Families and their Learning Environments

An Empirical Analysis

Kevin Marjoribanks

• Publisher: Taylor & Francis
• Year: 2017
• Print ISBN: 9780415787239
• eBook ISBN: 9781351847261
• Edition: 1
• Topic: Education
• Subtopic: Education General

First published in 1979, this study is one of the first works of educational research to include detailed assessments of family environments in an analysis of performance of children at their schools. Much of the research is based on data collected from families in Australia, Canada and England and the findings have been integrated with results from other family environments research. The study also explores social and psychological conceptual positions that will have relevance for further educational investigations.

This book will be of particular interest to those studying the relationship between family environments and education, as well as the sociology of education.

1 Family environments: frameworks for analysis

Although most educators would agree with the proposition that families are one of the most pervasive influences on children’s behaviours, only a relatively small number of educational researchers include detailed assessments of family environments in analyses of children’s school outcomes. A number of studies that have attempted to increase our understanding of the relations between measures of family environments and children’s academic performance are brought together in this book. The selection of studies, however, is quite idiosyncratic. I have presented much of my own research on data collected from families in Australia, Canada, and England. But in doing so I have endeavoured to integrate my findings with results from other family environment research. Also, I have attempted to explore social-psychological conceptual positions that might be relevant for further educational investigations. In this first chapter, general conceptual and statistical frameworks for examining family environments and their associations with children’s behaviours are presented, followed by a review of some of the seminal family environment studies from education-related research.

Conceptual framework for analysis

In social-psychological and educational empirical research, children’s behaviour has typically been examined in relation to one of the following three models: (a) the trait model, (b) the situationism position, and (c) the interactionism framework. In the trait model, factors determining behaviour are considered to be within the persons themselves. As Endler and Magnusson (1976, p. 6) suggest, in the ‘classical’ trait theory:

It is sufficient to merely study the individual. Situations are taken into account, but the provoking and restricting effect of situational factors on behaviour is not supposed to change the rank orders of individuals for a given trait. This means that the rank order of individuals for any given trait is supposed to be the same for different situations independent of the situational characteristics, except for errors of measurement.

Alternatively, the situationism position regards situational factors, or the stimuli in the situation, as the main determinants of individual behaviour. In some of the following chapters, a situationism position is adopted to investigate associations between children’s academic achievement and different measures of family environments. It is the basic assumption of the book, however, that behaviour is the result of an indispensable, continuous interaction between persons and the situations they encounter. That is, children’s school outcomes are explained most appropriately by examining an interactionism framework of behaviour. In this latter model it is assumed that not only do situations influence individuals but that individuals select and subsequently affect the situations with which they interact.

One of the earliest social-psychological forms of the interactionism framework was developed by Lewin in his field theory of personality. He proposes that ‘the dynamics of environmental influences can be investigated only simultaneously with the determination of individual differences and with general psychological laws.’ Therefore, ‘to understand or predict the psychological behavior (B) one has to determine for every kind of psychological environment … the momentary whole situation, that is, the momentary structure and the state of the person (P) and of the psychological environment (E). B = f(PE)’ (Lewin, 1935, p. 79). As Moos (1976, p. 20) says, it ‘made no sense to Lewin to analyze behavior without reference to the person and to his psychological environment’.

In a special adaptation of the interactionism framework, which is examined in Chapters 6 and 7, Thomas and Znaniecki (1958) propose that ‘the social results of individual activity depend, not only on the action itself, but also on the social conditions in which it is performed; and therefore the cause of a social change must include both individual and social elements’. They declare that the fundamental methodological principle of both social psychology and sociology is: ‘The cause of a social or individual phenomenon is never another social or individual phenomenon alone, but always a combination of a social and an individual phenomenon’ (see Thomas, 1966, p. 277).

In Lewin’s framework, the environment surrounding individuals is considered to be differentiated into regions. The regions may be defined along a continuum of ‘nearness-remoteness’ which is an indication of the extent of influence that one environmental region has on another. Also the boundaries of environment regions may be defined along the continua of ‘firm-weak’ and ‘fluid-rigid’. If boundaries are firm then the environment regions have minimal interaction; but if the boundaries are weak, then the regions may exert considerable influence on each other. A fluid region is one that responds quickly to any influence brought to bear upon it, while a rigid medium resists change (see Lewin, 1935; Deutsch and Krauss, 1965; Hall and Lindzey, 1970). The environment constructs introduced by Lewin are adapted in the present book to suggest that the family environment surrounding an individual may be conceived of as a set of nested environments, ranging from: distal social environments and distal family structure variables to proximal social-psychological and socio-linguistic environments.

In the book the distal social environments of children are defined by social-status characteristics and ethnic group membership, while distal family structure is assessed by family size, birth order, and crowding within the home. Although these two sets of environment variables are defined as distal situations, it is shown in Chapters 3 and 4 that family environment research becomes more sensitive, enriched, and valid when children’s school outcomes are examined not only in relation to proximal social-psychological measures but also in relation to global social environment characteristics.

Much of the impetus for educational research on family social-psychological environments was provided by Bloom (1964), who proposed that the environment may be regarded as providing a network of forces and factors which surround, engulf, and play on the individual. It is also suggested that the development of any particular human characteristic is related to a sub-set, or sub-environment, of the total set of environmental forces. In Figure 1.1, if A B C D represents, for example, the total set of family conditions surrounding an individual, then E F G H may represent a sub-environment of social-psychological process variables related to, say, the development of verbal ability. And JKLM may represent another sub-environment associated with, say, the development of achievement motivation. Then the task of researchers investigating relations between family environments and a particular human characteristic involves isolating, and then measuring, the pertinent sub-environment from the total set of environmental variables. It is shown later that families from different social-status and ethnic groups construct divergent sub-environments for children’s academic achievement, and that between social groups the boundaries of environment regions may have quite variable qualities.

FIGURE 1.1 Sub-environments of family environments

Within interactionism models of behaviour it is typically proposed that social conditions should be defined in relation to the meanings that the contexts have for individuals. It is suggested by Karabel and Halsey (1977, p. 58), however, that emphasis on the meanings that social conditions have for an individual ‘often fails to take adequate account of the social constraints on human actors in everyday life’. They propose that in the examination of behaviour there is a need to integrate structural and interactional levels of analysis. The research presented in this book is within the framework of structural analyses of family environments. Detailed and sensitive studies, using an interpretative methodology, of the relations between academic performance and the meanings that family environments have for children from different social status and ethnic groups, have yet to be completed. Moos (1979) proposes, however, that person variables such as ability levels, cognitive and emotional development, ego strength, and self-esteem determine the meaning that an environment has for an individual. In subsequent chapters, studies are examined that investigate relations between academic achievement and measures of attitudes, personality, and intelligence at different levels of social-psychological family environment variables. These person variables may be considered to represent, in part, the results of interactions between children’s perceptions of prior family environments and individual characteristics, and thus reflect, at least tentatively, the meanings that family environments have for children. But such measures are an initial approximation only of these meanings. Educational research is in need of investigations, using an interpretative framework, which examine relations between family learning environments and children’s behaviour. Such research will provide data to embellish the findings from studies that observe the structural elements of families.

Ideally, research adopting an interactionism framework would incorporate a dynamic or organismic rather than a reactive or mechanistic model of the individual (see Harré and Secord, 1972). In mechanistic models, interactions are between causes rather than between chains of causes and effects. That is, the mechanistic framework is concerned with uni-directional causality and it assumes that independent variables influence dependent measures. Endler and Magnusson (1976, p. 13) indicate that the dynamic model ‘is concerned with a reciprocal action (feedback) between environmental events and behavior. It is concerned with reciprocal causation so that not only do events affect the behavior of organisms, but the organism is also an active agent in influencing environmental events.’ But they go on to suggest that in the social sciences we have not yet developed fully the methodology and technology to investigate the nature of dynamic interacton. While some of the reported research in later chapters examines longitudinal data and approximates a dynamic model of behaviour, most of the studies investigated reflect a mechanistic framework. Although it is assumed throughout the book that situations are as much a function of the person as the person’s behaviour is a function of the situation, the research supports Endler’s (1976, p. 66) contention that ‘Concurrent with an investigation of dynamic interaction we must also examine how persons and situations interact (mechanistic interaction) in influencing behaviour’.

Statistical framework for analysis

Much of family environment research has relied on the use of restricted statistical techniques such as product-moment correlations which reveal only bivariate relations, and analysis of variance techniques that require the grouping of variables into levels. Regression models have been used, which concentrate on determining how much variance in children’s behaviour is related to the addition of family measures to the equations. But the findings generated from these regression models depend on the order in which family environment is entered into the models (e.g., see Coleman et al., 1966; Mosteller and Moynihan, 1972). Also, the regression models generally have not tested for the presence of interaction or curvilinear relations among the variables. In this book a variation of regression analysis is used in many of the studies that are examined. Regression surfaces are plotted using raw regression weights generated from equations of the form: Z = a X + b Y + c X Y + d X² + e Y² + constant, where Z represents a measure of children’s behaviour and X and Y are measures of family environments and person variables. Regression surface analysis allows a figural presentation of the relations between variables and is a useful methodology for testing an interactionism framework of analysis.

In the studies using regression surfaces simple random samples of families were not selected. Therefore in the analyses the design effect for each raw regression weight is generally estimated, using the Jack-knife technique (Mosteller and Tukey, 1968; Finifter, 1972). In the calculations the data on the families are divided randomly into sets of sub-samples, and then pseudo-values for the regression weights are computed by conducting the regression analyses on different combinations of sub-samples. Significance levels for the regression weights were recalculated using the formula: standard error of sample estimate = (design effect)½ × simple random standard error (Kish, 1965; Ross, 1976). Typically the application of the Jack-knife technique led to a reduction in the number of regression weights that were significant. Often the adjusted standard errors showed that the interaction and curvilinear terms were not significant. In such cases, a second stage of the regression analysis was conducted in which variables that no longer had significant associations with the measures of children’s achievement were deleted from the regression models. In the second analysis, the Jack-knife technique was used again to estimate design effects and adjust further the significance levels of the regression weights.

Seminal family environment studies

In the following section of the chapter a set of family environment studies is reviewed that have adopted various conceptual and methodological frameworks of analysis. The studies are considered to represent some of the most significant family environment research in educational literature. They have been categorized as either social-psychological or socio-linguistic family environment investigations. Within the former category, studies are divided further into those that have adopted variations of ‘traditional’ social science research methods (Burks, 1928; Rosen, 1956, 1959, 1961; Keeves, 1972, 1974), and those using intensive analyses of small samples of families (Kahl, 1961; Strodtbeck, 1958). The socio-linguistic studies include those labelled as: early-Bernstein, Hess and Shipman, Labov, and later-Bernstein.

Social-psychological family environments: ‘traditional’ research methodologies

1. In a Stanford doctoral dissertation presented in 1928, Burks completed a study of the relations between family environments and children’s intelligence test scores, which for its sophistication of sampling design and measurement has rarely been replicated. Burks indicated that ‘The investigation in hand approaches the aspect of the problem which concerns heredity and home environment through a comparison of mental test resemblances obtaining between parents and children on the one hand, with those obtaining between foster parents and their foster children on the other’ (p. 221). The sample included 214 children who were placed with foster parents before the age of 12 months and a control group of 105 children and their natural parents. Ages of the children ranged from 5 to 14 years.

A comprehensive set of measures was adopted for the study, including intelligence tests, schedules to assess affective characteristics, and family environment questionnaires. Each family participated for a period of four to eight hours. Parents and children completed the Stanford Binet test. Also included were a home information blank, a personal information blank, the Woodworth-Cady questionnaire to test for emotional stability, and a statement by parents on their children’s character. These scales are described below.

(a) The home-information blank was completed by the researchers; it consisted of the Whittier Scale for Home Grading and a culture scale constructed for the study. In the Whittier Scale five aspects of a home are evaluated, including: necessities, neatness, size, parental conditions, and parental supervision. Each of these aspects is scored on a five- or six-point scale. A score of six, for example, on parental supervision indicates that ‘care given the children and provision made for their welfare very exceptional’ while a two-point score suggests that there is a ‘lack of discipline because parents away large proportion of time. No partiality (in dealing with children) as far as known. Parents good, hard-working people.’ One further example is provided to indicate the nature of the Scale. A necessities score of one, assessed a family as having: ‘Wages of driver of small express and transfer wagon. Old ragged dirty clothes. Little food, very plain. Three small rooms in basement of cheap tenement house. Hardly bare necessities. Old, cheap, broken, wooden chairs and tables. No pictures or decorations. Bare floors. No comforts or improvement,’ wh...

Table of contents

• Cover
• Half Title
• Title Page
• Copyright Page
• Table of Contents
• Acknowledgments
• 1 Family environments: frameworks for analysis
• 2 The ‘Chicago’ and ‘British’ schools of family environment research
• 3 Social status and children’s outcomes: analytic models for research 55
• 4 Sibling variables: insignificant or meaningful family environment measures
• 5 Families, schools and children’s outcomes
• 6 Family environments, person variables and children’s achievement: an interactionist analysis
• 7 Ethnicity, family environment, and children’s achievement
• 8 Socialization influences on children’s outcomes: conclusions 178
• Appendix: Family environment schedule
• Bibliography
• Author index
• Subject index