its effect isconsistently not as strong. This may be because 1) absenteeism rate may reflectthe resistance to government-controlled schooling among the Muslims in theSouth, as suggested by the high variations in absenteeism rates among theschods there, or 2) the effect of presence I absence from school may be greaterwhen there are sharp ethnic differences, such as between the Muslim populationespecially in the four border provinces of the South and the Buddhist populationin the rest of the country. Schools may be the most important place whereinMuslim children are exposed to knowledge that may not be available otherwise.70Absence from school may have a greater impact on students’ academicachievement in the remote / periphery areas than in the ‘central’ areas becausechildren in the remote areas can only acquire ‘knowledge’ (required in thecurriculum and evaluated by the standardized tests) from school, while those inthe ‘central’ areas are probably more often exposed to ‘knowledge’ both insideand outside school. This may be why the Bangkok Metropolitan and Centralregions have the highest absenteeism rates and higher average achievementscores than the North and Northeast regions where absenteeism rates are low.The proportion of students receiving assistance with homework is found tohave a significant effect on life experience scores in the North. The lifeexperience curriculum includes general knowledge about such topics as thesociety, economy, basic science, and hygiene, that children in the remote areasmay only learn at school, whereas they can learn many other subjects both insideand outside of school. The North is a transitional region economically andculturally. Northern residents’ average SES is higher than their northeasterncounterparts, but lower than in other regions. Although the majority of northernresidents do not speak the Central Tha dialect, the proportion of those who do sois still larger than in the Northeast and the South. This may be why, of allaggregate students’ background variables selected in the study, the proportion ofstudents receiving assistance with homework has a stronger and more significantimpact on students’ achievement scores in that region.In sum, the national level analysis shows that the aggregate non-materialmeasures of students’ background have a strong effect on achievement scores.This suggests measures additional to SES to represent aggregate students’background for research in developing countries like Thailand. However, social,economic, and cultural conditions pertaining to each region result in variations inthe effect of these measures of aggregate students’ background. Aggregate71students’ SES has a strong impact in the more developed Central region and theleast developed Northeast region, while aggregate non-material measures ofstudents’ background have a rather strong influence on achievement in theSouth, where there are ethnic I religious differences from the rest of the country.4.5.2 The influence of school-related variables on students’ academicachievementThe third research question is how well do school and teacher characteristicspredict students’ achievement scores. Results from Thailand indicate that mostschool-related variables selected in this study have a much stronger effect onstudents’ achievement than found in previous research conducted inindustrialized countries (Table 4.8). The differences in results about the effect ofschool conditions may be due to the availability and distribution of schoolresources, and the role of teachers in imparting knowledge to students in adeveloping country like Thailand (or to the measurement problems associatedwith aggregate students’ SES).Both national and regional analyses show that most school-related variableshave a strong and significant effect on both academic and non-academicsubjects. The only exception is student-teacher ratios which does not have asignificant effect on any subject in any region.Generally, school size appears to be the most important school variable, Ithas a significant effect on academic subjects in all regions except the Centralregion. Its effect on the two non-academic subjects is generally small and notstatistically significant, although it has a significant effect on students’ workexperience scores in the Bangkok Metropolis, and on both non-academicsubjects in the South. At the national level, teachers’ qualifications has asignificant effect on students’ achievement in Thai; while the proportion of time72Table 4.8: The effect (B coefficients) of school resources on achievement ofprimary school students in Thailand, 1987Variable SubjectThai Math Life Work CharacterExperience Experience DevelopmentNationallySchool size .22* .21* .21* .07 .05Teachers with B.A. .15* .05 .05 .06 .06Student-Teacher ratios -.02 -.07 -.08 -.07 -.04Teaching aids available -.01 .08 .09 .11* .08% time teaching .07 .12* .14* .08 .13*% time checking homework -.01 -.01 .04 .04 .05Central RegionSchool size .31 .11 .25 .04 .07Teachers with B.A. .13 .02 .04 -.05 .16Student-Teacher ratios -.16 .04 .08 .06 -.13Teaching aids available .01 -.17 -.20 -.14 -.11% time teaching -.07 -.20 -.15 -.17 -.12% time checking homework .13 -.27 -.13 -.23 -.04Northern RegionSchool size 4Q* 31* .24 .12 .05Teachers with B.A. .06 .00 -.11 -.01 .05Student-Teacher ratios -.09 -.20 -.23 -.22 .03Teaching aids available -.02 .10 .18 .15 .16%timeteaching .17 -.00 .08 .07 -.02% time checking homework .03 -.13 -.04 .02 .04Northeastern RegionSchool size •43* -.01 .26* .13 .05Teachers with B.A. .11 -.04 .10 .04 .02Student-Teacher ratios .04 .02 -.02 .10 .06Teaching aids available .10 .07 .05 -.01 .03% time teaching .07 .42* •39* .18 .46*% time checking homework .01 .08 .13 .28* .22*Southern RegionSchool size .27 .40* 34* 43* .38*Teachers with B.A. .17 -.08 .01 -.06 .18Student-Teacher ratios .01 -.04 -.12 -.14 -.11Teaching aids available -.03 -.07 .12 .20 .08% time teaching -.16 .07 .06 -.08 .09% time checking homework -.08 .09 .05 -.15 .10Bangkok Metropolitan RegionSchool size .23 .48* 49* 55* .31Teachers with B.A. .36 .09 .00 .13 -.09Student-Teacher ratios .09 .01 .06 .08 -.07Teaching aids available -.00 .13 .10 .17 .22%timeteaching .01 .10 -.00 .17 .14%timecheckinghomework -.02 .17 .12 .11 -.08* statistically significant at the .05 level.73teachers allocate to teaching affects student& scores in mathematics, lifeexperience, and character development. Availability of teaching aids has asignificant effect on work experience scores.It is not clear how or why school size affects students’ achievement scores. Ingeneral, the correlations between school size and other school variables arerather low, but large schools tend to have more facilities available, and arelocated in cities. Thus, school size may be effective through pooling and moreefficient management of resources. Further, the strong effect of school size,which implies greater availability of educational resources, that is observed inThailand may be because school resources are more scarce here than inindustrialized countries (Fuller and Heyneman, 1989: 14). In general, students inThailand have fewer school resources available than those in industrializedcountries. A cross-nations comparison found that the budget per student forclassroom materials and other recurrent non-salary expenditures at the primarylevel in Thailand is 4 US dollars, while it is over 200 US dollars in the UnitedStates (Heyneman, 1984, cited in Fuller and Heyneman, 1989:15). To someextent this may reflect differences in the cost of living, but it may also reflect theoverall differences in the amount of school resources available between the twocountries.There may be a certain threshold limit to the effect school resources have onstudents’ achievement, whereby school resources do not exert any significanteffect after reaching a certain point (Mosteller and Moynihan, 1972; Schiefelbein,1979: 139-140; cf. Bridge et aL, 1979: 22-25). Generally, school resources havea consistent effect on Thai students’ achievement scores. The Central region isthe only region in Thailand where school-related variables do not have asignificant impact, which may be because the schools there are relatively large,74with more qualified teachers, and more teaching aids than schools in the otherregions.Regional analyses indicate that the proportion of time teachers spend onteaching and on checking homework has a very strong impact on students’achievement in the Northeast, but not in the other regions. The fact thatnortheastern students generally have low SES, do not speak the central Thaidialect, and have little assistance with homework from family members, mayindicate how important teachers are as the unique source of (Western)knowledge in the rural areas (see Saha,1 983). ‘Modern’ schools, introduced toThailand during the colonial period have replaced the traditional monasticschools, and have introduced kinds of knowledge that are new to most 4llagers.Students learn both spoken and written forms of the Central Thai dialect, as wellas subjects outside their daily life, from teachers educated in Western-styledcolleges and universities, instead of from the local Buddhist monks. Therefore,greater contact with teachers may help the students to perform better in school.The effect of teachers in enhancing students’ achievement may also beinfluenced by the social contexts of each region. Researchers in the Northeast(Keyes, 1991; Gurevich,1 972) observed that students usually pay high respect totheir teachers and do not question their authority or knowledge. However,research in the South found that Buddhist teachers feel alienated in the Muslimpondok schools while at the same time their Muslim counterparts feel that theyare discriminated against in obtaining teaching jobs (Dulyakasem, 1991: 147-148). The apparent lower degree of acceptance of schooling by communityresidents in the South may explain why the teacher-related variables do not haveas strong an impact on achievement scores there.The generally strong impact of school-related variables relative to that ofaggregate students’ SES observed in this study may be due partly to the poor75measurement of variables representing the latter (i.e., GPP per capita and forBangkok Metropolis, average household income). The higher level inaggregation of “aggregate students’ SES” may have attenuated the degree oftheir variation relative to that of school-related and other independent variables.The relatively strong effect of school-related variables should thus be cautioned.It appears though that this problem of obtaining reliable indicators of students’socioeconomic status for Thailand has remained consistent, as observed inprevious Thai research (Setapanich, 1982; Larpthananon and Wongkiattirat,1992) (see Chapter 2).In sum, school conditions in Thailand appear to have a stronger effect onaverage students’ achievement than found in previous American research. Thisshows that school conditions are very important, especially where aggregatestudents’ SES is low and few school resources are available. While researchconducted in industrialized countries (Coleman et al., 1966; Bridge et aL, 1979;Fuller, 1986: Appendix A) generally presents a pessimistic view of the effect ofschool-related variables on achievement of students from different backgrounds,the results from Thailand indicate that a more equitable distribution of schoolresources may help raise the achievement levels and thus reduce the inequalitiesin academic achievement, at least between schools.4.53 The effect of ‘overlapping’ variablesThe availability of textbooks and pre-schooling, are considered ‘overlapping’variables in this study because they seem to be indirect measures of aggregatestudents’ background, and at the same time have independent effects. Althoughstudents of higher SES backgrounds may have greater access to textbooks andpre-schooling than those from lower SES backgrounds, these two variables arealso subject to government intervention.76Nationally, the availability of textbooks has a significant effect on Thai, lifeexperience, and character development scores, while attendance at pre-schoolonly has a significant effect on character development (Table 4.9). Regionally,the number of textbooks has a significant effect on Thai and characterdevelopment scores in the Northeast, and on character development scores ofstudents in the Central region. It does not have any significant effect in theNorth, the South, and Bangkok Metropolis. Presumably there are enoughtextbooks available to students in the latter three regions. Pre-schooling has asignificant effect on character development in the South, but not on any othersubject or in any other region.Table 4.9: The effect (B coefficients) of overlapping variables on achievement ofprimary school students in Thailand, 1987Variable SubjectThai Math Life Work CharacterExperience Experience DevelopmentNationallyTextbooks .16* .10 .16* .10 .18*Pre-schooling .09 .08 .09 -.03 .15*Central RegionTextbooks .02 .24 .07 -.03 •35*Pre-schooling .07 .21 .27 .16 .22Northern RegionTextbooks .02 -.04 .00 .01 .03Pre-schooling .20 -.03 -.08 -.17 .09Northeastern RegionTextbooks .2 .13 .14 -.01 .31*Pre-schooling -.02 .10 -.02 -.16 .09Southern RegionTextbooks .05 .17 .12 -.05 -.10Pre-schooling .16 -.01 .26 .19 P34*Bangkok Meb-opolitan RegionTextbooks .05 -.02 .16 .10 .08Pre-schooling .25 .27 .11 -.14 .37* statistically significant at the .05 level.77The effect of textbooks may depend on ‘cultural’ differences among ethnicgroups. Having textbooks does not show a strong impact in the South, whichmay be partly due to the resistance of the Muslim populations to Thai schoolingand curriculum. An anthropologist working in the South in the 1970’s(Prachuabmoh, personal communication, 1993) was told by Muslim villagers thatpreviously parents did not allow their children to take (non Muslim) textbooksinside the house after they returned home from school. In that case, theavailability of textbooks may not translate into higher achievement.The strong effect of pre-schooling on character development scores in theSouth suggests that there is a greater tendency among southern students whoattended pre-primary schools to have adopted attitudes considered desirable inthe character development curriculum. Pre-primary schooling seems to have anindependent effect on achievement in character development; in other words,there is a low correlation between this variable and other aggregate students’background variables.4.6 Regional differences in variable effectsEmbedded in the previous three questions is the fourth research question,which is whether the effect of the two sets of variables differ among the fiveregions of Thailand or not. As the above discussion shows, specific economicand cultural conditions in each individual region result in differences in therelative effect of aggregate students’ background and school resources onachievement scores. This indicates that national-level results are not likely to beuseful when it comes to deciding which factors should be focused on in order toreduce the inequalities in achievement within regions.784.7 Comparisons of the relative effects of aggregate students’ backgroundand school conditions on academic- versus non-academic subjectsThe fifth research question is whether the effects of the two sets of variablesdiffer from subject to subject. Previous research has suggested that aggregatestudents’ background may have a greater effect on language-oriented subjects,while schooling may have a greater effect on mathematics or science subjects(Coleman, 1975: 382). In addition, it is postulated that in the Third Worldcountries, the content of school subjects is relatively new and thus parents maynot be able to help their children with school work, regardless of their familybackground (Saha, 1983: 85-86).In general, there does not seem to be much difference in factors affectingstudents’ achievement across the three academic subjects. Nationally,aggregate students’ background variables, especially the proportion of studentsspeaking Central Thai and average absenteeism rates have a consistently strongand significant effect. However, their effect is only slightly stronger than that ofthe school-related variables (see Tables 4.6, 4.7, and 4.8).There seems to be a clearer distinction between the two non-academicsubjects, wherein factors that have a significant impact on work experiencescores tend to have a weaker or no effect on character development scores.Achievement in work experience generally depends on the ability of students(students’ dialect) to understand their teachers’ instructions and the opportunityfor students to use the tools (absenteeism, availability of teaching aids). In turnstudents’ character development scores depend on both school- and non-schoolinputs, i.e., absenteeism, availability of textbooks, pre-schooling, and theproportion of time teachers spent teaching.One interesting observation is that SES has a consistent effect onachievement in Thai language. Aggregate students’ SES generally ranks among79the three strongest variables for achievement in Thai for all regions, exceptBangkok, although its effect is not usually statistically significant. Thisconsistency in the influence of SES, and the moderate correlation betweenaggregate students’ SES and the proportion of students speaking Central Thai(.559) suggests that high SES children may have a greater opportunity to learnthe Central Thai dialect than do low SES children, perhaps through better accessto telesion and the mass media.That there is not much difference in factors affecting achievement scores forthe three academic subjects may be due to the use of the Central Thai dialect inthe classroom. The consistently strong effect of dialect on both subjects at thenational level (see Table 4.7) indicates that students whose mother tongue is notCentral Thai invariably have difficuWes in learning, regardless of whether thesubject is language-oriented or not.School-related variables tend to exert greater influence on academic subjectsthan on the two non-academic subjects. This suggests that the contents ofacademic subjects are such that students’ access to resources and to teachers’(proportion of time spent teaching) have a strong impact on their achievementscores. However, the effect of school conditions on achievement scores in workexperience and character development is weaker than the effect of aggregatestudents’ background. This may be partly due to attempts by the Centralgovernment of Thailand to make the primary school curriculum more relevant tolocal conditions, during the Educational Reform of the late 1970’s (see Chapter1).Regional differences in variables affecting the two non-academic subjectssuggest that the government’s attempts to reduce inequalities in students’achievement may not be equally successful in all regions of the country. Schoolrelated variables have a significant effect on the two non-academic subjects in80the Northeast and the South. Perhaps, the contents of the two subjects are stillforeign to the people! students in the two regions, where few speak Central Thai.The finding that pre-schooling has a positive impact on Southern students’ scoresin character development suggests that there may be significant differencesbetween local conditions and school curriculum contents, whereby preparingstudents’ through pre-schooling programmes may raise the levels of students’academic achievement, at least in character development.In sum, there appears to be a distinction between academic- and nonacademic subjects in the factors affecting students’ achievement scores. Theuse of the Central Thai dialect as the sole language of instruction, and themodifications to the curriculum during the Educational Reform, may beresponsible for the consistent effect of aggregate students’ background on allthree academic subjects. At the same time, school conditions have a significantimpact on students’ achievement, especially in academic subjects.81Chapter 5Summary and ConclusionsThere is a long research tradition in the sociology of education, comparing therelative effects on students’ academic achievement of students’ background andschool conditions. Research conducted in the United States and other Westernindustrialized countries has consistently found that students’ background has afar greater impact than do school conditions (Coleman et aL, 1966; Jencks et a!.,1972; Rutter et a!., 1979), while the limited research conducted in developingcountries has obtained contrary results (Heyneman and Loxley, 1983; see Fuller,1985). As a result, opinion differs on the potential utility of using schoolresources for reducing differences in students’ academic achievement.This study has attempted to estimate which of the two claims apply to primaryeducation in Thailand. However, due to the aggregate nature of available data,the analyses in this study are limited to identifying the factors that affectdifferences in achievement scores at the school level. While it is possible thatvariations in achievement may be greater within schools, the between-schoolanalyses may still identify the school resources that can be adjusted in order toraise the level of achievement, because in most government-run redistributionprogrammes the school is the basic unit for (re)allocation of resources (NEC,1990).The research questions investigated in this study, their answers, and principalfindings are as follows:1. What is the effect of aggregate students’ SES on average achievementscores?82The effect of aggregate students’ SES (represented provincially by GPP percapita, and for the Bangkok Metropolis, by average household income) onaverage achievement scores in Thailand appears weaker and not consistentacross all regions, when compared to the results of previous American researchusing aggregate data (Armor, 1972). This may reflect differences betweencentralized and decentralized educational systems (Burstein et a!., 1980). Indecentralized systems (such as in the United States), the level of communitywealth has a significant impact on achievement as wealth enables schools toacquire facilities and teaching personnel, to set up student admission policies,and to allow curriculum differentiation. However, in Thailand the centralgovernment has control of the curriculum and of the allocation of schoolresources in all schools except private schools. In fact the Thai government hasattempted to reduce the inequalities in resources between schools, for example,by allocating more teachers to the poorer and more remote provinces (Ketudat,1984; Fry and Kaewdang, 1982). Therefore, aggregate students’ SES does notnecessarily reflect the availability of school resources.The interpretation of the effect of aggregate students’ SES in this study iscautious because 1) the data representing aggregate students’ SES are collectedat the provincial level, thus, variations within each province, or within schools, arenot evident, 2) the measure of aggregate students’ SES used may not capturethe actual variations in SES within the population of the Bangkok Metropolitanregion (see Setapanich, 1982), and 3) aggregate students’ SES may have a jointeffect with other aggregate students’ background variables in the South, such asethnicity, but no data are available to test this.2. Do the aggregate non-material characteristics of students in a school affect theaverage achievement scores of students in that school?83The aggregate non-material measures of students’ background (e.g. theproportion of students speaking the Central Thai dialect and rece.iing assistancewith homework, and the average absenteeism rate) have a strong effect in theSouth and the North. The influence of these aggregate non-material measuresmay partly reflect the differences in economic, social, and cultural conditionsamong regions. It is not possible to reliably assess the suggestion by previousresearch (Schiefelbein, 1979; Lockheed et a!., 1989) about non-materialmeasures being more relevant indicators of students’ background in a developingcountry, due to differences in units of analysis.3. Do the school and teacher characteristics predict students’ achievementscores?Of all selected variables, only school size has a consistently significant effectin all regions, except the Central region. This finding is very significant becauseprevious research in industrialized countries generally gives a pessimistic viewabout the potential of school resources, which can be manipulated by policymakers, to reduce the inequalities in achievement among students from differentbackgrounds (Jencks et al., 1972). That school size has a consistently strongeffect on achievement, net of other factors, may be because of the centralizedallocation and management of resources, which remain largely scarce in adeveloping country like Thailand (Fuller and Heyneman, 1989). However, it isnot known whether there is any threshold limit for school size. The absence of asignificant effect of school size in the Central region may be because mostschools in this region have more school resources and highly qualified teachersthan those in the other regions. In addition, it is not yet clear exactly what the84causal mechanism may be linking school size and students’ achievement scores.Larger schools, where scores are higher, do have more resources (NEC, 1 990a).4. Do the effects of aggregate students’ background (including aggregatestudents’ SES) and school-related variables vary among regions?The variations in aggregate students’ background, school conditions, andachievement scores among the five regions of Thailand are large and statisticallysignificant. Moreover, it is quite clear that the results from the national-levelanalyses differ from each regional-level analysis, and that there are noticeableregional differences in which of the variables have significant impacts onachievement. This indicates that results derived from national-level analysescannot be usefully applied to specific regions. Although the national-level resultsmay provide a comparison of the relative effects of aggregate students’background and school conditions on achievement scores across industrializedand developing countries (Heyneman and Loxley, 1983), they may haveincorrectly estimated the effect of different variables within the different regions ofthe country.5. Do the effects of aggregate students’ background (including aggregatestudents’ SES) and school-related variables vary by school subjects?For academic subjects, there is no difference in the effects of aggregatestudents’ background and school conditions on achievement scores. This ispresumably because these subjects, be they language related, or science based,are all taught in the same Central Thai dialect. The limited difference in factorsthat affect achievement scores is between the academic- and non-academic85subjects. For example, the effect of school-related variables is stronger forachievement in academic than in non-academic subjects. Students may learnacademic subjects only at school, while the subject matter of non-academicsubjects (work experience and character development) may be learned bothinside and outside school. However, some of the standardized tests, such as inlife experience, cover materials that can be learned both inside and outsideschool, making it difficult to assess which aspects of the subject are influenced byaggregate students’ background or by school conditions.Between regions there is little difference in the effect of aggregate student&background and school conditions on achievement scores. Exceptions are thataggregate students’ SES is more influential in the Central region, and school-related variables are more influential in the Northeast. This indicates that if onewants to find ways to improve the achievement levels of students, both aggregatestudents’ background and school conditions should be considered.Equality of educational opportunity in ThailandOne of the stated policies of the Thai government is to provide ‘equality ofeducational opportunity’ to every citizen regardless of sex, socioeconomicbackground, ethnic origins, or residence (Chantavanich et al., 1990: 15; NEC,1990). The meanings and indicators of equality of opportunity have evolved fromequality of school resources to the degree of effectiveness of school conditions inmoderating the effect of outside influences, especially those of student familybackground (see Coleman, 1990). However, in this study, due to a lack ofindividual-level data on student background, it was not possible to directlyestimate the effect of school resources relative to that of family SES. Instead, thedegree of equality of opportunity in Thai education was assessed indirectly by86looking at 1) the equality (availability) of school resources among differentregions, and 2) the effect of school resources, when aggregate measures ofstudents’ background are controlled for.The results show that the five regions of Thailand differ significantly inaverage achievement scores, especially in academic subjects. Moreover, thereis a high degree of inequality between the five regions in availability of schoolresources. School resources appear to promote higher average studentachievement, at least to the extent that the aggregate measures of students’background are held constant. Therefore, a more equitable redistribution ofschool resources may help reduce the regional differences in achievement.For policy planners, the results of this study have pointed out severalvariables that have a significant impact on average achievement scores ofstudents. The national policy that requires the use of Central Thai in schoolappears to have a very strong effect on students’ achievement scores, especiallyin the South. Facilitating the learning of students whose mother tongue is notCentral Thai, may be either by ensuring that teaching is done in the appropriatedialect (at least in the early grades), or by having students learn the Central Thaidialect very early (such as in pre-primary schools). Educational administratorsshould also consider how to improve the average attendance rate of students intheir schools since it appears to have a high negative impact on averageachievement scores. In addition, further investigation of the mechanismswhereby school size influences students’ achievement scores would be usefulsince this variable has the most consistent effect of all the variables examined inthis study.In sum, due to the lack of individual-level data in this study, it is not certainhow effective the Thai primary schools are in enhancing the achievement ofstudents from different backgrounds. However, the analyses in the present study87have indicated a high degree of inequality in school resources and a strong effectby some school-related variables on average achievement scores of students.These variables can be focussed on as a preliminary step toward bringing aboutequality of opportunity.Limitations of the data and analysesThe interpretation of findings from the present study is conditioned bylimitations in the data set. First, reliable measures of students’ background in adeveloping country like Thailand have been difficult to obtain, as also shown inprevious research (Setapanich, 1982; Larpthananon and Wongkiattirat, 1992).Although both economic and social I cultural aspects of aggregate students’background are considered in this study, more detailed measures arerecommended. It is not clear whether the effect of students’ dialect is only due tothe incompatibility of languages spoken at home and in school, or is compoundedby the differences in child rearing practices associated with certain ethnic groups.It appears that students’ ethnic background may be a significant factor, especiallyfor understanding the variations in achievement in the South, but no dataaddressing this problem are available in this study.Second, the data for this study are cross-sectional, and thus it is not possibleto assess how the effects on students’ achievement, of aggregate students’background and school conditions, may accumulate or change over time.Longitudinal data may be especially important in the case of students from ethnicminority groups, in order to understand how modern schooling has contributed totheir learning over the years.Third, academic achievement is the only school outcome of concern althoughit is possible that schools may influence other aspects of students’ lives, such astheir attitudes and social habits.88Fourth, the aggregate nature of data obscures the variations among studentswithin schools, and thus results observed in this study do not necessarily apply toindividual students. Moreover, student achievement is influenced by variousfactors, ranging from individual ability and effort to conditions in family,classroom, school, and community (Heyns, 1986); therefore, to reliably examinethe effects of students’ background and school resources, these two sets ofvariables should be measured at both individual and group levels (depending onresearchers’ objectives). However, multi-level research in sociology of educationhas not been conducted until very recently due to several constraints such as theavailability of data and simple to use statistical programmes (see Riddell, 1989;Heyneman, 1989).Finally, the analyses in this study are limited because it is not possible tocontrol for students’ initial ability or prior achievement. It may be the case thatschools are different from one another in terms of students’ ability, although thisis not likely. Additional variables in the BRIDGES dataset on students’ previousachievement, i.e., the proportion of students repeating in previous grades, andthe normalized scores of average GPA in Grade 5, appear to be too subjective tobe of use since each school has its own policies on grading and promotion ofstudents.Suggestions for future researchBoth quantitative and qualitative methods should be used to complementeach other in future research. Quantitative studies, such as this one, are usefulin pointing out which variables have a significant impact on students’achievement. However, they are unable to show the processes by which suchvariables exert influence on students’ achievement (see Mehan, 1992). Forexample, school size has been identified as having a more consistent effect than89all the other selected variables. However, it is not clear whether the effect is dueto the size of school itself, or to certain conditions associated with size.Qualitative studies can also be useful when independent variables do not havelinear relationships with achievement scores (e.g., class size), or when they aredifficult to quantify (e.g., interaction between teachers and students).Each region in Thailand differs economically, socially, and culturally;moreover, these regional variations appear to influence the relative effect ofaggregate students’ background and school conditions on students’ academicachievement. One line of future research that should be undertaken is toconduct longitudinal surveys, to investigate how the regional differences changeover time. Longitudinal data would also provide information to determine whetherthe effect of aggregate students’ SES will become stronger as a region developseconomically, which is happening in the Central region and in industrializedcountries. There are also regionally differing cultural factors to consider, such asthe resistance to government schooling among ethnic minority groups, whichmay intervene the effect of SES on achievement.There are a number of research topics, that could provide further informationon factors that contribute to Thai students’ academic achievement, such as therelevance of curricula to local conditions, the relationships between communityresidents and school teachers in different regions, and the changes in distributionpatterns of school resources across regions, to name but a few.ConclusionsBy examining the factors that affect differences in achievement scores at theschool level this study has attempted to estimate whether aggregate students’background or school conditions more strongly influence students’ achievement90during primary education in Thailand. To reiterate, the following observationshave been made:1. In Thailand, the effect of aggregate students’ SES on schools’ averageachievement scores is weaker and not as consistent as in the United States.2. In some regions of Thailand aggregate non-material measures of students’background have a strong effect, while in other regions aggregate students’SES has a strong and significant effect.3. Of all selected variables, only school size has a consistently significant effectin all regions, except one, the Central region.4. Inter-regional variations in aggregate students’ background, school conditions,and achievement scores are large and statistically significant. Moreover, it isquite clear that the results from the national-level analyses differ from eachregional-level analysis, and that there are noticeable regional differences inwhich of the variables has significant impact on achievement.5. There is little difference between regions in the effect of aggregate students’background and school conditions on achievement scores, therefore, theyboth should be considered if one wants to find ways to improve the averageachievement levels of students.These results present a convincing case that the standard conceptions aboutfactors that influence educational achievement in industrialized nations,particularly the U.S., do not necessarily apply in developing countries such asThailand. In turn, regional variations in the effects of aggregate students’background signify that one should not assume that models or results derivedfrom national-level analyses will apply to each region in the country. Differencesin economic, social, and cultural conditions, as well as availability and distributionof school resources between regions may influence how aggregate students’background and school conditions affect students’ academic achievement.91BibliographyAlexander, Karl L. and Martha A. Cook1982 “Curricula and Coursework: A Surprise Ending to a Familiar Story”American Sociological Review 47, (5), 636.Anderson, C. Arnold1967 “The international comparative study of achievement in mathematics”Comparative Education Review 11, 182-196.Armor, David J.1972 “School and Family Effects on Black and White Achievement: AReexamination of the USOE Data” in Frederick Mosteller and Daniel P.Moynihan (eds.) On Equality of Educational Opportunity. New York:Random House, 168-229.Avalos, Beatrice and Wadi Haddad1981 A Review of Teacher Effectiveness Research in Africa. India. LatinAmerica. Middle East. Malaysia. Philippines. and Thailand: Synthesis ofResults. Ottawa: IDRC.Bidwell, Charles E. and John D. Kasarda1975 “School District Organization and Student Achievement” AmericanSociological Review 40, 55-70.Blalock, Hubert M. Jr.1964 Causal Inferences in Nonexperimental Research. Chapel Hill: TheUniversity of North Carolina Press.Bohrnstedt, George W.1970 “Reliability and Validity Assessment in Attitude Measurement” in Gene F.Summers (ed.) Attitude Measurement. Chicago: Rand McNally &Company, 80-99.Bourdieu, Pierre and Jean-Claude Passeron1990 Reproduction in Education. Society and Culture. London: SagePublications.Bridge, R.Gary, Charles M. Judd, and Peter R. Moock1979 The Determinants of Educational Outcomes: The Impact of Families.Peers. Teachers, and Schools. Cambridge, Massachusetts: BallingerPublishing Company.Burstein, Leigh, Kathleen B. Fischer, and M. David Miller1980 “The Multilevel Effects of Background on Science Achievement: A CrossNational Comparison” Sociology of Education 53 (October), 215-225.92Chantavanich, Supang and Gerald W. Fry1985 “Thailand: System of Education” in International Encycloredia ofEducation. Oxford, UK: Pergamon Press, 5235-5240.Chantavanich, Amrung, Supang Chantavanich, and Gerald W.Fry1990 Evaluating Primary Education: Qualitative and Quantitative PolicyStudies in Thailand. Ottawa: IDRC.Chernichovsky, D. and 0. Meesook1985 School Enrollment in Indonesia. World Bank Staff Working PaperNo.746. Washington, D.C.: World Bank.Chiengkul, Witayakorn1983 “The Transformation of the Agrarian Structure of Central Thailand, 1960-1980” Journal of Contemporary Asia 13, (3), 340-360.Chowdhury, Sanjoy1989 “The economy” in Richard Tourret (ed.) Thailand. London: EuromoneyPublications in association with Bangkok Bank, Baring Securities, DusitThani Hotel Group, Lloyds Bank Fund Management, Merrill LynchCapital Markets, and Tilleke & Gibbins, R.O.P., 35-56.Cohen, Erik1991 Thai Sodety in Comparative Perspective. Collected Essays. Bangkok &Cheney: White Lotus.Coleman, James S.1968 “The concept of the equality of educational opportunity” Harvardeducational review (Winter), 7-22.1975 “Methods and Results in the lEA Studies of Effects of School onLearning” Review of Educational Research 45, 335-386.1990 Equality and Achievement in Education. Boulder, San Francisco, andLondon: Westview Press.Coleman, James S. et al.1966 Equality of Educational Opportunity. Washington, D.C.: GovernmentPrinting Office.Cummings, William K1977 “The effects of Japanese schools” in A. Ktoskowska and G. Martinotti(eds.) Education in a changing society. London: Sage Publications, mc,255-290.93Diller, Anthony1991 “What Makes Central Thai a National Language?” in Craig J. Reynolds(ed.) National Identity and Its Defenders - Thailand. 1939-1989. MonashPapers on Southeast Asia No. 25. Victoria, Australia: Aristoc Press Pty.Ltd, 87-1 32.Dogan, Mattei and Stein Rokkan (eds.)1969 Quantitative Ecological Analysis in the Social Sciences. Cambridge,Massachusetts and London, England: The M.I.T. Press.Dulyakasem, Uthai1991 “Education and Ethnic Nationalism: The Case of the Muslim-Malays inSouthern Thailand” in Charles F. Keyes (ed.), assisted by E. Jane Keyesand Nancy Donnelly Reshaping Local Worlds:Formal Education andCultural Change in Rural Southeast Asia. Monograph 36. New Haven,Connecticut: Yale University Southeast Asia Studies, 131-152.Duncan, Otis Dudley, Ray P. Cuzzort, and Beverly Duncan1961 Statistical Geography Problems in analyzing Areal Data. Illinois: TheFree Press of Glencoe.The Economist1991 “Thailand: North-east passage” 23 February-i March, 36.Eder, Donna1981 “Ability Grouping as a Self-Fulfilling Prophecy: A Micro-Analysis ofTeacher-Student Interaction” Sociology of Education 54 (July),151 -162.Foster, Philip1977 “Education and Social Differentiation in Less Developed Countries”Comparative Education Reiew 21 (June/October), 211-229.Fotheringham, A. Stewart and Peter A. Rogerson1993 “GIS and spatial analytical problems” International Journal ofGeographical Information Systems 7, (1), 3-19.Fry, Gerald1983 “Empirical Indicators of Educational Equity and Equality: A Thai CaseStudy” Social Indicators Research 12, 199-215.1992 “Thailand’s Political Economy: Change and Persistence” in Cal Clark andSteve Chan (eds.) The EvoMng Pacific Basin in the Global PoliticalEconomy - Domestic and International Linkages. Boulder and London:Lynne Rienner Publishers, 83-105.94Fuller, B.1986 “Raising School Quality in Developing Countries: What InvestmentsBoost Learning?” World Bank Discussion paper. Washington, D.C.:IBRD.1987 “What School Factors Raise Achievement in the Third World?” Review ofEducational Research 57(3), 255-292.Fuller, Bruce and Stephen P. Heyneman1989 “Third World School Quality: Current Collapse, Future Potential”Educational Researcher 18(2) (March), 12-20.Ganjanapan, Man1989 “Conflicts over the Deployment and Control of Labor in a Northern ThaiVillage” in Gillian Hart et al. (eds.) Agricultural Transformation: LocalProcesses and State in Southeast Asia. Berkeley: University ofCalifornia Press, 98-122.Goodman, Leo A.1959 “Some Alternatives to Ecological Correlation” American Journal ofSociology LXIV (6), 61 0-625.Grandstaff, Terry1986 “Thai Government Policy Toward Primary Education, 1932-1938” inRonald D. Renard (ed.) Anuson Walter Vella. Chiang Mal, Thailand:Walter F. Vella Fund, Payap University, and Honolulu, Hawaii: Center forAsian and Pacific Studies, University of Hawaii at Manoa, 279-294.Green, S.J.D.1988 “Is equality of opportunity a false ideal for society?” The British Journalof Sociology 39(1), 1-27.Gurevich, Robert1972 Khru: A Study of Teachers in a Thai Village. Ph.D. Dissertation.University of Pittsburgh.Hanks, Lucien M. (Jr.)1958 “indifference to modern education in a Thai farming community” Humanorganization 17(2), 9-14.Hannan, Michael T.1971 “Problems of Aggregation” in H.M.Blalock, Jr. (ed.) Causal models in thesocial sciences. Chicago: Aldine Publishing Company, 473-508.Hanushek, Eric A.1972 Education and Race: An Analysis of the Educational Production Process.Lexington, Massachusetts: Lexington Books.95Harvey, David W.1968 “Pattern, Process, and the Scale Problem in Geographical Research”Transactions of the Institute of British Geographers 45, 71 -78.Heyneman, Stephen P.1975 “Changes in Efficiency and in Equity Accruing from GovernmentInvolvement in Ugandan Primary Education” African Studies Review(April), 51-60.1976 “A brief note on the relationship between socioeconomic status and testperformance among Ugandan Primary School Children” ComparativeEducation Review (February), 42-47.1977 “Relationships between the Primary School Community and AcademicAchievement in Uganda” Journal of Developing Areas 11, (January),245-259.1979 “Why Impoverished Children do well in Ugandan Schools” ComparativeEducation 15 (2) (June), 175-185.1 980a “Differences between Developed and Developing Countries: Commenton Simmons and Alexander’s ‘Determinants of School Achievement”Economic Development and Cultural Change 28 (January), 403-406.1 980b “Planning the equality of educational opportunity between regions” inGabriel Carron and Ta Ngoc Chau (eds.) Regional disparities ineducational development: A controversial issue. Paris: IIEP, 115-173.1986 The Search for School Effects in Developing Countries: 1966-1986.Seminar Paper no.33. Washington, D.C.: International Bank forReconstruction and Development (IBRD).1989 “Multilevel Methods for Analyzing School Effects in DevelopingCountries” Comparative Education Review 33 (4), 498-504.Heyneman, Stephen and Dean T. Jamison1980 “Student Learning in Uganda: Textbook Availability and Other Factors”Comparative Education Review 24, pt.1 (June), 206-220.96---, and Xenia Montenegro1984 “Textbooks in the Philippines: Evaluation of the Pedagogical Impact of aNationwide Investment” Educational Evaluation and Policy Analysis 6(2), 139-150.Heyneman, Stephen P. and William A. Loxley1982 “Influences on Academic Achievement across High and Low IncomeCountries: A Reanalysis of lEA Data” Sociology of Education 55 (1), 13-21.1983a “The Distribution of Primary School Quality within High- and Low-Income Countries” Comparative Education Review 27 (1)(February),1 08-118.1 983b “The Effect of Primary-School Quality on Academic Achievement acrossTwenty-nine High- and Low-Income Countries” American Journal ofSociology 88 (6) (May), 1162-1194.Heyns, Barbara1974 “Social Selection and Stratification within Schools” American Journal ofSociology 79 (6), 1434-1451.1986 “Educational Effects: Issues in Conceptualization and Measurement” inJohn G. Richardson (ed.) Handbook of Theory and Research for theSociology of Education. Westport Connecticut: Greenwood Press, 305-340.Hum, Christopher J.1985 The Limits and Possibilities of Schooling (2nd ed.). Boston: Allyn andBacon, Inc.Husen, Torsten et al.1978 “Teacher training and student achievement in less developed countries”Washington, D.C.: IBRD (mimeo.)Ikemoto, Yukio1991 Income Distribution in Thailand: Its Changes. Causes, and Structure.Tokyo: Institute of Developing Economies.Inkeles, Alex1979 “National Differences in Scholastic Performance” Comparative EducationReview 23 (October): 386-407.97Jamison, Dean T., Barbara Searle, Klaus Galda, and Stephen P. Heyneman1981 “Improving Elementary Mathematics Education in Nicaragua: AnExperimental Study of the Impact of Textbooks and Radio onAchievement” Journal of Educational Psychology 73 (4), 556-567.Jamison, D. and M. Lockheed1987 “Participation in Schooling: Determinants and Learning Outcomes”Economic Development and Cultural Change 35 (2), 279-306.Jencks, Christopher et al.1972 Inequality: a reassessment of family and schooling in America. NewYork: Basic Books.Jimenez, E., M.E. Lockheed, and N. Wattanawaha1988 “The Relative Efficiency of Private and Public Schools: The Case ofThailand” World Bank Economic Review 2:139-164.Kaplan, Irving1980 “Chapter 2: The Society and Its Environment” in Frederica M. Bunge(ed.) Thailand. a country study. Washington, D.C.: Foreign AreaStudies, The American University, 49-118.Ketudat, Sippanondha1984 “Planning and implementation of the Primary Education Reform inThailand” Prospects XIV (4), 523-530.Keyes, Charles F. (ed.)1987 Thailand: Buddhist Kingdom as Modern Nation-State. Boulder, CO:Westview Press.1991 ReshaDing Local Worlds: Formal Education and Cultural Change inRural Southeast Asia. Monograph 36. New Haven, Connecticut: YaleUniversity Southeast Asia Studies.1991 “The Proposed World of the School: Thai Villagers’ Entry into aBureaucratic State System” in Reshaping Local Worlds. New Haven,Connecticut: Yale University Southeast Asia Studies, 89-130.Kidder, Louise H. and Charles M. Judd, with Eliot R. Smith1986 Research Methods in Social Relations. (5th ed.) New York: Holt,Rinehart and Winston.Knodel, John, Apichart Jam ratrittirong, and Niphon Thepawan1987 Revolution of family size in Thailand: Rapid decline of fertility growth inThird-World countries, translated by Napaporn Havanon. Bangkok,Thailand: Institute of Population Studies, Chulalongkorn University andInstitute of Demographic and Social Research, Mahidol University.98Komin, Suntaree1989 Social Dimensions of Industrialization in Thailand. Bangkok, Thailand:National Institute of Development Administration (NIDA). (November).Kostakis, Anastasia1987 “Differences among school outputs and educational productionfunctions” Sociology of Education 60, 232-241.Lam, Nina Siu-Ngan and Dale A. Quattrochi1992 “On the Issues of Scale, Resolution, and Fractal Analysis in the MappingSciences” Professional Geographer 44 (1), 88-98.Lareau, Annette1993 Home Advantage: Social class and parental intervention in elementaryeducation. London, New York, Philadelphia: The Falmer Press.Larpthananont, Phinit and Wattana Wongkiattirat1992 Investigation on measures of socioeconomic status in the Thai society.Bangkok: Division of Research, Chulalongkorn University. (March). (inThai).Leonor, Mauricio1982 “Access to primary schools in Thailand” in Peter Richards (ed.) BasicNeeds and Government Policies in Thailand. Singapore: Maruzen Asia,105-125.Levine, Donald M. and Mary Jo Bane (eds.)1975 The ‘inequality’ controversy: schooling and distributive justice. NewYork: Basic Books.Lockheed, Marlaine E., Stephen C. Vail, and Bruce Fuller1986 “How Textbooks Affect Achievement in Developing Countries: Evidencefrom Thailand” Educational Evaluation and Policy Analysis 8 (4), 379-392.Lockheed, Marlaine E. and Nicholas T. Longford1989 A Multilevel Model of School Effectiveness in a Developing Countries.World Bank Discussion Paper no. 69. Washington, D.C.: The WorldBank.Lockheed, M.E. and A. Komenan1989 “Teaching Quality and Student Achievement in Africa: The Case ofNigeria and Swaziland” Teaching and Teacher Education 5 (2), 93-113.Lockheed, Marlaine E., Bruce Fuller, and Ronald Nyirongo1989 “Family Effects on Students’ Achievement in Thailand and Malawi”Sociology of Education 62 (October), 239-256.99Mackler, B.1969 “Grouping in the Ghetto” Education and Urban Society 2, 80-96.McDermott, R.P.1977 “Social Relations as Context for Learning” Harvard Educational ReAew47, 198-21 3.Mehan, Hugh1992 “Understanding Inequality in Schools: The Contribution of InterpretiveStudies” Sociology of Education 65 (January), 1-20.Meyer, John W.1970 “High School Effects on College Intentions” American Journal ofSociology 76 (July): 59-70.Moore, Frank J., with Clark D. Neher1974 Thailand: its people. its society. its culture. New Haven: HRAF Press.Myers, Charles N. and Chalongphob Sussangkarn1992 Educational Options for the Future of Thailand: A Synthesis. The 1991Year-end Conference by the Chai Pattana Foundation and the ThailandDevelopment Research Institute Foundation. 14-15 December, 1991,Chon Bun, Thailand.Na Thalang, Ekkavidya (ed.)1970 Education in Thailand: A Century of Experience. A re’Ased version of theThird Academic Conference, held by Department of Elementary andAdult Education, Ministry of Education, Thailand. Bangkok: KarnsasanaPress, Mr. P.Tipayanetra Publisher.National Education Commission of Thailand (NEC)1974 Equality of Educational Opportunity: A Study of Primary Schooling inThailand. Bangkok: Office of the National Education Commission: Officeof the Prime Minister, Ministry of Interior, and Ministry of Education.1976 A study of primary schooling in Thailand. the final report: factors affectingscholastic achievement in Thailand. Thailand: Office of the NationalEducation Commission.1990 Factors affecting the quality of primary schools. Bangkok: NEC. (inThai).1992a The Seventh National Educational Plan (1 992-1 996). Bangkok: NEC. (inThai).1001 992b Allocation of Budget and Analysis of Educational Costs at the PrimaryLevel. Bangkok: NEC. (in Thai)National Statistical Office, Office of the Prime Minister, Thailand1986 Report of the socio-economic survey - Bangkok Metropolis. Bangkok:NSO. (in Thai).1991 Statistical Yearbook Thailand. No. 38. Bangkok : NSO. (in Thai).Niles, F.1981 ‘Social Class and Academic Achievement: A Third WorldReinterpretation” Comparative Education Review 25, 419-430.Norusis, Marija J.1988 SPSS/ PCi- V.2.0 base manual for the IBM PC I XT I AT and PS / 2.Chicago, Illinois: SPSS Inc.Passow, A. Harry, Harold J. Noah, Max A. Eckstein, and John R. Mallea1978 The National Case Study: An Empirical Comparative Study of Twenty-one Educational Systems. Stockholm, Sweden: Almqvist & WiksellInternational.Peaker, Gilbert F.1971 The Plowden Children Four Years Later. London: National Foundationfor Educational Research in England and Wales.Pedhazur, Elazar J.1982 Multiple Regression In Behavioral Research - Explanation andPrediction. (2nd ed.). New York: Helt, Rinehart and Winston.Pitsuwan, Surin1985 Islam and Malay Nationalism: A Case Study of the Malay-Muslims ofSouthern Thailand. Bangkok: Thai Khadi Research Institute,Thammasat University.Raudenbush, Stephen W., Somsri Kidchanapanish, and Sang Jin Kang1991 “The Effects of Preprimary Access and Quality on EducationalAchievement in Thailand” Comparative Education Review 35 (2), 255-273.Riddell, Abby Rubin1989 “An Alternative Approach to the Study of School Effectiveness in ThirdWorld Countries” Comparative Education Re’view. 33 (4), 481-497.Rist, Ray1970 “Social Class and Teacher Expectations: The Self-Fulfilling Prophecy inGhetto Education” Harvard Educational Review 40, 411 -451.101Robinson, W.S.1950 “Ecological Correlations and the Behavior of Individuals” AmericanSociological Review 15 (3), 351 -357.Rubinson, Richard1986 “Class Formation, Politics, and Institutions: Schooling in the UnitedStates” American Journal of Sociology 92 (3) (November), 519-548.Rutter, Michael, Barbara Maughan, Peter Mortimer, and Janet Ouston1979 Fifteen Thousand Hours: Secondary Schools and Their Effects onChildren. Cambridge, Massachusetts: Harvard University Press.Saha, Lawrence J.1983 “Social Structure and Teacher Effects on Academic Achievement: AComparative Analysis” Comparative Education Review 27, 1: 69-88.Samudavanija, Chai-Anan1987 The State. Bangkok, Thailand: Chulalongkorn University Press. (inThai).Sapianchai, Pote, Aroonsi Jitjang, Surang Phopruksawong, Patya Saihoo,Supang Chantavanich, and Utumporn Thongutai1988 “The Educational Research Environment in Thailand” in S.Gopinathanand H.Dean Nielsen (eds.) Educational Research Environments inSoutheast Asia. Singapore: Chopmen Publishers for Southeast AsiaResearch Review and Advisory Group in association with theInternational Development Research Centre of Canada, 185-234.Saradatta, Lamaimas and Chancha Savannathat1973 “Attitudes of Parents Toward Occupations and Education” in AmnuayTapingkae and Louis J. Setti (eds.) Education in Thailand: Some ThaiPerspectives. Washington: U.S. Government Printing Office, 48-50.Satha-Anand, Chaiwat1987 Islam and Violence: A Case Study of Violent Events in the Four SouthernProvinces. Thailand. 1976-1981. USF Monographs in Religion andPublic Policy. Tampa, Florida: Department of Religious Studies,University of South Florida.Selltiz, Claire, Lawrence S. Wrightsman, and Stuart W. Cook1976 Research Methods in Social Relations. (3rd ed.) New York: Holt,Rinehart and Winston.Setapanich, Nongram1982 Soclo-economic status, school resources and achievement: acomparative analysis among regions and types of schools in Thailand.Ph.D. Dissertation. University of Chicago.102Sewell, William H. and Robert M. Hauser1976 “Causes and Consequences of Higher Education: Modes of the StatusAttainment Process” in William H. Sewell, Robert M. Hauser, and DavidL. Featherman (eds.) Schooling and Achievement in American Society.New York: Academic Press, 9-28.Shiefelbein, Ernesto1980 “The Impact of American Educational Research on DevelopingCountries” in John Simmons (ed.) The Education Dilemma. Oxford:Pergamon Press, 135-142.Shiefelbein, E. and J. Simmons1981 “Determinants of School Achievement: A Review of Research forDeveloping Countries” mimeo. Ottawa: IDRC.Shiefelbein, Ernesto and Joseph P. Farrell1982 Eight Years of Their Lives. Ottawa: IDRC.Shiefelbein, Ernesto, Joseph P. Farrell, and Manuel Sepulveda-Stuardo1983 The Influence of School Resources in Chile: Their Effect on EducationalAchievement and Occupational Attainment. Washington, D.C. : TheWorld Bank.Simmons, John (ed.)1980 The Education Dilemma: Policy Issues for Developing Countries in the1 980s. Oxford: Pergamon Press.Simmons, John and Leigh Jexander1980 “Factors which Promote School Achievement in Developing Countries: AReview of the Research” in John Simmons (ed.) The EducationDilemma: Policy Issues for Developing Countries in the 1980s. Oxford:Pergamon Press, 77-95.Smith H. and P. Cheung1986 “Trends in the Effects of Family Background on Educational Attainmentin the Philippines” American Journal of Sociology 91(6), 1387-1408.Sudaprasert, Kamol, Vichai Tunsiri, and Ta Ngoc Chau1980 “Regional disparities in the development of education in Thailand” inGabriel Carron and Ta Ngoc Chau (eds.) Regional disparities ineducational development: Diagnosis and policies for reduction. Paris:Unesco, IIEP, 197-324.Thailand Development Research Institute1987 1987 Annual Report. Bangkok, Thailand: Thailand DevelopmentResearch Institute.103Theisen, Gary L., Paul P.W. Achola, and Francis Musa Boakari1983 “The Underachievement of Cross-national Studies of Achievement”Comparative Education Review 27 (1), 46-68.Thornton, Clarence, and Bruce K. Eckland1980 “High School Contextual Effects for Black and White Students: AResearch Note” Sociology of Education 53 (October), 247-252.Tuchrello, ‘Mlliam P.1989 “Chapter 2: The Society and Its Environment” in Barbara Leitch LePoer(ed.) Thailand. a country study (2nd ed.). Washington, D.C.: FederalResearch Division, Library of Congress, 60-120.Turner, Bryan1986 Equality. Chichester: Ellis Horwood Ltd., and London and New York:Tavistock Publications.Turton, Andrew1989 “Thailand: Agrarian Bases of State Power” in Gillian Hart et al. (eds.)Agricultural Transformation: Locai Processes and State in SoutheastAsia. Berkeley: University of California Press, 52-69.Watson, Keith1982 Education Development in Thailand. Hong Kong: Heinemann Asia.Weldon, K. Laurence1986 Statistics: A Conceptual Approach. Englewood Cliffs, New Jersey:Prentice-Hall, Inc.World Bank1 990a World Development Report. 1990. Oxford: Oxford University Press forthe World Bank.1990b “Thailand’s Education Sector at a Crossroads: Selected Issues.”Washington, D.C.: World Bank, Population and Human ResourcesDivision, Country Department II, Asia Region. mimeo.Wyatt, David K1969 The Politics of Reform In Thailand: Education in the Reign of KingChulalongkorn. New Haven and London: Yale University Press.1975 “Education and the Modernization of Thai Society” in G. William Skinnerand A. Thomas Kirsch (eds.) Change and Persistence in Thai Society.Ithaca and London: Cornell University Press, 125-149.104Appendix A: List of educational regions, provinces, and districts in thestudy:Educational Code Coderegion Province Number District(Amphur) Number01 Nonthaburi 01 Muang (major) 01Bang BuaThong 02Sai Noi 03Pak Kret 0402 Pattani 02 Muang (major) 01Sai Bun 02YaRing 03KokPo 04Sub-district ThungYang Daeng 0503 Song Khla 03 Muang (major) 01Ranod 02Na-thawee 03Thaepha 04Haad Yai 05Sub-district KuanNiang 0604 Ranong 04 Muang (major) 01Kra Bun 0205 Phet Bun 05 Muang (major) 01Cha-aam 02KhaoYoi 0306 Sing Bun 06 Muang (major) 01Inn Bun 02Bang Rajaan 03Phrom Bun 04Uthai Thanee 07 Muang (major) 01Nong Khaa Yang 0207 Phitsanulok 08 Muang (major) 01Wang Thong 02Phrom Phiram 03Nakhon Thai 04Kamphaeng Phet 09 Muang (major) 01Sai Ngam 02Khlong Khlung 03Khanuworalaksa Bun 0408 Chiang Mai 10 Muang (major) 01Phrao 02Chiang Dao 03Cm Koi 04San Kamphaeng 05Doi Saket 06Faang 07Saraphee 08105Appendix A: (cont’d) List of educational regions, provinces, and districts in thestudy:Educational Code Coderegion Province Number District(Arnphur) Number08 cont’d Chiang Rai 11 Muang (major) 01Mae Sai 02Wiang Pa-Pao 03Phaan 04Chiang Saen 0509 Loei 12 Muang (major) 01Phuu rua 02Chiang Khaan 0310 Mook Dahaan 13 Muang (major) 01Nong Sung 02Roi Et 14 Muang (major) 01ThawatBuri 02Pone Thong 03Muang Suang 04AatSaa Mat 05Kaset Wisai 0611 SiSaket 15 Muang (major) 01Kan Tharalak 02KhuKhan 03Uthum porn Phisai 04Sub-district Bung BoonO512 NaKhonNaaYok 16 Muang (major) 01Paak Phlee 02BanNaa 03Rayong 17 Muang (major) 01Baan Khaii 02PluakDaeng 0313 Bangkok Metropolitanl8 Phra Nakhorn 01Yaan Nawaa 02Thon Bun 03Phasee Jaroen 04Bang Khun Thian 05Meen Bun 06106Appendix B Outlier casesAn outlier or an outlying observation is a data value that is so far from theother data values that it should be presented separately in order to avoid amisleading result (see Weldon, 1986: 101, 146-147; Norusis, 1988: B211). Inthis study, the average scores reported for the Central and Northeast regionsexclude two outlier provinces, Sing Bun and Loei. In addition, subsequent within-region analyses for the Central and Northeast regions exclude the two provincesas they may distort the results of the relative effect of students’ background andschool resources on achievement scores.Sing Bun province was identified as an outlier in the Central region, since forsome unknown reason, the average scores there are more than two standarddeviations higher than average scores from the other provinces in the sameregion (Table 8.1). Moreover, there appears to be a large variation inachievement scores among students in this province, as indicated by the highstandard deviations. The average GPP per capita for Sing Bun (Baht 16,160) issomewhat lower than the average of other Central region provinces (Baht21 ,040). ‘t’ test results show that students in Sing Bun have attended pre-schoolsfor a longer period of time and currently have lower average student-teacherratios than students in the other central provinces. Pre-schooling and student-teacher ratios may contribute to high students’ achievement scores. Which ofthese factors is I are responsible for the outlying status of Sing Bun is not clearas there are not enough cases from this province to conduct reliable regressionanalyses.Table B.1: Average standardized achievement scores of Sing Bun and otherprovinces in the Central Region. (Standard deviations are inparentheses.)Sing Burl Other ProvincesThai language 1.42 (.614) -.00 (.541)Mathematics 1.49 (.882) .03 (.622)Life experience 2.07 (.937) -.03 (.479)Work experience 2.54 (1.144) .00 (.525)Character development 1.70 (.945) -.04 (.406)Loei was excluded from the Northeast region analyses, because its’ GPP percapita is much more than two standard deviations from that of the otherprovinces in the region. It also has lower achievement scores than the otherthree provinces sampled for the region (Table B.2). The high GPP per capita inLoei may be due to its tourism industry and trading activities with neighboringLaos. The relatively high average income of Loei residents may not necessarilylead to high levels of students’ achievement because parents may take theirchildren to work with them, resulting in the children not having enough time tostudy. ‘t’ test results show that students in Loei have lower absenteeism ratesand fewer textbooks than students in the other Northeastern provinces. It is notpossible to determine what factors affect students’ achievement in Loei, due tothe inadequate number of cases.107Table B.2: Average achievement scores and GPP per capita (in 1987) of Loeiand other provinces in the Northeastern Region. (Standarddeviations are in parentheses.)Loei Other ProvincesGPP /capita (Baht) 10,170 6,717 (260)Thai language -.53 (.565) -.32 (.516)Mathematics -.32 (.481) -.30 (.571)Life experience -.52 (.472) -.25 (.584)Work experience -.29 (.605) -.07 (.598)Character development -.42 (.265) -.10 (.593)AppendixCCorrelationmatricesofvariablesselectedinthestudyVariabIeThaiMathUfeWorkCharactGPP/DialectAbsent-HelpinNo.ofPre-SchoolTeachsStudents/TeachingTimeMarkingExper.Exper.Develop.CapitaeeismHomeworkTextsschoolSize.withBATeacher.AidsTeachingHomeworkThai1.0000.6846.7523.6056.5200.2912.4205-.0516.1880.3088.3233.3695.1760.0491.1330*.0232.2453Math1.0000.8136.6551.5848.2716.3616-.0749.1786.2684.2939.3260.0957.0109.1705’.0654.2051LifeExperience1.0000.8008.7001.2360.3631-.1027.2574.3087.3419.3387.0777-.0084.1901.0745.2366WorkExperience1.0000.6440.0197.2577-.1228.1573*.1489*.1346*.1149.1172-.0821.1705.0400.0920CharacterDevelopment1.0000.1011.2168.1369*.1559*.2688.2899.1607*.0569-.0400.1560*.0841.1347*GPPperCapita1.0000.5559.2727.1586*.3490.2517.3745.0786.1970.1015-.0570.4023Dialect1.0000.3260.1606*.3443.3708.3564.2034.0357.1695-.1220.3943Absenteeism1.0000-.0421.1310*.0481.1306*.0277.0962.0067-.0728.1220HelpinHomework1.0000.0738.2288.1808-.0649.0272.0696.0662.1489*NumberofTextbooks1.0000.3871.2342-.0866-.0168.1795-.0299.2230Preschooling1.0000.3869-.0862.0513.1604*-.1078.3902SchoolSize1.0000-.0083.3573.0525-.0668.4789TeacIierswthB.A1.0000-.0812.1684-.0155-.0068Students/Teacher1.0000-.0291.0767.1161TeachingAids1.0000-.0898.0280TimeTeaching1.0000.3390**MarkingHomework1.0000+Foradescriptionofvariables,seeChapter3*p<.01<.001-L 0 0,


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



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"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items