UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Order effects in the measurement of social development Miki, Laureen Wynne Yumi 1992

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

Item Metadata


831-ubc_1992_fall_miki_laureen_wynne_yumi.pdf [ 2.52MB ]
JSON: 831-1.0054484.json
JSON-LD: 831-1.0054484-ld.json
RDF/XML (Pretty): 831-1.0054484-rdf.xml
RDF/JSON: 831-1.0054484-rdf.json
Turtle: 831-1.0054484-turtle.txt
N-Triples: 831-1.0054484-rdf-ntriples.txt
Original Record: 831-1.0054484-source.json
Full Text

Full Text

ORDER EFFECTS IN T H E M E A S U R E M E N T OF SOCIAL D E V E L O P M E N T  by Laureen Wynne Yumi Miki  B.A., The University of British Columbia, 1986  A THESIS S U B M I T T E D IN P A R T I A L F U L F I L L M E N T O F T H E R E Q U I R E M E N T FOR T H E D E G R E E O F . M A S T E R OF A R T S in T H E F A C U L T Y OF G R A D U A T E STUDIES (Department of Educational Psychology and Special Education, School Psychology)  We accept this thesis as conforming to the required standard  T H E UNIVERSITY OF BRITISH C O L U M B I A October 1992 (c) Copyright Laureen Wynne Yumi Miki, 1992  In  presenting  degree freely  this  at the  thesis  in  partial  fulfilment  of  University  of  British  Columbia,  I agree  available for  copying  of  department publication  this or of  reference  thesis by  this  for  his thesis  and study. scholarly  or for  her  Department  of  The University of British Vancouver, Canada  Date  -6  (2/88)  Jet ^  Columbia  ^ mi-  purposes  requirennents that  agree  may  representatives.  financial  permission.  I further  the  It  gain shall not  be is  that  the  permission  granted  allowed  an  advanced  Library shall make  by  understood be  for  for  the that  without  it  extensive  head  of  my  copying  or  my  written  Abstract The purpose of the present study was to examine the role of order effects between two measures of social negotiation: the Hypothetical Interpersonal Negotiations Strategies Interview (Schultz, Yeates & Selman, 1989) and the Real-Life Interpersonal Negotiations Strategies Interview (Schultz and Selman, 1989). The presence of order effects would raise doubts as to the validity of the above interviews. Thirty-six subjects (12 grade fours, 24 grade sevens) were administered the two interviews in one of two orders, HINSI-RLINSI, or RLINSI-HINSI. The findings did not support the main hypothesis: no order effects were seen in the overall scores of the two interviews. However, a secondary hypothesis, that order effects may be seen in interview questions involving adults and unfamiliar people, was supported. These findings suggest that whereas no overall order effects occur when administering the two interviews, caution should be used when interpreting responses given under specific types of social contexts.  Table of Contents  Abstract  ii  Table of Contents  iii  List of Tables  iv  List of Figures  V  Acknowledgements  vi  Introduction  1  Literature Review  9  Method  22  Results  34  Discussion  43  References  49  List of Tables  Table  Title  Paae  1.  Levels of Perspective-Taking & Negotiations Strategies  14  2.  Examples of HINSI Dilemmas  24  3.  Standard Questions and Probes of the HINSI  25  4.  Examples of Real-Life Dilemmas Generated by Subjects  29  5.  Standard Questions and Probes of the RLINSI  30  6.  Descriptive Statistics of the Sample  36  List of Figures  Figure  Title  Page  1.  Mean Levels of HINSI & RLINSI overall scores  38  2.  Mean Levels of HINSI & RLINSI scores by generation  40  3.  Mean Levels of HINSI & RLINSI scores by familiarity  41  4.  Mean Levels of HINSI & RLINSI depicting gap within the  5.  unfamiliar contect  45  Mean Orientation of HINSI & RLINSI overall scores  47  Acknowledgments There are many people who made it possible to finish this thesis. My committee members each made a unique contribution. Billie Housego gave me her constant support, Nand Kishor made certain mathematical concepts accessible and Larry Walker inspired me, through his teaching and his genuine concern for the education of his students, to pursue the study of psychology beyond the undergraduate level. Thanks are also in order to Bob Uttl who gave me his expertise without reserve and more time and patience than I'm sure I deserved. I would never have arrived at this point academically if not for my wonderful family. Each member has stood by me over the many years I have been attending university and they have all given me their unquestioning support for which I will be eternally grateful. My friends are to be thanked as well. It was their warm and enticing company that often tempted me, successfully, to gleefully neglect my work. It was also their understanding that allowed me the time alone to complete this thesis on time. Heartfelt thanks as well to two special people: one who taught me the value of subjective experience and the importance of metaphor not only in art but in everyday life as well. Without those lessons, my understanding of psychology would have been forever limited and very lifeless. Then there was the other whose quiet acceptance gave me the stability to focus throughout the somewhat stressful periods I have undergone during graduate school. There is one more very special person who taught me the importance of asking "why" for everything that I do, a simple word which has helped me immensely, both within and without academia. This same person, through patient tenacity, also made it clear that living life involves more than passive observation....it sometimes involves jumping off mountains - literally and metaphorically.  Introduction  The Assessment of Social Development Increasingly, in the research on social development, the need has been emphasized to assess the construct from both a competence and a performance perspective (Yeates et al., 1990). The term "competence" is used to describe an individual's highest level of responding, while "performance" denotes the level at which actual and overt responses occur. The two constructs of competence and performance have often been translated into the terms thought and behaviour, respectively. Thoughts, when unencumbered by the need for action, are considered to reveal an individual's highest level of competence, or, understanding of a construct and behaviour is a reasonable way to discuss an individual's performance. This emphasis on understanding social development from both perspectives stems from research findings (Schultz & Selman, 1988) which suggest there is often a difference in how an individual discusses hypothetical social dilemmas (thought) and how that same individual would behave in an actual real-life setting (behaviour). This suggests that knowledge of both thought and behaviour, and the relationship that exists between the two, is necessary in order to create a more accurate picture of social development. This is important not only for how we view the construct of social development itself, but ultimately, for developing effective strategies or programs concerned with the enhancement of social development when applied to the social education of average populations, populations at-risk or for those which currently suffer from extreme deficits in social thought and/or behaviour. In the area of social-cognitive development, a competence measure typically involves an interview or questionnaire designed to elicit an individual's highest level of understanding regarding social issues. A performance measure may involve natural observations of real-life behaviour, behavioural ratings by a significant other, or may rely on the interview technique which asks individuals to report how they have actually behaved in past social situations. The Hypothetical Interpersonal Negotiation Strategies Interview (HINSI), (Schultz et al., 1989) was  developed to examine competence as it occurs in the specific domain of social negotiation. The Real-Life Interpersonal Negotiation Strategies Interview (RLINSI) (Schultz & Selman, 1989), is a complementary interview designed to measure the level of behaviour used in real-life situations of social negotiation. Together, these interviews are considered to capture the important relationship between thought and behaviour within the realm of social negotiation (Schultz and Selman, 1989). While research has shown that there is a signficant correlation between the HINSI and the RLINSI (Schultz & Selman, 1989), there is a surprisingly dubious relationship between the RLINSI and behavioural ratings of social negotiations (Fleischer, 1989) Further, the relationship between the HINSI and the RLINSI has been misrepresented in other research (compare discussions of this relationship by Fleischer, 1989, and Schultz & Selman, 1989). Despite this evidence, the authors of the RLINSI suggest that it is a valid measure of social behaviour (Schultz & Selman, 1989). Relatively little research has been done on the RLINSI as it presently exists (Schultz & Selman, 1989; Fleischer, 1989), and consequently, there is an even greater paucity of information regarding its use in combination with the HINSI. One of the problems with this lack of research is that concerns with methodology have yet to be addressed. To date, research involving the two measures has administered them in the same order, first the HINSI, followed by the RLINSI. Because these interviews are very similar in structure, the influence of order effects is suspect. Christensen (1988) notes in his book on experimental methodology that it is imperative to control for the influence of extraneous variables on the dependent variable when conducting research. Without the necessary degree of control over known extraneous variables, such as order, the validity of research findings must be questioned. The focus of this study is to examine the effect that order may have when these two interviews are used in combination. If, in fact, order effects do occur, then one must seriously question the validity of the interviews when they are used together.  Adolescence and Social Development Considerable research has shown that social development plays an important role in psychological adjustment (Cowen, Pederson, Babigian, Izzo, & Trost, 1973). Adolescence is a particularly important period in social development. It is a time when individuals not only face new and qualitatively different kinds of social experiences with peers and adults such as dating and part-time work but also when evaluations of social thought and behaviour by friends and peers may be particularly harsh. Peers and friends play a major role in adolescent social development as they provide the opportunity to experiment with new types and styles of social interaction (Reisman, 1985). The arena where most of this interaction occurs is the school. For almost all their adolescent years, students will spend a large portion of their time in this institution. Furthermore, research has suggested that relationships in school have an influence on later psychological adjustment (Parker & Asher, 1987). For this reason alone, the need for the school to play a role in fostering social development is not only obvious but imperative as well. Indeed, social development is one of the mandates of Curriculum 2000, the document derived from a provincial Royal Commission (1989) aimed at examining the role and scope of education in British Columbia. This document clearly states the intention of the school to take greater responsibility for the social development of students at all levels. While this objective is very progressive and commendable, the question of how the school is planning to promote such development immediately arises. Taking on such a responsibility implies more than mere concern. Systematic assessment and intervention is required which must be preceded by an understanding of the construct of social development and its measurement and manipulation. By adopting a particular model of social development, educators will receive the structure and organization necessary to guide their ideas and actions. When searching for a model of social development, a major concern is its theoretical foundation. This is important because a theory serves as an heuristic, guiding ideas and positing predictions. Without such a basis, educators formulate their ideas on idiosyncratic experiences  and intuition and while it may be that the product of such efforts is fruitful, this method is too unsystematic to confidently suggest the cause of difficulties or to provide a guide for replication. For these reasons, without providing a reliable, theoretical base for a program of social development, it is very difficult to justify its use by other schools or even other teachers. In searching for a model with a sound theoretical base, there are various global aspects of social development to consider, such as what the construct actually implies, and how it should be observed or measured. While a growing body of research on the subject frequently uses the term social competence to describe normal social development, there is still relatively little agreement on what the term social competence actually means. Yeates and Selman (1989) maintain that, on a short-term basis, social competence is heavily context-bound and therefore, its definition will change from context to context. They further explain that social competence, as an indicator of long-term adjustment, may be better understood as a term to describe the outcome of an individual's adaptation to the social environment. From this perspective, Yeates and Selman (1989) define social competence, as the development of the social-cognitive skills and knowledge... that mediate behavioural performances in specific contexts, which are in turn, judged by self and others to be successful and thereby increase the likelihood of positive psychosocial adjustment... In the end, social competence does not depend on a specific cognitive process or behavioural manifestation, or even on short-term judgments of success in a particular behavioural domain. Instead, it depends on the degree to which such factors lead to positive social adaptation...(p. 66).  The above statement presents a framework with which to view the construct, suggesting that a particular defintion of social competence will vary depending on the social context and on whether the focus of interest is on the short- or long-term implications. In attempts to observe or measure social competence, two important aspects must be captured. The first one is behaviour. It would be impossible to leave this component out of any assessment of social competence for, by its very definition, the word "social" implies that it cannot occur in an individual vacuum. It must observe the real and actual behaviour that occurs within interpersonal interactions. The second aspect that a form of measurement must capture is  social cognition. A n understanding of the cognitive processes behind social interactions is imperative for appropriately interpreting an individual's behaviour. As an example, observations of a student's behaviour with peers may be classified as aggressive. In and of itself, this description may be quite accurate. The danger, however, in relying solely on behavioural observations lies in the tendency of people to infer intent. In the case of aggressive behaviour, the label often changes to "hostile" which denotes malintent. By investigating the cognitive basis for such behaviour, the chance of misclassification can be reduced. In this instance, such an investigation may reveal that the motivation for this behaviour lies with the student's strong need for peer affiliation and approval. In light of this knowledge, the description of "hostile" becomes inappropriate. Rather, the student's behaviour may be better understood as a series of ineffective social strategies for gaining friendship instead of being viewed as hostile. This example highlights the importance of understanding the intentions behind a behaviour especially with its implications for designing interventions. While interventions which are strictly behavioural may seem more appealing due to the speed with which they can be applied, including cognitive changes may stand a better chance of generalizing to other related behaviours in other, related domains (Gresham, 1985). Robert Selman and his colleagues have been working on a model of social development. This model focuses on a particular domain or context of social functioning - how individuals negotiate with others in the face of conflict. Selman and his associates believe that while this aspect of social interaction captures a large proportion of all social interactions, it also relates to only a specific context, which, as mentioned previously, is important to operationalizing a shortterm definition of social competence. The Interpersonal Negotiation Strategies (INS) model (Selman, 1981, Selman & Demorest, 1984) is based on the developing ability to coordinate the perspectives of self and others when formulating, choosing and implementing social strategies. These researchers posit that, initially, human social behaviour is guided by only a single perspective, that of the self. However, by adolescence, individuals are able to consider the perspectives of others while simultaneously considering their own. This development is  demonstrated by the more collaborative nature of their social interactions. The Hypothetical Interpersonal Negotiations Strategies Interview (Schultz, Yeates & Selman, 1989), hereafter referred to as the HINSI, elicits information on this perspective-taking ability by presenting various hypothetical social conflicts or dilemmas between the self and one significant other. Through standardized questions following presentation of the dilemmas, the individual is asked to identify the conflict, generate various solutions to it and describe the perspectives of those involved. This interview is carried out in a non-threatening environment and is designed to elicit the highest level of understanding of which the individual is capable. A shortcoming of this assessment measure, however, is that we can never be sure that the level of response given in the interview will reflect the level of response seen in actual, everyday behaviour. For this reason, a second interview was developed (Gurucharri et ai., 1984). Its most recent form was used in recent research (Schultz & Selman, 1989) and shall hereafter be called the Real-Life Interpersonal Negotiations Strategies Interview, or the RLINSI. This interview asks the individual to recall personal, real-life situations when various types of conflict arose. The use of this second interview is crucial to assessment for it allows the examiner to inspect the responses to the hypothetical dilemmas (thought) and self-reported situations (behaviour) and the match between the two. Research to date, albeit limited, has demonstrated a significant correlation between responses of the first interview and responses in the second (Schultz & Selman, 1989). Selman and his associates have used the INS model in applied settings with successful results (Selman & Schultz, 1988; Selman & Arboleda, 1986; Lyman & Selman, 1985). While this research has only been conducted in individual therapy, it may be of particular interest for schools, not only as a means of assessment but also because it contains inherent suggestions on how to incorporate the model into teaching styles which could be incorporated with relative ease across the curriculum (Yeates & Selman, 1989). Although use of the two interviews holds promise for implementation in the school, there are still some problems that limit confidence with regard to validity. While a considerable amount of research has confirmed the reliability and validity of the HINSI, Selman and his  colleagues rely heavily on the assumption that the RLINSI is not only a valid means of assessing actual behaviour but also that it is not influenced by the preceding interview. To date, all research using these interviews has presented the HINSI first followed by the RLINSI. For this reason, it is impossible to know whether or not responses given during the RLINSI are somehow influenced by the HINSI. It may be that when the HINSI is administered first, it somehow cues the individual on how to respond during the RLINSI. If there is a significant degree of influence, the validity of these measures, when used in combination, must be questioned. No prior research examining the effect of order on these two interviews was found, although one researcher has commented on its possible role noting that, "children's descriptions of actual dilemmas they had experienced might also be influenced by a "set" created by the preceding hypothetical dilemmas and probes" (Fleischer, 1989). Even in related areas of research in social development, there is scant information on the effects of order. In 1976, two separate papers were published by the same group of researchers (Parsons et al., 1976, and Feldman et al., 1976) which examined the effect of order of information presentation on children's moral judgements. These researchers found that order played a significant role when examining the importance of intent- or consequence-based moral decisions. In other related areas of research, such as Piagetian class inclusion tasks (Gold, 1987) and children's responses to gender-constancy questions (Siegel & Robinson, 1987), order effects were also noted. These results were found with young subjects, ranging in age from 3 to 8 years. Only one study was found which dealt with slightly older children and adolescents (Grueneich, 1982). While the results of this research revealed the presence of order effects when children made moral judgements, age differences between third- and sixth-graders were also observed when the order of information regarding moral stories was varied. Based on the above information, there is sufficient evidence to question the resilience of the RLINSI to order effects. The purpose of this study, therefore, is to assess the vulnerability of the RLINSI to order effects by administering it both before and after the HINSI to two separate groups. By examining the issue in this way, more information will be provided  regarding the amount of confidence that can be placed in the RLINSI. This is important because these results will ultimately affect the confidence we have in the assessment technique of the INS model. In chapter two, a brief history of the field of social competence and its current state of affairs will be presented. Additionally, a more in-depth examination of the research regarding the HINSI and the RLINSI will be included. The chapter will end with a more detailed explanation of the aim of this research. Chapter 3 will describe the design and method of the study. Chapter 4 will present the statistical results from the investigation and Chapter 5 will discuss the implications of the findings.  Literature Review  A Brief History of the Study of Social Competence Social-cognitive development within an individual encompasses the growth in thought and behaviour which occurs in interaction with others. Psychological theories of human social development include Freud, Erikson, Watson, Bandura and Piaget, but it was not until the last several decades that the term "social competence" came into widespread use and became its own specialized field of study. Since that time numerous other terms and definitions of this construct have been employed. Social competence, interpersonal competence, social skills, communicative competence, psychosocial competence and relational competence are just some of the terms used to describe this construct (Spitzberg & Cupach, 1989). This abundance of terms suggests there is a lack of consensus on a definition. In fact, so many different views of the nature and definition of social competence exist that Phillips (1984) likened the task of defining it to "climbing a greased pole," (p. 24). Likewise, Conger and Conger (1982) had this to say about the problem of definition. Undaunted by the absence of specific definitions or criteria we have developed instruments that purport to select and measure people on "it"; indeed, we have treatment programs for people who do not have "it", and still the mysterious "it" eludes specification, (p. 314) It is no wonder, then, that this aspect of human functioning is studied in so many ways. While some researchers prefer to study social competence in terms of overt behaviours ranging from such molecular behaviours as length of eye gaze, gestures and voice tone (Dillard & Spitzberg, 1984) to more molar behaviours such as those involved in entering a new group (Putallaz & Gottman, 1981) or being popular (Dodge, 1983), others have studied social competence from a cognitive perspective, examining the ways in which we reason about the social world. This view focuses on our thoughts about social relationships and the ways in which we choose to deal with them.  Presently, the social-cognitive approach is receiving much attention in the research literature which may reflect the general trend away from strictly behavioural approaches. However, most social-cognitive researchers realize that the role of behaviour in social interactions cannot be ignored, and have therefore incorporated behavioural perspectives into their work. In this regard, there seems to be general consensus within the field. However, there is much less consensus in another area of social-cognitive research which has been responsible for creating a dichotomy in research approaches. This concerns the issue of how one should regard the construct of social-cognition itself. One school of research approaches the study of social-cognition by focusing attention on the social-cognitive processes purportedly involved in social functioning. These researchers examine one component of social competence, namely, the information processing used in social interactions (e.g., Spivak & Shure, 1974). The process, in general, described in this literature involves the following steps (Rose-Krasnor, 1988): 1) selection of a social goal 2) taking relevant information from the environment 3) producing new strategies or retrieving old ones 4) selecting a strategy 5) implementing the strategy.  Researchers supporting this model posit that we move through this process in our social interactions. In the case that an individual achieves the desired outcome after Step 5, no further processing is required. However, in the event that the outcome is unsatisfactory, the individual may return again to any of the preceding steps in further attempts to achieve the goal. This perspective on social competence maintains that problems in social development arise when one or more of these steps is carried out in an inefficient, unproductive or dysfunctional way (Piatt et al., 1974). Intervention therefore, focuses on training in one or more of these steps. While this type of research has provided much information particularly with regard to intervention  techniques, it lends no frameworic for assessing the developmental status of a child's social competence (Yeates & Selman, 1989). Without knowledge of the developmental level of an individual's ability, there is little that can be said regarding its proximity to normal development. The other school of social-cognitive research involved in the forementioned dichotomy focuses on a structural model of social competence. The emphasis here is on the developmental sequence of social thought and as such, describes it in terms of stages following the ideas of Piaget (1934) and Kohlberg (1969). Robert Selman and his colleagues have carried out extensive work in the field of the structure of development in social cognition. Their research began with the study of the developing ability of an individual to consider and coordinate the perspectives of significant others (Selman, 1980). In order to describe the increasing sophistication of this ability in developmental terms, Selman proposed a stage model which had the same theoretical underpinnings as the stage theories of Piaget (1934) and Kohlberg (1969). Thus, the stages involved in perspective-taking develop in an invariant sequence with each higher stage a structured whole being qualitatively different and more highly organized than the last (Selman & Jaquette, 1977). In cross-sectional and longitudinal research (Selman, 1980) empirical evidence was found to justify this claim. However, though this model provided an overall structure with which to gain a developmental perspective on social development, there were several shortcomings. First, the model was very global in its outlook and ignored the role of the socialcognitive processing in perspective-taking. This is to say that it paid little attention to the stepby-step information-processing used in social interactions and therefore, had very little to contribute when discussing the actual social interactions that constitute social development. The second shortcoming of the model was that it offered little in the way of relating social cognition to social behaviour. Pelligrini (1985) stated that although Selman's model described the content of thought, it could not account for how this content was actually used in interaction. Finally, the model did not adequately acknowledge the importance of context. While people may consistently demonstrate a certain level of perspective-taking under ideal circumstances, different contexts may affect this ability substantially (Yeates & Selman, 1989).  The INS Model To address the criticisms levelled against it, Selman and his colleagues continued research with this model which was named the Interpersonal Negotiation Strategies, or INS model (Selman, 1981). This model was a departure from the previous perspective-taking model because, rather than focusing on levels of perspective-taking, it emphasized the development of the strategies used by an individual to deal with conflict with significant others such as peers or adults. In doing so, the model was limited to a specific context or domain of social interaction but, at the same time, included a wide and significant range of social behaviour (Yeates & Selman, 1989). Perspective-taking was now regarded as the underlying structure which guided the reasoning for choosing a particular strategy. In response to criticisms from the functional school of social cognition, the INS model incorporated its version of the social informationprocessing involved in social negotiation. In one study (Selman et al., 1986) the following processes were used: 1) defining the problem 2) taking action 3) justifying and indicating the consequences of the strategy 4) understanding of the feelings expressed. In developing the HINSI (Schultz et al., 1989), the processes highlighted were: 1) defining the problem 2) generating alternative strategies 3) selecting and implementing a strategy 4) evaluating the outcome. As will be discussed shortly, this was a vast improvement to the INS model because by incorporating information-processing into the model, the individual steps involved in social cognition could now be assessed and existing strategies for intervention, based on social information-processing, became accessible.  The need to link cognition to behaviour was also taken into account by this model because now, it not only looks at the underlying levels of perspective-taking involved in negotiating strategies but it also evaluates the actual strategies used or suggested for use in both self-reported and hypothetical situations. Establishing the relationship between thought and behaviour is of crucial importance to the INS model because a major purpose of this model is to identify individuals who demonstrate a particular type of "gap" that may exist between the two. It is understanding this relationship which creates a more complete picture of an individual's social development, and will ultimately determine the approach to intervention (Selman & Schultz, 1988). For example, consider the following types of gaps that may exist between thought and behaviour (Schultz & Selman, 1989) and the implications these gaps carry for treating troubled children and adolescents: Type 1 - when the developmental level of thought is greater than that of self-reported action. Type 2 - when the developmental level of self-reported action is greater than that of thought. Type 3 - when developmental levels of thought and self-reported action are similar but low. Treatment for a Type 1 profile would probably entail teaching an individual various social skills that will allow the level of action to approximate the level of thought. Treatment for a Type 3 gap will probably involve attempts to raise the thought level before attempting to raise action levels. The Type 2 profile is very uncommon (Schultz & Selman, 1989) and seemingly contradicts theories of development. It is possibly due to the puzzling nature of this type of profile that treatment approaches were not discussed in the literature reviewed. Table 1 outlines the four levels of social perspective-taking and their corresponding levels of interpersonal negotiation strategies (Yeates & Selman, 1989).  Table 1 Levels of Perspective-Taking and Interpersonal Negotiations Strategies  LEVEL 0 Perspective-taking - Egocentric and undifferentiated At Level 0, physical and psychological characteristics of persons are not clearly differentiated. This confusion between objective/physical and subjective/psychological features is seen in the failure to distinguish between acts and feelings, and unintentional and intentional behaviour. Subjective perspectives are not differentiated, so that the possibility that another person may interpret the same behaviour differently is not recognized. INS - Impulsive Strategies at the first level involved primarily impulsive and physical behaviour to get what one wants or to avoid harm. They are based on Level 0 perspective-taking skills, which do not differentiate subjective perspectives or distinguish between actions and feelings. Strategies at this level use either unreflective force to achieve a goal, or unreflective obedience or withdrawal to protect oneself. LEVEL 1 Perspective-taking - Subjective and unilateral The key conceptual advance from Level 0 to Level 1 is the clear differentiation of physical and psychological characteristics or persons. Each person is acknowledged to have a unique, subjective, and covert psychological life. Subjective states of others, however, are thought to be directly observable. The relating of perspectives, moreover, is accomplished in one-way, unilateral fashion, in terms of the perspective of only one actor and the impact of social interaction on him or her. INS - Unilateral Strategies at the next level consist primarily of unilateral attempts to either control or appease the other person. They depend on Level 1 perspective-taking skills, which differentiate subjective perspectives, but do not allow them to be considered simultaneously. Thus, strategies at this level involve either willful one-way orders to assert power, control the other person, and satisfy oneself, or "will-less" submission to the power, control and wishes of the other person. LEVEL 2 Perspective-taking - Self-reflective and reciprocal The major conceptual progressions as compared to Level 1 is the ability to step outside oneself mentally and take a second-person perspective on one's thoughts and actions, along with the realization that others can do so as well. People are understood to be capable of acting in opposition to their thoughts and feelings. Differences between perspectives are seen relativistically. A reciprocity occurs, wherein the perspectives of self and other are both appreciated, albeit not in relationship to one another. INS - Reciprocal Strategies at this level involve attempts to satisfy the needs of both participants in reciprocal fashion through trades, exchanges and deals. They rely on Level 2 perspective-taking , which not only differentiates between subjective perspectives, but also allows those perspectives to be considered simultaneously. Reciprocal strategies consciously use either psychological influenced to changes the other person's mind, or psychological compliance to protect one's own interests by making them secondary to the other person's.  Table 1 (cont'd) Levels of Perspective-Taking and Interpersonal Negotiations Strategies  LEVEL 3 Perspective-taking - Third person and mutual Level 3 reflects an advance over Level 2 in that children come to be able to step outside not only their own immediate perspective but indeed outside the self as a system. They can begin to take a truly third-person perspective. The self is seen as both actor and object, as are others. The perspective on relationships simultaneously includes and coordinates the perspective of self and other(s), and the system is seen from a generalized perspective. Reciprocal perspectives are not only acknowledged, but seen to be in need of mutual coordination. INS - Collaborative Strategies at the highest level involve attempts to collaboratively change both one's own and the other person's wishes in order to develop mutual goals. These strategies rest on Level 3 perspective-taking skills, which permit the ability to coordinate the selfs and the other's perspectives in terms of the relationship between them or from a thirdperson viewpoint. At this level, strategies use self-reflection and shared reflection to facilitate the process of dialogue that leads to compromise and the construction of mutually satisfactory resolutions. They demonstrate concerns for a relationship's continuity and the understanding that solutions to immediate problems have a bearing in that regard.  Over the last several years, Selman and his colleagues have continued to work with the INS model, investigating its relation to troubled adolescents (Selman & Demorest, 1984; Gurucharri, Phelps, & Selman, 1984; Beardslee, Schultz, & Selman, 1987), natural social situations (Abrahami, Selman, & Stone, 1981; Selman et al., 1983), IQ and varying interpersonal contexts (Selman et al., 1986) and psychodynamic processes (Schultz & Selman, 1989). At present, the INS model is being employed in "pair therapy", Selman's term for clinical intervention for troubled children and adolescents (Selman & Demorest, 1984; Lyman & Selman, 1985). The main tools of the INS model are the HINSI and the RLINSI which are quite similar to Kohlberg's technique for assessing moral development (Kohlberg, 1969). The method consists of using the HINSI and the RLINSI to examine social competence and performance. It begins with administration of the HINSI. During this interview, an individual is told a series of hypothetical situations in which there is conflict between the protagonist and one significant other. Within these situations, the social context is varied along two dimensions involving the nature of the relationship with the significant other. One dimension on which the social context is varied is by generation, or the age of the signiflcant other. Therefore, the significant others in some of the dilemmas are peers while others are adults. The second dimension along which social context is varied is familiarity, meaning that the significant other is either familiar or relatively unfamiliar to the protagonist. By these variations, the four possible combinations of social context by generation and familiarity are: familiar peer, unfamiliar peer, familiar adult, and unfamiliar adult. The hypothetical situations are followed by a series of standard questions designed to elicit the subject's highest level of understanding of the interpersonal aspects of the situation, in terms of perspective-taking, information-processing and negotiation strategies. Each interview is audiotaped, transcribed, and scored. The HINSI also examines the consistency of the style or orientation of an individual's suggested solutions to the conflicts. Schultz and Selman (1989) have noted that an individual may deal with a social conflict in one of three  ways: 1) in a self-transforming fashion where the self is changed in order to meet the demands of the other, or 2) in an other-transforming style where the self attempts to change the other to meet his or her own goals, or 3) in a collaborative fashion where the self tries to work with the other in order to meet both goals or to create a mutual one. When an individual's orientation cannot be established or is unclear, it is labelled indeterminate. The second part of assessment which examines social behaviour, or social self-reported action, employs use of the RLINSI where subjects are asked to relate several real-life situations from the recent past when they experienced some kind of social conflict. They are also asked to describe the strategies they used to resolve these conflicts. This measure is also audiotaped, transcribed and scored and is used as an indicator of actual behaviour. For both interviews, various scores are derived, the most global ones being an overall score which reflects both pure and transitional stage levels and an overall level score for orientation. The RLINSI is an important measure because its responses are contrasted with those of the HINSI in order to indicate the type and degree of gap that exists between thought and behaviour (Schultz & Selman, 1989). As previously described, the gaps which indicate cause for concern exist when thought scores are higher than self-reported action scores, when selfreported action scores are higher than thought scores and when thought and self-reported action scores are both low. The second pattern is quite rare. Knowing the type of gap that an individual possesses will have implications for the design of intervention programs and thus, researchers and clinicians alike must rely heavily on the validity of this measure. While the use of the RLINSI seems crucial to assessment, the self-report method itself is fraught with all sorts of threats to reliability and validity because it is susceptible to the effects of extraneous factors. One of these factors is social desirability. The body of research on social desirability suggests that our desire to promote ourselves as reasonably good, rational beings may tempt us to report our behaviour as we would have preferred to have acted rather than how we actually did act (Hughes, 1984). This creates doubt as to whether the self-report measure accurately represents actual behaviour and consequently how accurately it can predict actual  behaviour. In Selman's own preliminary work with this measure (Selman et al., 1983), he examined children's self-reports regarding conflicts and interpersonal issues experienced during a 3-month activities club. While these researchers felt this was a natural way to capture an accurate picture of real behaviour, the subjects were aware, at least to some degree, that their ideas and opinions were being evaluated. For example, the subjects were told that the adults in charge were interested in learning how children learn to cooperate. In round-table discussions to elicit such information, microphones, tape recorders and occasionally video recorders were used in full view. Additionally, throughout each meeting, one of the adults observed individual children, recording observations into a tape recorder while moving about the room (Selman et al., 1983). Another factor which may influence responding on the RLINSI is that of order effects. In research using the RLINSI, the HINSI has always been administered first. This may be a problem when one considers the nature of the hypothetical interview. During this first interview, subjects receive information that may somehow serve as a cue to how they should respond in the following interview. Additionally, because these cues provide information to the subject, they will probably serve to increase the level of response on the RLINSI. This source of influence may come from the nature of the dilemmas, the content of the standard questions, the mere presence of the questions themselves, or all three. The main purpose of this study is not to explain how the preceding interview may affect responses on the second interview, although this will be discussed, but to establish whether or not such effects exist at all. If order effects do occur, it is hypothesized that the RLINSI scores will be artificially increased by the preceding HINSI. In an earlier study of Selman's (1984), a relationship between the overall Interpersonal Understanding Score (lUS) and types of self-reported strategies was hypothesized but not analyzed statistically. It was not clear from the journal article describing the study whether the self-report was administered before or after the hypothetical situations interview but due to the longitudinal design of the study, all subjects who gave self-reports had been given the hypothetical interview twice previously. A more recent study which used the RLINSI (Schultz  & Selman, 1989) administered it after the hypothetical situations interview. The results showed correlations of .45 between thought and self-reported action. The authors discuss the issue of the RLINSI's validity citing some preliminary evidence for its support. However, as previously mentioned, the accuracy of this evidence is questionable and the issue of order effects was not discussed. An additional aspect of the hypothesis is that order effects will be unidirectional. In other words, it is hypothesized that while the RLINSI scores will be artificially inflated by the preceding HINSI, the HINSI scores will not be influenced by the preceding RLINSI. The rationale behind this directional hypothesis is based on the assumption that the HINSI purportedly measures the ceiling or highest level of responding of an individual. This interview should not, therefore, be susceptible to influence by the RLINSI when it follows it. While research shows that the correlations between thought and self-reported action are significant but not strong (Schultz & Selman, 1989), controlling for order effects may change these results. Another route of investigation in regard to order effects involves analysis at a more specific level of the two interviews. In a recent article on the INS model (Yeates et al., 1990), it was reported that individuals use higher-level problem-solving when social contexts involve peers rather than adults and when the social contexts involve familiar, as opposed to unfamiliar, others. This suggests that the role of experience is important in determining the level with which an individual will respond to standard questions and probes. Because children and adolescents tend to have more experience in dealing with peers and familiar people, it makes sense intuitively that, when questioned about social interactions with adults or unfamiliar people, they will have had less opportunity for social interaction and therefore, less opportunity to mature in their responses regarding conflicts with these people. It seems then, that due to this lack of experience, children and adolescents will be somewhat unsure of how to respond, in thought or behaviour. If this is the case, then the effect of order may be particularly salient in these social contexts. For this reason, an additional hypothesis of this study was that order would also have an effect within the contexts of generation and familiarity, namely that RLINSI scores would be  artificially inflated by the preceding HINSI when dilemmas involved adults and unfamiliar others. The importance of having confidence in the self-report measure for both research and clinical application cannot be underestimated. While the most recent research involving the INS model no longer includes use of the RLINSI and has concentrated on using teacher ratings of behaviour, possibly due to their stronger correlation to the HINSI, the role of the RLINSI should not be ignored. While some researchers see its primary use as providing a more structured opportunity for an individual to discuss real-life social conflicts (Yeates et al., 1990), its benefits are more far-reaching. When carrying out any kind of psychological assessment, it is obvious that the more information one has about an individual, the more detailed the picture will be of psychological functioning. While behaviour ratings conducted by a significant other add important knowledge about social functioning, so do individuals' perceptions about how they view their own behaviour. These perceptions could be extremely beneficial to understanding the nature of the thought-self-reported action gaps described previously and why they occur. For example, a common thought-self-reported action gap occurs when thought levels are developmentally normal and behaviour levels are low. Therapy to decrease the gap between these two usually involves teaching the individual a series of social skills to raise the level of their behaviour. The design of this therapy would be more effective by knowing how the individual perceives his or her own behaviour. The RLINSI provides such information. Therefore, if an individual reports higher level responses than teacher ratings, the therapist may consider that part of the reason for the thought-self-reported action gap lies with the individual's perceptions regarding his or her own behaviour. This knowledge will help direct therapy toward individual needs and therefore, greater effectiveness. For these reasons, the RLINSI should be an integral part of assessment using the INS model. Because of its contribution, it should not be discarded in favour of other measures. However, until some of the forementioned problems are investigated, confidence in its use will continue to elude researchers and clinicians alike.  In terms of orientation scores and order effects, it was also hypothesized that orientation would be affected. However, there was no body of research on which to determine the direction such an influence may take, since only recently has orientation been included in the INS model (Selman & Schultz, 1988). In previous research, orientation has been scored to the degree that it was other-transforming. Therefore, higher scores in orientation indicate that an individual tends to be more other-transforming. While there was a negative correlation between levels of thought and degree of other-transformingness in one study (Schultz & Selman, 1989), there is not enough research to conclude that this is the normal pattern of occurrence and therefore, we cannot conclude that other-transformingness is necessarily less-desirable than self-transforming or collaborative orientations. At this point then, it would be prudent to suggest that the "acceptability" of a particular orientation depends more on the developmental level with which such orientations are used, rather than on the orientation alone. For this reason, it is difficult to suggest with any certainty how orientation figures in the INS model with regard to gaps in thought and self-reported action, and therefore, due to the limited information available on orientation, the hypothesis that order effects would be seen was limited to the overall orientation, without specifying whether such an influence would cause scores to be increased or decreased. Based on the above overview, the following study will examine the various relationships between: 1) overall level scores on the RLINSI given either before or after the HINSI, 2) level scores on the RLINSI, for adults, given either before or after the HINSI, 3) level scores on the RLINSI, for unfamiliar others, given either before or after the HINSI, and 4) overall orientation scores on the RLINSI given either before or after the HINSI. While an examination of how order effects influence scores is outside the scope of this experiment, the discussion that follows the method and results sections will include such considerations.  Method  Subjects The sample consisted of 24 grade seven students and 12 grade four students for a total of 36 subjects altogether. In the older group, there were 14 females and 10 males, and in the younger group there were 10 females and 2 males. Altogether, there were 24 females and 12 males. The mean age for grade seven and grade fours were 12.9 and 9.9 years, respectively. The grade seven subjects were drawn from two urban elementary schools (School A = 14, School B, n = 10). The grade four subjects were all drawn from School A. In terms of ethnicity, 66.7% of the entire sample was White (grade sevens = 62.5%, grade fours = 75%) and 33.3% was comprised mainly of children of Asian descent (grade sevens = 37.5%, grade fours = 25%). A l l children demonstrated an acceptable command of the English language and no evidence of mental disability was evident during the interviews. A l l subjects participated with the full written consent of the parent or legal guardian. No subject received any kind of remuneration for the participation.  Assessment Measures 1. Hypothetical Interpersonal Negotiation Strategies Interview (HINSI) During this interview, each subject was presented with 4 of 12 possible hypothetical dilemmas from the Interpersonal Negotiation Strategies Interview Manual (Schultz, Yeates, & Selman, 1989). A l l the dilemmas contained some type of interpersonal conflict between a protagonist and one other person. Each dilemma involved a protagonist and one other person from one of four different social contexts: unfamiliar adult, familiar adult, unfamiliar peer, familiar peer. While the peer dilemmas involved settings both in and out of school, the adult dilemmas involved only teachers within the school setting. For each dilemma the subject was asked eight standard questions designed to elicit an understanding of the nature of the problem.  the perspectives of those involved, and possible solutions to the situation. Tables 2 and 3 list examples of the four types of dilemma and the standard questions, respectively.  Table 2 Examples of the four types of dilemmas used in the HINSI by generation and familiarity  Peer - Familiar Randy and Tom are friends. They have been assigned to work together on a science project in school and only have two days to finish the project. They meet after school and Randy says he wants to start working on the project right away, but Tom wants to play Softball first. Peer - Unfamiliar One day a new kid in class named Denise says she's cold and asks Peggy to lend her a sweater that Peggy has but isn't wearing. The next day when Denise returns the sweater there is a hole in it that Peggy is sure wasn't there the day before. Adult - Familiar In class one day, Ben's teacher, Mr. Davis asks if he could stay after school to help plan the class's Halloween party. Ben would like to help but he's supposed to go play soccer with friends that afternoon. Adult - Unfamiliar On day, Stephanie's class has a substitute teacher. Stephanie remembers that she is supposed to leave school early for an important doctor's appointment but she forgot to bring the note from her mother. When Stephanie asks is she can leave, the substitute teacher says that she can't go without a note. Note: the names, connating the sex of the characters are changed to match the sex of the subject.  Table 3 Standard questions and probes of the HINSI  1.  What is the problem here? Why is that a problem?  2.  How does (protagonist) feel? Why does he/she feel that way? How do you think (other person) feels? Why does he/she feel that way?  3.  What are all the things (protagonist) could do to solve his/her problem with (other person)? How would that solve the problem? What else could he/she do?  4.  What would be the best way for (protagonist) to solve the problem with (other person)? Why is that the best way to solve the problem?  5.  How would they feel if (protagonist) did that? Why would they feel that way?  6.  What could go wrong with that solution? Why would that mess things up?  7.  What would (protagonist) do next if that happened? Why would he/she do that?  8.  How would (protagonist) know if he/she had really solved the problem?  Each interview was audiotaped, transcribed, and scored according to the Interpersonal Negotiations Strategies Interview Manual (Schultz, Yeates & Selman, 1989) to determine the subject's underlying level of perspective-taking and problem-solving ability across each of the four social contexts. The manual provides prototypical definitions and examples for each level of development in every dilemma and scoring involves matching the subject's responses to those in the manual. The scores had a possible range of 0 to 3, reflecting the 4 stages of development listed in Table 1. In this experiment, such scores are referred to as HINSI-Level scores. An overall HINSI-Level score for each subject was derived by averaging scores across the four dilemmas and is capable of reflecting both pure and transitional levels (Schultz et al., 1989). Subjects were considered in between stages or "in transition" from one level to the next higher one when overall scores received a decimal point of over .5. An overall score reflects the level of interpersonal understanding of an individual. This overall score is based on the level of perspective-taking that is demonstrated, the level of the strategies suggested to solve the conflict and the level of information-processing demonstrated. Scores can also be broken down to examine responses across specific types of social situations (i.e., generation and familiarity), or the specific levels of information-processing involved. However, because the RLINSI does not yield data on all the steps involved in the INS model of information-processing, such scores will not be used for the analyses. In addition to the above score, subjects were also rated, across the four dilemmas, according to their orientation. This was measured by the degree to which their solutions tended to give priority to the needs of the self, or the degree of "other-transformingness". This information was described as the "interpersonal orientation", or HINSI-Orientation score of the subject and was derived from the types of strategies the subject suggested to solve the problem. The strategies fall into one of three categories: other-transforming, self-transforming, collaborative, or indeterminate. Other-transforming orientations place the needs of the protagonist ahead of the other, self-transforming orientations place the needs of the protagonist second to the other, collaborative orientations consider the needs of the protagonist and the other  equally, and indeterminate orientations are those in which the priority of needs between two people involved in a situation cannot be established. The scores for orientation were based on the degree of "other-transformingness" and thus, other-transforming orientations received a score of 1, self-transforming orientations received a score of 0 and collaborative and indeterminate orientations received a score of .5. This interpersonal orientation score was given for each of the four contexts and was also averaged across the dilemmas. This interview lasted approximately 20 - 30 minutes. Interrater reliability for the averaged overall scores for this interview was high (level scores, r=.87, orientation scores, r=.93). This is similar to the level of reliability found in a previous study (Schultz & Selman, 1989).  2.  Real-Life Interpersonal Negotiation Strategies Interview (RLINSI) During this interview, subjects were asked to report real-life conflicts they had  experienced in the past across the same four social contexts used in the HINSI. Initially, subjects were only asked who the conllict involved, what they did, and what was the outcome. After conflicts across all four social contexts had been related, the standard questions were asked. This was done so that responses could be scored before any of the standardized questions were asked. It was thought that this manipulation would have the most effect on subjects who received the RLINSI first. This group, would respond without the effect of the preceding HINSI or the effect of standardized questions and probes. In accordance with past research (Schultz & Selman, 1989), subjects also responded to a hypothetical situation in which they were the protagonist. This was included in order to make standardized comparisons across subjects and to provide data for subjects who were unable to generate experiences in certain contexts. Within the context of adults, subjects were initially asked to recall conflicts involving teachers in order to make it as similar to the HINSI dilemmas as possible. However, when a subject was not forthcoming, the adult context was expanded to include any adult with whom the subject had had an interaction. As with the HINSI, the peer interactions could involve peers within and without  the school setting. This interview was also audiotaped, transcribed and scored in the same way as the HINSI. For the RLINSI, only six questions were asked for each dilemma. The two HINSI questions that were excluded from the RLINSI were the ones which asked the subject to generate possible and best solutions to a dilemma. Because the RLINSI was used as a measure of what a subject has already done and therefore involves only past events, these questions were considered inappropriate. Although the standard hypothetical dilemma included in this interview does not fall into the category of a past event, no other approach to providing standardized data was considered possible. Because these questions were deleted, no comparisons between the two interviews on information-processing were possible. Tables 4 and 5 list examples of the types of dilemmas subjects generated within each of the four social contexts, including the standardized dilemma, and the standard questions , respectively.  Table 4 Examples of dilemmas generated for the RLINSI across generation and familiarity  Peer - Familiar The subject's good friend becomes angry when the subject neglects to tell a boy in her class that her good friend likes him. Peer - Unfamiliar A classmate copies the work of the subject, gets caught by the teacher, but refuses to admit to the teacher or to the subject that the assignment was plagiarized. Adult - Familiar The classroom teacher loses a special piece of art work of the subject and refuses to admit it or to give the subject any marks for the missing work. Adult - Unfamiliar An adult at the ticket counter of a movie theatre refuses admission to the subject for a PG13-rated movie two days before the subject's thirteenth birthday. Standard Dilemma What would you do if your best friend said to you that you he/she didn't want you to hang around with another good friend of yours anymore?  Table 5 Standard questions and probes of the RLINSI  1. 2. 3.  What was the problem? Why was that a problem? How did you feel? How did (other person) How did you know that you had really solved the problem?  As no manual or scoring method was available for this interview, despite repeated attempts to contact the author, the procedure and criteria for scoring both level and orientation were based on the INS manual. The RLINSI-Level score reflects the subject's average score in perspective-taking ability across the four social contexts and the standardized dilemma. Again, overall scores reflect both pure and transitional stage levels. Orientation was scored as in the INS manual. In the Schultz and Selman study (1989) using the RLINSI, the scoring procedure differed from the procedure used to score the HINSI. First, the dilemmas were separated into the generational contexts of adult and peer. Three developmental scores and one orientation score were given for each context. Scores were given for an individual's predominant, lowest, and highest level of responding. The predominant level score was given twice the weight of the highest and lowest level scores. The Schultz and Selman justification for using this scoring procedure, which differed from the scoring procedure of the HINSI, was "because of its greater frequency of occurrence in the adolescents' interpersonal negotiations," (1989, p. 142). This justification was regarded as weak, mainly due to the fact that although this same kind of reasoning could have been applied to the scoring of the HINSI as well, it was not. Based on this, and in order to make comparisons between scores of the two interviews more comparable, this scoring procedure was not used. Instead, the RLINSI-Level score reflects the average score of all responses, as was done for the HINSI. The RLINSI-Orientation score, as with the HINSIOrientation, reflects the average degree of "other-transformingness" across the different social contexts. This score was taken from the initial telling of the real-life conflicts. Interrater reliability for the averaged overall scores for this interview was high (level scores, r=.84, orientation scores, r=.85). This is similar to the level of reliability found in a previous study (Schultz & Selman, 1989).  Design and Procedure The experiment was a52-factor design with the between subjects variable being the order of assessment (HINSI-RLINSI = Order 1, RLINSI-HINSI = Order 2), and the within-subjects variable being the type of interview. The independent variables were the order of interview and the actual interviews themselves. The dependent variables were the HINSI and the RLINSI scores previously discussed. A l l subjects were interviewed individually in the school after classes had finished for the day. Each subject was randomly placed in one of two groups where either the HINSI was given first, followed by the RLINSI or vice versa. Before the interview began, some background information from the subject (age, sex, grade, and ethnic background) was taken. This time was also used to establish rapport with the child. All subjects were given the following introduction at the beginning of the HINSI: "What I would like to do now is tell you a few very short stories about girls (or boys, depending on the sex of the subject) your age. After I read one story, I would like to ask you some questions about it, to see what you think about it. There are not right or wrong answers, I just want to know how you think, OK? I'm going to tape record our discussion but I wanted to remind you that no one but me will ever be able to listen to the tape, not your parents, your teachers or your friends, OK?" This introduction was changed slightly for the RLINSI. The two interviews were then conducted. During the HINSI, for each of the four social contexts, there were three possible dilemmas. The dilemmas were assigned to each subject randomly. After the interviews were conducted the purpose of the experiment was explained and the subjects were thanked for their participation.  Statistical Analyses The purpose for conducting the experiment was to search for the presence of order effects between the two interviews. It was hypothesized that the HINSI, designed to elicit a subject's highest level of cognitive reasoning, creates a "set" for level and orientation responses and thereby biases the responses on the RLINSI. Further, the hypothesis stated that, when the HINSI interview preceded the RLINSI, scores on the latter interview would be artificially inflated. A n additional hypothesis was that even if no order effects were seen in the overall scores of the interviews, order effects might be discovered for only certain types of dilemmas namely, at the levels of generation and familiarity. These dilemmas were ones which involved specifically, the adult context but not the peer context and for the unfamiliar but not the familiar context. To test these hypotheses, t-tests were first used to examine the difference scores between overall scores of the HINSI-Level and the RLINSI-Level. A significant discrepancy between the difference scores on Order 1 and Order 2, namely, that the average difference score on Order 1 would be significantly smaller than the average difference score on Order 2, would suggest that order effects were present. The data were further analyzed using several repeated measures A N O V A s to test the main effects of order, interview type, and possible interaction effect between the order and interview type on the developmental level scores. The rationale for using repeating measures design arises from the fact that both interviews measured the developmental level achieved by subjects using the same level scores but under two different contexts, hypothetical and real-life. It can be argued that the most obvious way to analyze these data is a 4-way A N O V A with order as the between-subject variable, and type of interview, generation and familiarity as the within-subject variables. This analysis would examine all of the above-stated hypotheses, as well as the possible 4-way interaction among the variables. The data were not analyzed this way for two reasons. First, there were too few data points in some cells for this type of analysis as not all subjects generated data for all types of dilemmas. Second, even if each subject had been able to provide complete data, the sample size was too small to make this analysis meaningful.  Results  This chapter will begin with a report of the descriptive statistics. These statistics are important not only for providing information about the particular sample of this experiment, but also to demonstrate the degree to which this sample is similar to those involved in previous, related research. This has an important bearing on the implications of the overall findings. If the present sample behaved in a way similar to samples involved in past research, then the confidence with which the results can be generalized is increased. However, if the sample of the present experiment is not comparable to previous samples, then more caution must be used when commenting on the relevance of these results to related research. The subsequent discussion will examine the implications of the results from the various statistical analyses performed on the data in order to ascertain whether or not order effects occurred as either a main effect or in interaction with other variables.  Univariate Descriptive Statistics The sample means of this research were similar to those of previous research using the same interview measures. Table 6 presents the descriptive statistics of the sample both combined across age groups and separated by age group. For the overall and the grade 7 sample, the HINSI-Level mean of 1.59 (standard deviation, .22) and 1.67 (standard deviation, .24) compares with HINSI-Level means of 1.61 for subjects aged 11-10 (Fleischer, 1989), 1.57 for subjects ages 13-9 (Schultz & Selman, 1989) and 1.58 for subjects aged 12-11 (Yeates, Schultz, & Selman, 1988) . The mean score of the grade 4 sample of 1.54 (standard deviation, .14) is slightly higher than past findings of 1.38 for subjects aged 9-2 (Yeates, Schultz, & Selman, 1988).  For the RLINSI-Level scores, the present sample's average score of 1.45  (standard deviation, .23) overall and 1.51 (standard deviation .22) for grade sevens is very similar previous research reporting scores of 1.46 for subjects aged 11-10 (Fleischer, 1989). The grade 4 mean for RLINSI-Level of 1.31 (standard deviation, .13) is slightly lower. No previous  research reports scores for either HINSI-Orientation or RLINSI-Orientation and therefore, comparisons are not possible. Because the size of the grade four sample was very small, due to lack of participation, and because the level and orientation scores between the two age groups were similar, it was decided to combine them in order to increase the size of the sample. Justification for combining the groups revolved around the belief that, if order effects exist, they should occur across age groups. Although it would be interesting to analyze the grade four data to see if order effects occur differentially across age groups, the sample was too small to allow this kind of analysis. For these reasons, the data were analyzed for order effects with both age groups combined.  Table 6 Sample means and standard deviations for the HINSI and RLINSI measures, including level and orientation, grade 7 and grade 4 combined and separated  Overall Grade 4 Variable M SD M HINSI - L L59 .22 1.54 RLINSI-L L45 .23 1.31 HINSI - 0 .56 .12 .58 RLINSI - 0 .59 .26 .64 Note. L = level score, possible range from 0 - 3 O = orientation score, possible range from 0  SD .14 .18 .11 .31 -I  Grade 7 M 1.67 1.51 .55 .54  SD .24 .22 .13 .33  To complete comparisons between present and previous research, correlational analyses were completed on level and orientation scores for both the HINSI and RLINSI interviews. The correlation between the scores of the HINSI-Level and the RLINSI-Level was .28, Q< .06. While this is somewhat different from the Schultz and Selman (1989) finding of .45, this difference is not statistically significant, z = .87, n.s. The correlation between the orientation scores of the HINSI-Orientation and RLINSI-Orientation was -.13, n.s. The comparable Schultz and Selman (1989) correlation was .04. As one further link between the present sample and those of the past, the mean score of the HINSI-Level was generally higher than the RLINSI-Level. Fleischer (1989) found that 65% of the subjects had higher HINSI-Level scores while the present study showed that 72% of subjects received higher HINSI-Level scores than RLINSI-Level scores. The above descriptive statistics show that the sample of the present study is comparable to those samples used in past research involving the HINSI and RLINSI.  Order Effects Analyses The first set of analyses performed on the data involved only the level scores of both interviews. Figure 1 displays the HINSI-Level and RLINSI-Level scores when they were administered first and second. Although there was a slight increase in RLINSI-Level score when it was administered second as opposed to when it was administered first, a planned comparison showed that this difference was not significant, t(34) = .95, p_< .20, one-tailed. The data on developmental level scores were further analyzed with a 2 X 2 repeated measures A N O V A with order as the between-subject factor and the type of interview as the within-subject factor. This analysis showed that, overall performance did not differ between the orders, F(l,34) = .05, n. s., overall performance was higher on the HINSI-Level than on the RLINSI-Level, F(l,34) = 9.94, £ < .01, and any interaction between the order and type of interview was not significant, F(l,34) = 1.83, n. s.  Figure 1. Differences between HINSI and RLINSI overall scores when administered first and second. The figure shows no apparent influence of order on level scores for either interview.  In answer to the hypothesis regarding order effects and the context of generation, Figure 2 displays the scores of the HINSI-Level and RLINSI-Level interviews when they were administered first and second for the adult context and peer context separately. For the adult context, a planned comparison showed that RLINSI-Level scores were higher when this interview was administered second than when it was administered first, t(32) = 1.88, £ < .05, one-tailed. These level score data were further analyzed with two 2 X 2 repeated measures A N O V A s with order as the between-subject factor and type of interview as the within-subject factor for the adult and peer context separately. The following results showed that for the adult context, overall performance did not differ between the orders. F(l,32) = .24, n. s., overall performance was higher on the HINSI-Level than on the RLINSI-Level, F(l,32) - 15.15, Q< .001, and the interaction between the order and type of interview was significant, F(l,32) = 4.76, E< .05. A simple effect analysis showed that the performance on the HINSI-Level did not differ when this interview was administered first or second, F(l,63) = 1.39, n. s. For level scores regarding the peer context, the analysis showed that overall performance did not differ between the orders, F(l,34) = .00, n. s., overall performance was higher on the HINSI-Level than on the RLINSI-Level, F(l,34) = 6.45, £< .05, and, the interaction between the order and type of interview was not significant, F(l,34) = .11, n.s. Figure 3 displays the scores of the HINSI-Level and RLINSI-Level interviews when they were administered first and second for the unfamiliar context and familiar context separately. For the unfamiliar context, a planned comparison showed that the RLINSI-Level scores were higher when this interview was administered second than when it was administered first, t(29) = 2.52, £< .01, one-tailed. These level score data were further analyzed with two 2 X 2 repeated measures A N O V A s with the order as the between-subject factor and the type of interview as the within-subject factor, for the unfamiliar and familiar context separately. For the unfamiliar context, this analysis of level scores showed that overall performance did not differ between the orders, F(l,29) = 1.83, n. s., overall performance was higher on the  HINSI  RLINSI  ADULTS  HINSI  RLINSI  PEERS  Figure 2. Score on the HINSI and the RLINSI when administered first or second across adult and peer contexts. The figure shows that level scores on the RLINSI were increased by the preceding HINSI for the adult dilemmas only.  HINSI  RLINSI  UNFAMILIAR  HINSI  RLINSI  FAMILIAR  Figure 3. Score on the HINSI and the RLINSI when administered first or second across unfamiliar and familiar contexts. The figure shows that level scores on the RLINSI were increased by the preceding HINSI for the unfamiliar dilemmas only.  HINSI-Level than on the RLINSI-Level, F(l,29) = 19.75, |2< -001, and the interaction between the order and type of interview was significant, F(l,29) = 4.63, j2< .05. The simple effect analysis showed that the performance on the HINSI-Level did not differ when this interview was administered first or second, F(l,57) = .07, n. s. For the familiar context, the analysis of level scores showed that overall performance did not differ between the orders, F(l,33) = .84, n. s., overall performance was higher on the HINSILevel than on the RLINSI-Level, F(l,33) = 7.63, n < .001, and the interaction between the order and type of interview was not significant, F(l,33) = . 15, n. s. In regard to the effect that order may have on orientation scores, no hypothesis was stated. A 2 X 2 repeated measures A N O V A was run on the overall orientation scores. The analysis showed that overall performance did not differ between the orders, F(l,34) = .04, n. s., overall performance was on the HINSI-Orientation was not higher than on the RLINSIOrientation, F(l,34) = .51, U.S., and the interaction between the order and type of interview was not significant, F(l,34) = 1.85, n.s.  Discussion  This study was designed to search for possible order effects in the administration of two measures of social development, the HINSI and the RLINSI. More specifically, it was hypothesized that the HINSI would artificially inflate the scores of the RLINSI. If significant effects were found, it would justify the need for caution when interpreting the scores of these two measures when used together. More importantly, it would change the impact that these scores may have in regard to diagnosis and/or treatment of those identified as having difficulties in their social development. In terms of overall scores, no order effects were observed although it appears from the data that a basic trend in that direction exists. While it is only conjecture at this point, future research using a larger sample and thereby yielding greater power, might well observe a stronger, significant extension of these trends. A closer look at the effect of order on more specific social contexts of the interviews revealed statistically significant order effects within the adult and unfamiliar contexts. The difference created by order within the contexts of adults and unfamiliar others was .15 and .24, respectively. While these findings were statistically significant, the question arises as to whether such differences are clinically significant. The implications that order effects may have for clinical purposes are twofold. First, when scores on the RLINSI-Level are artificially inflated by the preceding HINSI, one may be led to believe that a child or adolescent is functioning within normal limits in terms of interactions with adults or unfamiliar people. However, when the RLINSI is administered first, scores will be lower and may drop to a point which merits clinical concern. One of the benefits emphasized by the INS model is its ability to assess an individual's level of functioning within certain social contexts. The effect of order on the contexts of adults and unfamiliar others makes such identiflcation difflcult, mainly because the person evaluating the responses to the RLINSI will be unsure whether the higher or lower scores paint a more accurate picture of the individual.  The second set of implications that order effects have for clinical purposes, is in identifying the size and nature of the gap that exists between thought and self-reported action. In the Schultz and Selman study (1989), a difference of .25 of a level, approximately one standard deviation, was used to establish a gap between thought and self-reported action. In the present study, the difference between RLINSI-Level scores in Order 1 and Order 2, in the adult context amounted to a difference of .16 of a stage which did not equal one standard deviation (Order 1 sd = .23, Order 2 sd = .25). The difference between RLINSI-Level scores in Order 1 and Order 2, in the unfamiliar context, however, amounted to a difference of .24 of a stage which was approximately equal to one standard deviation (Order 1 sd = .23, Order 2 sd = .29). According to the criterion set by Schultz and Selman (1989), this could make the difference in the type of profile an individual receives. For example, an individual receiving the two interviews in the traditional order (HINSI-RLINSI) may be identified as displaying a thoughtself-reported action gap that was relatively small, similar to the Type 3 profile previously described. However, if this same individual was interviewed in the reverse order (RLINSIHINSI), this thought-self-reported action gap would be significantly increased and the individual might be classified as having a Type 1 profile. Figure 4 shows the relative differences created by order in the unfamiliar context. While the implications this may have for treatment (i.e., treating them as having a Type 3 profile versus having a Type 1 profile) are not ones that will foreseeably cause further harm for an individual with existing social difficulties, they may well affect the efficacy of such treatments, and thereby limit the amount of confidence that therapists may have in the INS model. The seriousness of order effects in this regard will only be illustrated when further research is conducted on the effects of therapy based on the INS model. No significant results, at the overall level, were found in regard to the effect of order on orientation. Further analyses on orientation were not conducted for reasons previously described. To briefly reiterate, it is difficult to tell at this point in time whether one mode of orientation is necessarily better than the others. Rather, such qualitative judgements should occur only in conjunction with the known developmental level of an individual which will  ! •  1st  2nd  •  1.7i  HINSI-RLINSI  RLINSI-HINSI  ORDER  Figure 4. Differences in the gap between thought and action level scores between the two orders within the context of unfamiliar others. The figure shows that the gap is considerably larger when the RLINSI is administered first.  indicate tlie sophistication of the reasoning behind a given strategy or self-reported action. Because so little is known about the construct of orientation in the INS model, it would be very difficult to say which way the hypothesized "set" created by the preceding interview would influence scores on this dimension. Figure 5 depicts the pattern of results found on orientation. It is difficult to interpret what this pattern indicates. It may be that, if orientation is subject to the influence of order, the influence will come more from the demand characteristics of the interview technique itself, rather than from a preceding interview. Such demand characteristics might come from a perceived need by an individual to satisfy the interviewer by consistently responding with what he or she believes to be an acceptable orientation. Aside from the fact that the effects of demand characteristics were outside the domain of this experiment, more research is needed before any hypotheses can be formulated regarding the orientation with which an individual might feel obligated to respond. It would have been informative to assess whether order effects occur at more specific levels of the interviews, namely at the level of unfamiliar adults. Unfortunately, the small sample and the inability of some subjects to generate real-life conflicts in this specific context, led to a lack of data and prevented further analysis of this kind. One of the important questions that remains unanswered is why the preceding HINSI had a significant effect on dilemmas involving adults and unfamiliar others in the RLINSI. There are several areas that could be investigated in future research. One line of reasoning might suggest that the standard questions and probes given in the HINSI created a format which somehow allowed the subjects to organize their thoughts and thus, relate their own, personal conflicts in a more sophisticated manner. Once the subjects knew the direction that the questions and probes were taking, this may have cued them to respond with higher levels of reasoning for the RLINSI, once they had had the experience of answering the questions and probes of the HINSI. Another line of reasoning would suggest that the type of dilemmas used in the HINSI somehow influenced the subjects in the type of conflicts they chose to relate, and were possibly ones in which they had used a higher level of reasoning. The HINSI dilemmas, by  Figure 5. Differences between HINSI and RLINSI overall orientation scores when administered first and second. The figure shows that orientation scores were not influenced by order.  and large, are ones which avoid content which is heavily emotionally-laden. Recent research using the INS model (Schultz & Selman, 1989; Fleischer, 1989; Yeates et al., 1990) has begun to examine the role of emotion in interpersonal negotiations and has generally found that emotionally-laden conflictual situations tend to lower the level with which an individual will react in a real-life situation. If, in fact, the HINSI dilemmas prompt individuals to relate similar, real-life situations, these conflicts would be less emotionally-laden and thereby may function to increase the level with which they respond. A third and final line of reasoning to explain how order effects influence individuals is slightly different in its perspective. It may not be the HINSI specifically that is creating the order effects at all but merely the presence of some initial discussion of interpersonal negotiations that influenced responses on the RLINSI. This would suggest that the RLINSI itself, as an interview technique, is not robust. Interestingly, it seems that the researchers following the INS model have begun to favour use of the INS Rating Scale mentioned previously, when trying to identify levels of social behaviour. This may be due to the stronger correlations the rating scale has with the HINSI (Fleischer, 1989; Yeates et al., 1990). It will be interesting to see if future research continues to reject use of the RLINSI in favour of the INS Rating Scale. As a final note, it is difficult to ascertain which scores rendered by the RLINSI depict the true state of affairs of an individual. Does the preceding HINSI allow an opportunity for "practice" that individuals may need to more accurately relate their real-life conflicts, or does the RLINSI, when administered first, capture the more accurate observations of an individual's reallife behaviour? While this question remains unanswered for the time being, there is one idea which is certain: more research involving the psychometric properties of the RLINSI must be conducted if its use is to be continued. While the field of psychology abounds with tests and assessment measures, there are few which stand up to rigorous tests of reliability and validity. It is the responsibility of researchers and therapists alike, to ensure that such measures are sound, both psychometrically and clinically, before they are used in applied settings.  References  Abrahami, A., Selman, R.L. & Stone, C. (1981). A developmental assessment of children's verbal strategies for social action resolution. Journal of Applied Developmental Psychology, 2, 145 - 163. Beardslee, W.R., Schultz, L.H., & Selman, R.L. (1987). Level of social-cognitive development, adaptive functioning, and DSM-III diagnoses in adolescent offspring of parents with affective disorders: Implications of the development of the capacity for mutuality. Developmental Psychology, 23(6), 807-815. British Columbia Ministry of Education. (1989). Enabling Learners: Year 2000 A Framework for Learning. British Columbia: British Columbia Ministry of Education. Christensen, L . B . (1988). Experimental Methodology (4th ed.). Toronto: Allyn and Bacon, Inc. Conger, L C . & Conger, A . J. (1982). Components of heterosocial competence. In J.P. Curran & P . M . Monti (Eds.), Social Skills Training (pp. 313-347). New York: Guilford. Cowen, E.L., Pederson, A., Babigian, H., Izzo, L.D. & Trost, M . A . (1973). Long-term followup of early detected vulnerable children. Journal of Consulting and Clinical Psychology, 42(3), 438-446. Dillard, K . & Spitzberg, B . H . (1984). Global impressions of social skills: Behavioral predictors. Communication Yearbook, 8, 446-463. Dodge, K . A . , Schlundt, D.C. Schocken, I. & Delugach, J.D. (1983). Social competence and children's sociometric status: The role of peer group entry strategies. Merrill-Palmer Quarterly, 29 309-336. Feldman, N.S., Klosson, E.C., Parsons, J.E., Rholes, W.S. & Ruble, D.N. (1976). Order of information presentation and children's moral judgments. Child Development, 47, 556-559. Fleischer, L . (1989). Intrapsychic process, interpersonal reasoning, and the development of mutuality in préadolescent friendship. University Microfilms International, 9014303.  Gold, R. (1987). Class inclusion performance: Effect of intermingling the subsets. British Journal of Developmental Psychology, 5, 343-346. Gresham, P.M. (1985). Utility of cognitive-behavioral procedures for social skills training vv'ith children: Acriticial review. Journal of Abnormal Child Psychology, 13(3), 411-423. Grueneich, R. (1982). The development of children's integration rules for making moral judgments. Child Development, 53, 887-894. Gurucharri, C , Phelps, E., & Selman, R.L. (1984). Development of interpersonal understanding:A longitudinal and comparative study of normal and disturbed youths. Journal of Consulting and Clinical Psychology, 52(1), 26-36. Hughes, H . M . (1984). Measures of self-concept and self-esteem for children ages 3-12 years: A review and recommendations. Clinical Psychology Review, 4(6), 657-692. Kohlberg, L. (1969). Stage and sequence: The cognitive-developmental approach to socialization. In D. Goslin (Ed.), Handbook of Socialization Theory and Research (pp. 347380). Chicago: Rand McNally. Lyman, D.R. & Selman, R.L. (1985). Peer conflict in pair therapy: Clinical and developmental analyses. In M . W . Berkowitz (Ed.) Peer Conflict and Psychological Growth. New Directions for Child Development, no. 29 (pp. 85-102). San Francisco: Jossey-Bass. Parker, L G . & Asher, S.R. (1987). Peer relations and later personal adjustment: Are lowaccepted children at risk? Psychological Bulletin, 102, 357-389. Parsons, I.E., Ruble, D.N., Klosson, E.C., Feldman, N.S. & Rholes, W.S. (1976). Order effects on children's moral and ^c\ntytmtni']\iùgmtn{sJ)evelopmental  Psychology, 4, 357-358.  Pellegrini, D.S. (1985), Social cognition and competence in middle childhood. Child Development, 56, 253-264. Piaget, J. (1960). The general problems of the psychobioiogical development. In J.M. Tanner & B. Inhelder (Eds.) Discussions on Child Development, (Vol 4, pp. 3-27). London: Tavistock. Piatt, J.J., Spivack, G., Altman, N., Altman, D. & Peizer, S.B. (1984). Adolescent problemsolving thinking. Journal of Consulting and Clinical Psychology, 42, 787-793.  Putallaz, M . & Gottman, J. (1982). Conceptualizing social competence in children. In P. Karoly & J.J. Steffen (Eds.), Improving Children's Competence: Advances in Child Behavioral Analysis and Therapy (pp. 1-33). Lexington, M A : Lexington Books. Reisman, J.M. (1985). Friendship and its implications for mental health or social competence. Journal of Early Adolescents 5(3), 383-391. Schultz, L . H . , & Selman, R.L. (1989). Bridging the gap between interpersonal thought and action in early adolescence: The role of psychodynamic processes. Development and Psychopathology, 1, 133-152. Schultz, L . H . , Yeates, L.O., & Selman, R.L. (1989). The Interpersonal Negotiation Strategies Interview Manual. Unpublished manual. Harvard University. Seigel, M . & Robinson, J. (1987). Order effects in children's gender-constancy responses. Developmental Psychology, 23(2), 283-286. Selman, R.L. (1980). The Growth of Interpersonal Understanding. New York: Academic Press. Selman, R.L. (1981). The development of interpersonal competence: The role of understanding in conduct. Developmental Review, 1, 401-422. Selman, R.L. & Arboleda, C. (1986). Pair therapy with two troubled early adolescents. McLean Hospital Journal, 10(2), 84-111. Selman, R.L., Beardslee, W., Schultz, L.H., Krupa, M . & Podorefsky, D. (1986). Assessing adolescent interpersonal negotiation strategies: Toward the integration of structural and functional models. Developmental Psychology, 22(4), 450 - 459. Selman, R.L. & Demorest, A.P. (1984). Observing troubled children's interpersonal negotiation strategies: Implications of and for a developmental model. Child Development, 55, 288-304. Selman, R.L. & Jaquette, D. (1977). Stability and oscillation in interpersonal awareness: A clinical-developmental analysis. In Keasey, C.B. (Ed.) Nebraska Symposium on Motivation (pp. 262-304). Nebraska: University of Nebraska Press. Selman, R.L. & Schultz, L . H . (1988). Interpersonal thought and action in the case of a troubled early adolescent. Toward a developmental model of the gap. In S. Shirk (Ed.), Cognitive Development and Child Psychotherapy (pp. 207-246). New York: Plenum.  Selman, R.L., Schorin, M.Z., Stone, C.R. & Phelps, E. (1983). A naturalistic study of children's social understanding. Developmental Psychology, 19(1), 82-102. Spitzberg, B.H. & Cupach, W.R. (1989). Handbook of Interpersonal Competence Research. New York: Springer-Verlag. Spivak, G. & Shure, M . (1976). The Problem-Solving Approach to Adjustment: A guide to Reseaarch and Intervention. San Francisco: Jossey-Bass. Yeates, K . O . & Selman, R.L. (1989). Social competence in the schools: Toward an integrative developmental model for intervention. Developmental Review, 9, 64-100. Yeates, K.O., Schultz, L . H . , & Selman, R.L. (1990). Bridging the gaps in child-clinical assessment: Toward the application of social-cognitive developmental theory. Clinical Psychology Review, 10, 567-588.  


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