16

Click here to load reader

Socialization and Attribution Processes: Actual versus Perceived Similarity among Parents and Youth

  • Upload
    vern-l

  • View
    215

  • Download
    0

Embed Size (px)

Citation preview

Page 1: Socialization and Attribution Processes: Actual versus Perceived Similarity among Parents and Youth

Socialization and Attribution Processes: Actual versus Perceived Similarity among Parents andYouthAuthor(s): Alan C. Acock and Vern L. BengtsonSource: Journal of Marriage and Family, Vol. 42, No. 3 (Aug., 1980), pp. 501-515Published by: National Council on Family RelationsStable URL: http://www.jstor.org/stable/351895 .

Accessed: 03/08/2014 11:48

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .http://www.jstor.org/page/info/about/policies/terms.jsp

.JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range ofcontent in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new formsof scholarship. For more information about JSTOR, please contact [email protected].

.

National Council on Family Relations is collaborating with JSTOR to digitize, preserve and extend access toJournal of Marriage and Family.

http://www.jstor.org

This content downloaded from 82.40.156.237 on Sun, 3 Aug 2014 11:48:52 AMAll use subject to JSTOR Terms and Conditions

Page 2: Socialization and Attribution Processes: Actual versus Perceived Similarity among Parents and Youth

Socialization and Attribution Processes: Actual Versus Perceived Similarity

Among Parents and Youth*

ALAN C. ACOCK** University of Oklahoma

VERN L. BENGTSON*** University of Southern California

In the socialization of attitudes, which is the better predictor of youths' orientations: what parents actually think (stated attitudes) or what their children think they think (attributed attitudes)? Mother, fathers, and youths from 466 family triads stated their own opinions on nine political and religious questions; youths were also asked to predict their fathers' and mothers' responses (attributed attitudes). Results of regression analysis support symbolic interactionists' and attribution theorists' argu- ments regarding attitude socialization. Stated opinions of parents have little direct effect on youths' orientations except as they are perceived (attributed) by their chil- dren. Youths perceive mothers and fathers as more conservative or traditional than their parents'stated responses justify. Youths greatly exaggerate parental agreement. This "polarized misattribution" is independent of the gender of the youth; it reflects the "generational stake" of each cohort involved in socialization interaction.

During the past two decades, the degree of attitude contrast or congruence between par- ents and their young adult offspring has been a topic of considerable debate. Popular litera- ture has repeatedly postulated the reality of a "generation gap" between parents and post- adolescent youth in attitudes and values. At- tempts by social scientists to examine the validity of this popular belief have produced many contradictions, as is suggested in the

propositional review by Troll and Bengtson (1978). Some research implies considerable attitude-behavior differences between parents and their children (Connell, 1972; Gallagher, 1974; Payne et al., 1973; Tedin, 1974). Other research suggests generally high levels of inter- generational similarity (Aldous and Hill, 1965; Hyman, 1969; Troll et al., 1969; Hill et al., 1970; Kandel and Lesser, 1972; Thomas, 1974; Bengtson, 1975; Lerner and Knapp, 1975; Acock and Bengtson, 1977; Acock and Bengtson, 1978).

Family interactions, especially those in- volving the socialization of attitudes and behaviors, can be viewed as an application of one of the most general questions in social psychology: the extent to which one person (X) influences the behavior or orientations of another (0).

The family is a particular type of small group, with special functions (such as sociali- zation) which involve both long-term inter- personal influence as well as continual

*This study is a collaborative effort and the order of names does not signify unequal responsibility or contri- bution. The research has been supported by the National Institute of Mental Health (Grant #MH-18158). We ac- knowledge with gratitude the comments of Professors Gunhild Hagestad and Neal Cutler to an earlier draft of this paper and the computer assistance to Mitsuyo Hanada.

**Department of Sociology, University of Oklahoma, Norman, Oklahoma 73019.

***Laboratory for Social Organization and Behavior, Andrus Gerontology Center, University of Southern Cali- fornia, Los Angeles, California 90007.

August 1980 JOURNAL OF MARRIAGE AND THE FAMILY 501

This content downloaded from 82.40.156.237 on Sun, 3 Aug 2014 11:48:52 AMAll use subject to JSTOR Terms and Conditions

Page 3: Socialization and Attribution Processes: Actual versus Perceived Similarity among Parents and Youth

negotiations regarding conformity and change between generations (Bengtson and Black, 1973). Of specific interest is the path of influ- ence in the "social construction of reality." One line of theory-that of the behaviorists- suggests it is the actual opinions or goals of X which constitute the model, or influence. A second, however, suggests it is the orienta- tions attributed to him/her by O which are more predictive of O's orientations; these are the more "cognitively" oriented theorists, for example, symbolic interactionists.

BACKGROUND AND FOCUS OF THE RESEARCH

In an attempt to resolve contradictions in the available studies concerning socialization outcomes, we have phrased a somewhat dif- ferent research question. Rather than ask, "To what extent is the generation gap real?" we ask, "Where is the reality of the genera- tion gap?" We hypothesize two processes to account for the apparent inconsistencies in the literature. First, to the extent a genera- tion gap exists, it is greater in the intrafamil- ial attributions of the parents and children than in their stated or actual attitudes. In other words, the perceived generation gap is far greater in the minds of the two genera- tions than in their stated attitudes and opin- ions. This suggests that, if we compare the stated opinions of parents and their children, we should observe less difference than if we compare the stated opinions of the children to the perceptions the children have about their parents' opinions.

The second process involves the location of the children on a particular attitude dimen- sion relative to their own generation. We hypothesize that the relative location of the child depends more on the child's perception of his or her parents' position than on what the parents actually believe. This second process takes into account the child's opinion relative to other members of the child's generation. The first process, by contrast, at- tempts to account for the absolute level of agreement between parent and child.

The distinction between these two pro- cesses is important in evaluating intergenera- tional research, since various studies have measured either the stated opinion of the parent or the child's perception of that opin- ion, but rarely both. The stated opinions of the parents are important to the socialization

of children since some studies suggest chil- dren end up with similar stated opinions, through processes termed identification, modeling, or imitation. The perceptions (or misperceptions) the children have about their parents' opinions are also important to the socialization process. These perceptions are important in determining how liberal or con- servative the child is relative to his/her own generation.

While the distinction between stated and perceived attitudes has not been clearly stated in the intergenerational literature, several strains of behavioral theory suggest its poten- tial relevance. Fifty years ago W. I. Thomas (1972) suggested that humans base an action on their "definition of the situation" which constitutes the ultimate reality of their action. Symbolic interactionists (e.g., Stryker, 1965; Kinch, 1963) have consistently considered that people are influenced directly by the meaning (attribution) they give to the responses of others and only indirectly by the actual responses of others. Cognitive theorists appear to share the basic assumption that the actual attitude and behavior of the Other are "peripheral" factors in comparison to per- ceptions of them. The behavior or attitude of Other becomes meaningful and has conse- quences only as it is perceived or defined by the individual actor (see discussion in Shaw and Costanzo, 1970: 277-285; 291-294). At- tribution theorists, currently occupying a major role in social psychology, place special emphasis on the importance of attributed attitudes, relative to actual orientations held by others, in models of interpersonal influ- ence (Sherif et al., 1965; Jones and Harris, 1967; Kelley, 1973).

Given this thrust by so many branches of social scientific thinking, it is puzzling that so many researchers have continued to empha- size actual attitudes or behaviors rather than perceived or attributed orientation in their investigations. Researchers in reference groups have either used the modal or mean attitude score of the reference group (e.g., Shibutani, 1955) or they have been somewhat defensive about using perceived reference group attitude (Acock and DeFleur, 1972). In family research, the limitation of using only estimates or attributions of other family members' orientations has been pointed out with respect to intergenerational comparisons (Cutler, 1977a; Troll and Bengtson, 1978).

502 JOURNAL OF MARRIAGE AND THE FAMILY August 1980

This content downloaded from 82.40.156.237 on Sun, 3 Aug 2014 11:48:52 AMAll use subject to JSTOR Terms and Conditions

Page 4: Socialization and Attribution Processes: Actual versus Perceived Similarity among Parents and Youth

There is at present little empirical evidence to support-or disconfirm-the key assump- tion of cognitive theorists that it is the per- ceived orientation that constitutes the pri- mary predictor variable. In family research, where many surveys have relied on the report of only one family member for characteriza- tions of Others' opinions, this appears to be a particularly crucial issue to test.

Empirical evaluation of possible socializa- tion consequences of actual versus perceived orientations leads to a series of interrelated research questions in examining various intrafamilial comparisons: mother-child, father-child, and father-mother (see Figure 1). These may be symbolically represented as chains or pathways of influence between the actual (i.e., stated) response of the mother (AM), the actual response of the father (AF), the perceived response of the mother (PM), the perceived response of the father (PF), and the actual response of the child (AC). Five analytical questions may be formulated re- flecting various chains of influence:

1. Actual versus perceived orientations: which are the more effective predictors of responses of post-adolescent children: (a) the actual opinions of mothers and fathers (AM--AC and AF AC); or (b) the perceived parental opinions (PM* AC and PF-AC)? How high is the level of parental prediction in each of these influence chains?

2. Accuracy of child's attribution: to what extent are the perceived parental atti- tudes (as estimated by the child) similar to the actual parental responses? How predictive are the links AM- PM and AF- PF?

3. Perceived and actual similarity of the parents: do children perceive greater similarity between the parents (PM*- PF) than actually exists (AM-AF)?

4. Maternal versus paternal predictivity: are the orientations of one parent more predictive of the child's opinions than the orientations of the other parent? Does it matter whether these are actual

FIGURE 1. PATH MODEL OF CHAINS OF INFLUENCE (AS DETAILED IN TABLES)

e

PM (Perceived

Respon•ne

of Mother)

7,3

a3 $53

(Actual Response AM

.7X- ------5• of Mother)

r12

2•3

AC (Actual Response

(Actual of Child) Response AF X of Father)

$42 $54

4 PF (Perceived Response of Father)

August 1980 JOURNAL OF MARRIAGE AND THE FAMILY 503

This content downloaded from 82.40.156.237 on Sun, 3 Aug 2014 11:48:52 AMAll use subject to JSTOR Terms and Conditions

Page 5: Socialization and Attribution Processes: Actual versus Perceived Similarity among Parents and Youth

or perceived orientations? Are the links (AM--AC and PM--AC) stronger than the links (AF--AC and PF--AC)?

5. Cohort contrasts in attitudes: at the group (cohort) level of analysis, how dif- ferent or similar are the means of the parents (actual and perceived) com- pared to the means of the youth sample? Can misperceptions be traced to the child perceiving the parental attitudes as more divergent, or more similar, to their own than they really are? This is a different way of stating the problem of attribution between generations, using group means of lineage (within-family) comparisons.

STATISTICAL DESIGN

The first four research questions can be addressed in terms of three standardized regression equations (path equations) where data are available from a sample of lineage- related mothers, fathers, and post-adolescent youth:

X = 331X + e3, (1.1) X4= l42YX2 + e4, (1.2) X= p3sX1 +5 p2X2 + p,3x3 + psdX4 + es, (1.3) where:

X, = actual response of mother (AM), X, = ictual response of father (AO),

X3 = perceived response by child of mother (PM),

X4 = perceived response by child of father (PF), X5 = child's own response (AC), Pj = a standardized regression coefficient (path

coefficient) of the effect of variablej on vari- able i, and

ei = random error associated with variable i.

This system of equations provides direct answers to each of the first four research questions. The fifth research question is addressed by a comparison of group means and standard deviations. To examine possible sex-lineage contrasts, each analysis has been replicated for the sub-samples of male and female children.

Equation 1.3 indicates two chains of influ- ence on the children (see Figure 1) derived from the two strains of theory reviewed above. The first chain comes from the parents' actual attitudes (X, and X2). According to a strictly behavioristic perspective, these may be postu- lated to have direct effects on the child's opin- ion.

A second path of influence comes from the perceived or attributed opinions of the mother and father (X3 and X4). According to interactionists or attribution theorists, par- ents' attitudes would affect their children in- directly, through the perception of the chil- dren. Thus, from this point of view, much of the impact of the actual opinions of the par- ents should be mediated by the children's perceptions.

METHODS The data for this analysis are part of a

larger study of three generations drawn from a population of 840,000 members of a Los Angeles area medical care plan (Bengtson, 1975). The population represents a predomi- nantly working- to middle-class background. We will use only the middle-aged parents and their young adult children, the second (G2) and third (G,) generation of respondents in this research program.

From the 2,044 cases in the initial three- generation sample (response rate 70 percent to a self-administered questionnaire mailed in 1972), a subsample of 446 parent-child triads (father-mother-child) were selected in such a way as to maximize the number of cases (triads). Although each triad is unique, some individuals are represented in more than one triad; where more than one youth (age 16-26) from a family responded, both youths are included in the analysis. Thus, two triads may include the same father-mother dyad.

The logic of using triads which involve rep- lications of individuals requires justification. Such sampling-with-replacement may attenu- ate the variance in the parental scores (lead- ing to a lowering of correlations and signifi- cance levels) and may raise questions about sample representativeness. We chose this strategy for two reasons. First, each triad is still a uniquely defined father-mother-youth unit. This reflects the observation noted in qualitative family studies that each parent- child unit is unique; that parent-child rela- tions and influences vary even within the same family. Second, the logic allows us to avoid the potentially more serious sampling bias of underrepresenting large families, in which there is more than one youth in the 16- 25-year-old age range.

504 JOURNAL OF MARRIAGE AND THE FAMILY August 1980

This content downloaded from 82.40.156.237 on Sun, 3 Aug 2014 11:48:52 AMAll use subject to JSTOR Terms and Conditions

Page 6: Socialization and Attribution Processes: Actual versus Perceived Similarity among Parents and Youth

However, it is necessary to test empirically the potential bias that may have resulted from our sampling logic. Thus, each statis- tical analysis presented in the following sec- tions was replicated on a subsample com- posed of "non-replicated" triads. A given mother and father enter into the subsample only once. When data from two or more chil- dren are present, one child is selected ran- domly. The N for this replication analysis varied between 225 and 241 triads (depending on missing data). Results of the replication indicated no systematic contrast from the general patterns described below (tables are available from the authors), as is documented in footnotes.

The children in the triad sample (N 446) ranged in age from 16 to 26 years with a median age of 19. The fathers were from 37 to 67 years (median age 46); mothers from 34 to 67 years (median age 43). Although the sample is not a national probability sample, it is generally representative (i. e., considerably diverse with respect to socioeconomic status, education, geographic distribution, and political and religious group identification). All grandparents were members of the same Los Angeles-based health plan, but their chil- dren and grandchildren live throughout the country. Many other studies of parent-child relationships have, by comparison, somewhat more serious limitations. They tend to be either based on availability samples of college students and their parents, or people living in a small geographic area and hence more homogeneous and decidedly less mobile than our sample.

The variables which allow for comparisons of actual and perceived attitudes are pur- posively diverse (see Table 1 for key words identifying each question; the complete ques- tionnaire is available from the authors). Attitudes range from major political dimen- sions along which generational cleavages are often depicted ("The United States should be ready to answer any challenge to its power, anywhere in the world") to matters of relig- ious belief ("God exists in the form in which the Bible describes Him"). The nine items re- ported below, with four response alternatives (strongly agree to strongly disagree), were answered separately by fathers, mothers, and children. Later in the questionnaire, the chil- dren were asked to predict their mother's and father's response to the same question.

The expected effect of using single-item indicators rather than attitude scales is an at- tenuation of the correlations reported (Born- stedt and Carter, 1971), although the effects on the partial regression coefficients are ambiguous (Cohen and Cohen, 1975). Because of this, we expect some of our find- ings to reflect less parent-to-child predictivity than those reported elsewhere, where more reliable measurement procedures have been employed (for example, see Acock and Bengtson, 1978).

Before reporting our results we must point out two significant limitations in the analysis to follow. The first concerns the cross-sec- tional research design employed. To infer socialization influence processes from indica- tors of parent-child similarity collected at just one point in time is obviously risky; attitudes of both parents and children are subject to change over time, and a snapshot cross-sec- tional assessment may misrepresent patterns of similarity or dissimilarity. A second and related limitation concerns the linear and unidirectional model employed in our data analysis. In interpreting our results, it should be kept in mind that socialization is a bi- lateral, not unidirectional, process involving continual and inevitable negotiation and in- fluence (see Hagestad, 1977; Bengtson and Black, 1973; Bengtson and Troll, 1978). Full examination of the influence process should consider effects children have on parents in the formation and change of attitudes over time, as well as the more common parent-to- child model of influence considered here.

RESULTS

The system of equations 1.1 to 1.3 above were solved using multiple regression. This was done for each of the nine variables, allow- ing nine separate answers to each of the first four research questions. Space limitations preclude presenting complete results for each equation on each of the questions. Table 1 provides the essential summary of the multi- ple regression analysis; more detailed infor- mation such as unstandardized coefficients are available from the authors.

1. Stated versus Perceived Parental Orien- tation as Predictors of Youth

Which predicts the child's response more accurately: the perceived response of the

August 1980 JOURNAL OF MARRIAGE AND THE FAMILY 505

This content downloaded from 82.40.156.237 on Sun, 3 Aug 2014 11:48:52 AMAll use subject to JSTOR Terms and Conditions

Page 7: Socialization and Attribution Processes: Actual versus Perceived Similarity among Parents and Youth

(T3 z

Q C? z

H

0

z 0 2$

0

r• r•

0

H

oi td

tA C= tA?

tA

Pt

*

C)

zN

00 i

"F

~Q cri

0o

0

QX

0

o

o X

?,,,

QX P QX

L-4

_-q W) r--q ( C O -- rtt _-q C L() -q

00 0L() N q m00 Itc 00 m

clq C) crr lq m m tn

r -'- 00

C() C) 00 C) C) C)~

C~q m Cq :t m crr m~

O Oc (Nf \Ci \CC r(N C~

C,-4 C) ~C~ C)] C4,4 C',4'

C',4 M 'It C',4 M M tn

- t t m 00 m m 00 t -

*** ***** 00 O\ O\ (N ec] 00 ~t (NCc

e-r O ?- -

OO Er 0 00

(N O\ Cc '-- (N (N in 00 '-r

(N UO O\ 00 ~- I- 00 (N

t-r 00 0? r - 00 f

(N- t" 00 (N e f 00Cc

a) 4-J u 444

;-4 a- C13 0D o )

c, 0

0 ~ ~ r-

II

a)

on

cis

?

cis

o

t-

II

0

a)

a) q

o oo

Ia

ci

a) Q

ot-

c o

,. o

ao

o c c- a

cis 0

a) a)

.5 .;Z c

a)

mother and father as attributed to them by the child (PM and PF) or their actual stated opinions (AM and AF)? There are a variety of ways of answering this research question. Each answer points to the greater predictivity of responses attributed to the parents, rather than the orientations reflected in the mothers' and fathers' own responses.

The first approach is to examine the ex- planatory power of the actual responses as opposed to the perceived responses of the par- ents. Table 1 indicates that, in each case, R,.,12 (column 9) is smaller than R,.,34 (column 10). Note that these statistics are the multiple correlations when the actual responses of the parents (AM and AF; i.e., R.,12) are used, in contrast to when the perceived parental re- sponses (PM and PF; i.e., R s.34) are em- ployed.

To test for possible sex-linkage contrasts, separate analysis of subsamples composed of father-mother-son triads and father-mother- daughter triads were carried out. The results were essentially similar. For example, the mean value of R,.,2 for the total sample of triads is .31, compared to .32 for the sub- sample of sons and .29 for the subsample of daughters. The mean value of R,.34 is .43 for the combined triad sample, compared to .41 for the sons' and .44 for the daughters' sub- samples. In short, no sex differences emerged in this comparison.

On some variables the greater predictivity of attributed responses are particularly domi- nant. In the case of "blacks push too hard" (Item 3), the

RS.34

= .42 (attributed com- pared to R,.,2 = .28 (stated). Also, for Item 2, "U.S. should answer any challenge," and Item 4, "Government should not interfere with business," the values of

R,.34 are much

larger than the values of Rs.12 (.43 to .27, and .41 to .22).

While these results demonstrate the su- premacy of the perceived responses in a pre- dictive sense, we need to go a second step to determine whether the actual and perceived orientation of parents are complementary predictors. On the one hand, we can examine the symbolic interactionists' contention that reality consists in shared perceptions (Kinch, 1963). From this position it would be reason- able to expect that, when one controls for the perceptions, the actual orientations of the parents are irrelevant. That is, the actual per- ceptions contain no information that adds to our ability to predict the child's responses ex-

506 JOURNAL OF MARRIAGE AND THE FAMILY August 1980

This content downloaded from 82.40.156.237 on Sun, 3 Aug 2014 11:48:52 AMAll use subject to JSTOR Terms and Conditions

Page 8: Socialization and Attribution Processes: Actual versus Perceived Similarity among Parents and Youth

cept through how the parents are understood by their children. Such a position is graphic- ally represented in Figure 1 by the arrows (X, to X3 to Xs) and (X2 to X4 to Xs).

On the other hand, against this highly cognitive (or "subjective") approach is the behaviorist perspective which stresses the im- portance of the objective orientations of others (indicated here by their stated response) as the most relevant stimuli or pre- dictors. Behaviorists who see attitudes as no more than behavioral responses to stimulus situations (for an early statement of this posi- tion see Doob, 1947) might thus view the ob- jective opinion of the parents as the direct source of influence. Social learning, then, may reflect a direct modeling of parental re- sponses to stimuli without regard to the inter- pretations children make of the parental re- sponses (Bandura, 1969). The resulting direct effect is represented by the longer arrows in Figure 1 (X, to X, and X, to X,).

The two positions outlined above can be examined by comparing the multiple correla- tion between the perceived parental responses and the child's response R5.34 (column 10, Table 1) to the multiple correlation based on both perceived and actual parental responses, R5.1234 (column 11). Results indicate that, in most cases, the actual responses of the par- ents add little predictivity compared with the child's perceptions (attribution).

This same issue can be considered by in- specting standardized coefficients presented in columns 3 and 4 of Table 1. These indicate the impact of parental variation on the child when the perceptions are controlled. Note that most of these Beta weights are small and not statistically significant. Thus, the par- ents' stated attitudes have only a limited direct effect on the children's responses.

At a more general level, it should be noted that the level of parental predictivity overall is much higher than is conventionally reported in the literature on the generation gap. The multiple R's (see column 11 of Table 1) range from .36 to .60. These results suggest that parental influence-through attributed re- sponses, especially those of the mother-is considerable. This level of prediction is higher than Ix;ually found because we have incorporated both actual and attributed atti- tudes as well as combining the influence of both the mother and the father into a single model of the influence process.

2. Accuracy of the Child's Attribution Re- garding Parents' Attitudes

The second research question concerns the similarity between perceived parental re- sponse (furnished by the child) and the actual parental response. How correctly do children predict how their parents will respond to questions concerning social and political is- sues? The data reviewed above indicate that perceptions are better predictors of the child's orientations than are the actual opin- ions of the parents. If the child's perceptions are accurate, there is less cause for concern on the part of researchers relying on data from only one generation. If, on the other hand, the children do not attribute the cor- rect opinions to their parents, this is of major significance to socialization theory and re- search practice. Our basic question concerns how well the actual response of the parents can predict the response attributed to them by the child: in symbolic terms, the congru- ence between AM- PM and AF-PF.

The results indicate that the level of accur- acy is generally low. Even recognizing the potential reliability problems involved in the use of single item indicators, Table 1 reflects low coefficients in terms of (3 ,3 (column 1, actually the equivalent of r,,3) and (3 42 (col- umn 2, the equivalent of r42). Most of the 3 's are only moderate.' Interestingly, the highest level of predictivity comes in the case of the item on religion (.50 and .53) which is per- haps the most explicitly socialized of the opinions we have examined.

Many previous studies have shown a weak level of predictivity between the stated opin- ions of parents and their offspring. Such re- sults may be explained in part by the present data, which indicate that the children simply do not know their parents' opinions. For ex- ample, in response to the political-economic item concerning government regulation of business (Item 4), the estimates children have of their parents' opinions are only slightly

'To test potential bias from our sampling logic (i.e., our decision to include all possible triads), the results were repeated for a nonreplicated subsample in which each father and mother (thus each family) was repre- sented only once. The results are similar for this subsam- ple (N = 225 to 241 triads depending on the item). On the nonreplicated subsample, the mean value of P31 is .32; in Table 1 it is also .32. The mean value of1P42 for the "nonreplicated" subsample is .31, identical to that in Table 1 for the total sample of triads.

August 1980 JOURNAL OF MARRIAGE AND THE FAMILY 507

This content downloaded from 82.40.156.237 on Sun, 3 Aug 2014 11:48:52 AMAll use subject to JSTOR Terms and Conditions

Page 9: Socialization and Attribution Processes: Actual versus Perceived Similarity among Parents and Youth

related to the actual opinions of their parents ( 31 = .11 and 42 = .15).

Sex of the child does not appear to alter these results. When male and female children are considered separately, the beta weights are generally similar. While the mean value of ( 31 is .30 for the total sample reported in Table 1, the mean for the male child subsam- ple is .27 and for the female child subsample it is .33. This indicates only a minimal differ- ence in the accuracy with which boys estimate the opinions of their mothers as opposed to the accuracy with which girls estimate the opinions of their mothers. Similarly, the mean value of (3 42 for the total sample is .31 while, for the male subsample, it is .32 and, for the female subsample, it is .30.

The comparison of the sex-specific linkages produces little in the way of differences. In correlational terms, at least, the daughters' perceptions of their mothers' opinions are only slightly more correlated with their mothers' actual opinions than are the sons' perceptions of their mothers' opinions. Simi- larly, sons' perceptions of their fathers' opin- ions are only slightly more correlated with their fathers' actual opinions than are the daughters' perceptions. The evidence does not demonstrate strong same-sex specification of the accuracy of attribution of attitudes.

3. Perceived Versus Actual Similarity of Parents

The third research question concerns the actual similarity between the mother and father as contrasted to the child's representa- tions of their similarity. This issue is impor- tant for at least two reasons. First, if the parents do not agree between themselves, the extent of their agreement with young adult children is problematic. Moreover, the com- mon practice of relying on the father's opin- ion as the parental opinion in sociopolitical orientations (Troll and Bengtson, 1978) im- plies either a substantial covariation, r,2, between the father and mother, or, that only the father is important as an influencing agent. Second, if the children perceive the parents as reflecting greater agreement than they actually do, this adds further evidence to their misattribution (question 2).

From the correlations reported in column 8 of Table 1 it is clear that the children perceive their parents as similar to one another in atti- tudes. In the light of the reliability limitations

of single item indicators, the high correla- tions are all the more remarkable. The more liberal the mother is perceived, the more liberal the father is seen; the more conserva- tive the mother is viewed, the more conserva- tive the father is seen. In fact, the perceived mother-father similarity is so high, mean r34 = .77, as to present a serious collinearity problem in addressing the fourth research question comparing attributed responses of the father and the mother, respectively, as predictors of the child's opinion.

While the children believe both parents are highly concordant in their attitudes, the actual responses of the parents indicate a relatively low level of agreement. One ex- treme example of this involves Item 4 con- cerning governmental interference with busi- ness. The responses attributed to the two par- ents are highly correlated in the eyes of the children (r34 - .82), while the actual paren- tal responses show very low similarity (r2 = .02). Overall, the mean value of r12 is .30 (actual similarity) compared to the mean value of r34 of .77 (perceived similarity). Thus, postadolescent children greatly exag- gerate the attitudinal similarity of their par- ents.2

Our data demonstrate an interesting problem for those who have relied on a strictly correlational approach to "justifying the existence of a generation gap" (see Fried- man et al., 1972). The fact of the matter is that correlations between the two parents tended to be on the same order as those between either parent and his or her child. Separate analysis of the male child and female child subsamples produced virtually identical results. Neither male nor female chil- dren are better able to estimate their parents' opinions. Both male and female children report misattributed unanimity to their par- ents.

4. Maternal Versus Paternal Influence The fourth question, concerning relative

predictivity of fathers and mothers, can be assessed in terms of the relative influence of

2For the nonreplicated subsample in which each father and mother appears only once, the mean value of r12 is .30, identical to that in Table 1 for the total sample of triads. The mean value of r,, for the nonreplicated subsample is .78 compared to .77 for the total triad sam- ple in Table 1.

508 JOURNAL OF MARRIAGE AND THE FAMILY August 1980

This content downloaded from 82.40.156.237 on Sun, 3 Aug 2014 11:48:52 AMAll use subject to JSTOR Terms and Conditions

Page 10: Socialization and Attribution Processes: Actual versus Perceived Similarity among Parents and Youth

the responses that their children attribute to them. We will not discuss the relative influ- ence of their actual responses for two reasons. First, the direct influence as measured by P s, and 3 52 (columns 3 and 4 in Table 1) is typically small. Second, actual attitude predictivity has been extensively examined elsewhere, although with varying results (Jen- nings and Langdon, 1969; Connell, 1972; Niemi, 1973; Tedin, 1974; Lerner and Knapp, 1975; Acock and Bengtson, 1978; Cutler, 1977b).

Focusing on the relative predictivity of the perceived responses, the overall results (, s3 for mother and 3 14 for father in Table 1, col- umns 5 and 6) suggest the dominant role of the mother in predicting the child's orienta- tions. This result is consistent with our pre- vious analysis comparing parents in terms of actual responses (Acock and Bengtson, 1978). The mother is not only more predictive in some expressive areas, but her influence extends beyond this to most areas of sociali- zation including sociopolitical and work orientations. Comparing the coefficients of the response attributed to the mother, p

03, to

the response attributed to the father, p 54, the father has less of an impact on all nine variables we investigated. The greatest differ- ences are on Item 4, "Government should not interfere in business" (mother, 13

3=

.41; father, 54 = -.03) and Item 2, "U.S. should answer any challenge" (mother, 3 s, = .39; father, 3 = - .01). The smallest differences are on Item 3, "Blacks push too hard and fast" (mother, 'P3 -= .23; father, 13 54 = .16) and Item 7, "Marijuana is harm- less" (mother, [3

3 = .15; father, 3 ,p

.09).3 It is of interest, moreover, that no sex- linkage differences emerged; similar patterns were found in both the male-child and the female-child subsamples.

Overall, the pattern of these data question the common notion that the father plays the dominant role in the attitude socialization within the family. This conclusion should be qualified in view of a statistical problem that appears unavoidable. The response attribu-

ted to one parent is so highly correlated with the response attributed to the other parent (r34 in Table 1, column 8), that it is difficult to determine their separate effect. Neverthe- less, the response attributed to the father adds little to the prediction of the response of the child. This finding covers a wide range of substantive issues and does not appear to de- pend on the gender of the child.

5. Cohort Contrasts and Similarity

Any discussion of generations in socializa- tion and social change must deal with three conceptually and operationally distinct phe- nomena: (a) within-family (lineage) compari- sons; (b) cross-cohort (age aggregate) com- parisons; and (c) within-cohort (generation unit) comparisons (see Bengtson and Cutler, 1976). The first two are particularly crucial in empirical studies of intergenerational influ- ence.

Throughout this paper we have empha- sized that we are concerned with predictivity between generations within the family, rather than absolute agreement. This distinction, which was brought to the attention of sociolo- gists by Robinson (1957), has often been ig- nored by researchers investigating parent- child socialization. The data presented above using a correlational approach indicate that the relative liberality of the parent, with respect to the parent's generation, is associ- ated to the relative liberality of the child with respect to the child's generation. That is, the most liberal parents (relative to their cohort) have the most liberal children compared with other youth in the sample. However, this does not mean that parents and children neces- sarily agree (are at the same levels of liberal orientations). If each child were twice as liberal as his or her parent, the correlation (predictivity) would achieve unity while the agreement would be naught.

The importance of this distinction between predictivity and agreement is clear when we look at the cohort level of analysis. While the perceived (attributed) responses were most predictive of the youth's responses (research question 1), analysis at the cohort level shows that the actual responses of the parents are more similar to those of the youth than are the parents' perceived responses (attribu- tions).

Table 2 reports the mean anct standard de- viation for each variable, while Table 3 reports

3For the nonreplicated subsample of unique families, the maternal influence is slightly lower. In two variables the beta weight for the father is actually slightly larger than for the mother. For Item 3, "Blacks push too hard," the

P3s3 is .20 for the mother. On Item 9, "God

exists as the Bible describes Him," coefficients are .29 for the father and .26 for the mother.

August 1980 JOURNAL OF MARRIAGE AND THE FAMILY 509

This content downloaded from 82.40.156.237 on Sun, 3 Aug 2014 11:48:52 AMAll use subject to JSTOR Terms and Conditions

Page 11: Socialization and Attribution Processes: Actual versus Perceived Similarity among Parents and Youth

TABLE 2. MEANS, STANDARD DEVIATIONS, AND SIGNIFICANT DIFFERENCES IN ACTUAL AND PER- CEIVED RESPONSES

1. 2. 3. 4. 5. 6. Actual Actual Actual Perceived Perceived N

Variable Key Words Fathers' Mothers' Children's Mothers' Fathers' (Triads) Responses Responses Responses Responses Responses

X2/s2 X1/s1 X5/s5 X3/s3 X4/s4 1. Duty to work 3.15 3.02 2.41 3.30 3.38 405

.94 .98 1.09 .78 .82 2. U.S. answer any challenge 2.90 2.56 2.29 2.73 2.97 412

1.10 1.12 1.05 .98 1.03 3. Blacks push too hard 2.39 2.03 1.86 2.52 2.77 417

1.01 .94 .92 1.01 1.06 4. Government not interfere business 2.47 2.27 2.31 2.72 2.81 391

.96 .93 .85 .81 .90 5. Punish college demonstrators 2.80 2.54 1.84 2.77 2.96 415

1.04 1.02 .88 .95 .97 6. Maintain law and order 2.77 2.50 2.25 2.97 3.08 421

1.05 1.11 1.00 .86 .89 7. Marijuana is relatively harmless 1.65 1.52 2.61 1.57 1.54 407

.94 .86 1.13 .86 .85 8. Welfare people are lazy 2.43 2.27 2.27 2.89 3.07 414

.99 1.01 1.02 .96 .97 9. God as Bible describes 3.00 3.05 2.62 3.17 3.00 375

1.08 1.10 1.16 1.03 1.06

*The higher the mean, the more traditional (conservative) the direction of response, except for item seven where the lower the mean, the more traditional the category.

the mean absolute difference over the nine variables for selected comparisons. For exam- ple, the difference between actual fathers' re- sponses and actual children's responses ap- pears in row one of Table 2. Because of the characteristics of the sample, significance tests are not reported but can be computed from the data presented in Table 2 on the basis of the means and standard deviations.4

TABLE 3. MEAN ABSOLUTE DIFFERENCES OVER THE NINE VARIABLES FOR SELECTED COM- PARISONS

Mean Absolute Comparison Difference 1. Actual Fathers' Mean minus

Actual Children's mean

Ix2 -x5! .56 2. Actual Mothers' Mean minus

Actual Children's mean

Ix1 - X51 .40 3. Actual Fathers' Mean minus

Actual Mothers' mean 1X2 - X11 .21

4. Perceived Fathers' Mean minus Actual Children's mean

I4 - X5 1 .81 5. Perceived Mothers' Mean minus

Actual Children's mean

Ix3 - x51 .70 6. Perceived Fathers' Mean minus

Perceived Mothers' mean IX4 - X51 .15

4 A discussion of the cohort level similarities and dif- ferences is not complete without some attention to the standard deviations reported in Table 2. There is a syste- matic tendency for the children to have a smaller stan- dard deviation (indicating within-generation cohesion) than their parents. At the same time, the responses the children attribute to the parents are not only more con- servative in direction but also have a smaller standard deviation or spread than the actual responses of the par- ents. Thus, misattribution of youth also extends to per- ceiving less differentiation within the parental generation in attitude than actually exists. In some instances, these distortions are substantial. Item 7, "The most important problem facing this country is to maintain law and order," provides a clear example: The standard deviation for the fathers' actual opinion is 1.05, for mothers', it is 1.11. By itself this indicates greater cohesion within the generation of children (S.D. = 1.00) than within the generation of parents. However, in terms of attribution, for the children's perceptions of the mothers', the stan- dard deviation is 0.86, and for the perceptions of fathers', it is 0.89. While the parents are more diversified as a

group in their attitudes, the children do not accurately perceive this diversification. (A word of caution: The standard deviation may be biased because of the duplica- tion of individual parents in the same triads.)

510 JOURNAL OF MARRIAGE AND THE FAMILY August 1980

This content downloaded from 82.40.156.237 on Sun, 3 Aug 2014 11:48:52 AMAll use subject to JSTOR Terms and Conditions

Page 12: Socialization and Attribution Processes: Actual versus Perceived Similarity among Parents and Youth

The most consistent finding is that the "generation gap" is far more apparent in the minds of the children-the responses attribu- ted to the parents by youth-than it is evi- denced in comparing the actual opinions of the two generations. Across the nine variables the mean response of the children's genera- tion (X,) is far closer to the mean of the par- ents (X, for mothers and X2 for fathers) than it is to the mean of responses attributed to the parents (X, for perceived mother's opinion and X4 for perceived father's opinion).

An example of how the "generation gap" is predominantly a matter of perception is illu- strated by Item 8, "People on welfare are lazy." The mean actual response for the group of fathers' is 2.43 while the mothers' and youths' mean are both 2.27. Thus, the children's level of agreement with this state- ment is quite similar to that of the parents. An inspection of how the children thought their parents would respond, however, re- flects substantial differences. The children's mean score for how they predicted their mother would respond is 2.89, for their father, 3.07. The youth, as a group, expect parents to agree with the statement much more than they actually do.

One might term this phenomenon "misat- tributed polarization." Children perceive their parents' attitudes as highly polarized from their own (see rows 4 and 5 in Table 3). The children consistently perceive their parents' attitudes as more conservative or tra- ditional than they really are; fathers' even more so than mothers'. The parents' actual opinions, however, are much closer to the childrens' though the mean absolute differ- ence is still substantial (see rows 1 and 2 in Table 3). Overall, in eight of the nine varia- bles, the children perceive the mothers as more traditional than they actually are.5

In general, then, these survey results do not give support to the experimental findings concerning attitude assimilation and contrast (Hovland et al., 1957). The experimental findings reported that subjects perceived statements as more similar to their own posi- tion on issues that did not involve a clear cleavage (assimilation), and more dissimilar when a clear cleavage was involved (contrast).

We found a persistent misperception that in- volved polarization or contrast. Moreover, the single item for which there was not an exaggerated perception of parental conserva- tism involved marijuana-an issue of clear cleavage between the generations.6

When male and female children subsam- ples were examined, there was some gender- specific differentiation of these overall results. Male children had more liberal re- sponses than female children on six of the nine variables. However, the differences were generally small, the mean absolute difference between the male and female children being .16. All of the mean absolute differences reported in Table 3 are within .02 of the cor- responding means in the male and female subsamples. For example, the mean absolute difference between the perceived fathers' mean and the actual children's mean, X 4-X ', is .81 in Table 3. The corresponding mean absolute difference for the male child subsample is .82 and, for the female child subsample, it is .81. Thus, there is no evi- dence of specification by the sex of the child, although, it is acknowledged that the male children are somewhat more liberal or less traditional than the female children.

DISCUSSION

Our data indicate that children do not per- ceive their parents' stated opinions accurately. Yet the perceived (attributed) parental atti- tude is a surprisingly high predictor of the post-adolescent youths' own attitude. What can account for this? Answers may be found in considering, first, factors which may con- tribute to misattribution or misperceptions of attitudes between generations, and second, in the meaning of generational contrasts-real or imagined-within a family.

The present findings suggest that youth perceive their parents' opinions to be sharply divergent from their own. In this sense it may be said that the "generation gap" in attitudes

s The results are identical for the nonreplicated parent subsample.

6 The assimilation and contrast arguments of social judgment theory (Sherif and Hovland, 1961) are sup- ported by small group research (Hovland et al., 1957). Especially important for present consideration are the patterns of acceptance and rejection. Sherif et al. (1965) have proposed that the opinion a person has serves as an "anchor" for the placement of other opinions. By impli- cation, we would expect children to misperceive their parents' attitudes as polarized (in cases of substantial disagreement) or as assimilated (in cases of substantial agreement).

August 1980 JOURNAL OF MARRIAGE AND THE FAMILY 511

This content downloaded from 82.40.156.237 on Sun, 3 Aug 2014 11:48:52 AMAll use subject to JSTOR Terms and Conditions

Page 13: Socialization and Attribution Processes: Actual versus Perceived Similarity among Parents and Youth

is real, inasmuch as differences are perceived to exist. But the children do not perceive their parents' opinions accurately: Attributed differences are greater than actual differ- ences. From a sociology of knowledge per- spective (Berger and Luckmann, 1961), the "generation gap" may be considered a social construction of reality, albeit reflecting misattribution of actual orientations. At the cohort level, the substantial differences between the youth's actual opinions and the attitudes they attributed to their parents indi- cates a belief that the contrasts exist. Several other studies support the generalization that youth and middle-aged respondents both assume there to be generational contrasts in opinions (Bengtson, 1971; Flacks, 1971; Yankelovitch, 1972; Lerner and Knapp, 1975). Thus, one must begin with the taken- for-granted belief that there are cohort con- trasts between generations in attitudes; this social construction of reality creates the ex- pectation of generational differences and leads to misattributing opinions to the child's own parents.

Why is it taken for granted that generation contrasts exist? One reason for misattribu- tion, suggested by many life-cycle theorists, concerns the difference in ontogenetic or de- velopmental level between youth and their middle-aged parents, and the resulting pro- cesses of negotiation and conflict that occur within families (Davis, 1940; Bengtson and Black, 1973). Similar to Mannheim's (1952) notion of "fresh contact," this perspective indicates that the concerns of adolescents negotiating their emancipation from parents, and coming into contact with the values and institutions of adult society, are oriented toward contrasting goals. Thus, for youth, needs in the service of psychosocial emanci- pation may lead to conflict; indeed, if differ- ences do not exist, they may be manufactured as youth attempt to define a distinctive iden- tity (Bettleheim, 1965). Bengtson and Kuy- pers (1971) suggest such contrasts reflect dif- ferences in the "generational stake" between parents and youth. Each generation has an investment in the generational bond. But, for youth, the "stake" is more toward maximiz- ing a sense of separate identity; for parents, the investment pays off in maximizing conti- nuity. In short, misattribution may be a normal consequence of ontogenetic develop- ment, rooted in the psychosocial needs of dif-

ferent generations who are at contrasting points on the life-cycle.

A second factor within family misattribu- tion may be lack of communication between generations. Issues such as we have examined in this paper may be infrequently discussed within families. Support for this argument is found in the high similarity children attribute to mothers' and fathers' attitudes, whereas the actual opinions of the two parents were only moderately related. Moreover, while the parents' generation is approximately as heter- ogeneous (see the standard deviations in Table 2) as the youth's, the children misat- tribute homogeneity to the parents.

A third reason for misattribution is deliber- ate misrepresentation: parents shielding children from their actual attitudes, and the children accepting, then perpetuating, the myth. That the parental generation is per- ceived as more traditional or conservative than their actual opinions indicate may re- flect what parents wish their children to think. We have termed this "polarized misat- tribution." Actual similarity is seen as sub- stantial difference.

Polarized misattribution is substantial on items concerning government regulation of business, black power, and U.S. readiness to answer challenges to its power. At the cohort level the parents are thought of as very con- servative, whereas their actual responses are similar to those of their children. Parents, who may be only partially committed to the work ethic, may misrepresent their feelings because they want their children to work hard in school. Or, they may fear that their chil- dren will become seduced by radical or liberal orientations. Suspecting that some rebellion from parental opinions is inevitable, they may present themselves as highly conventional. Caution, and the presentation of traditional attitudes, may be the implicit socialization agenda for many parents of the youth in our sample.

Despite the strong evidence for misattribu- tion between generations, the data point to parent-child predictivity as well. The fact that attributed parental opinions were highly predictive of the child's opinions appears to support the argument of transmission from generation to generation within the context of family socialization. One major implication from these data is that children are strongly influenced by parental attitudes-although it

512 JOURNAL OF MARRIAGE AND THE FAMILY August 1980

This content downloaded from 82.40.156.237 on Sun, 3 Aug 2014 11:48:52 AMAll use subject to JSTOR Terms and Conditions

Page 14: Socialization and Attribution Processes: Actual versus Perceived Similarity among Parents and Youth

is the perceived, not the actual, parental atti- tudes which are influential. The distinction between attributed versus actual attitudes is crucial, for it appears to explain contradic- tory findings in previous research concerning parent-child similarity.

Such an interpretation of these data sug- gests considerable support for symbolic inter- actionists and attribution theorists who argue that one's perception of another's behavior is far more predictive of the other's influence on him or her than is the other's actual behavior or attitude. Indeed, within the family, the actual opinions of parents appear to have limited direct bearing on children's orienta- tions. The chain of influence is rather through perceived (attributed) parental opinions to the child's actual opinions. If so, socialization theorists must in the future pay much greater attention to attribution processes in charting intergenerational influence patterns.

SUMMARY

Results of the data analysis can be sum- marized as follows. First, attributed (or per- ceived) responses of parents were much more predictive of post-adolescent children's atti- tudes than were the parents' actual opinions. In fact, attributed parental responses pre- dicted children's opinions at an unexpectedly high level, in light of previous studies suggest- ing low similarity in actual parent-child com- parisons of opinions. Second, there was con- siderable error in the children's attributions of parental responses, with the children tend- ing to perceive parents as much more conser- vative or traditional than the parents' actual responses would indicate. Third, the children perceived mothers and fathers to be highly similar in their opinions, whereas the actual responses of mothers compared to fathers were only moderately correlated. Fourth, mothers' responses-whether attributed or actual-were more predictive than the fathers' responses of the children's opinions. Fifth, at the cohort level, there appeared a "polarized misattribution" on the part of youth-children perceiving parental attitudes as contrasted to their own, and more conser- vative than they actually were. In none of these five issues do substantial male-female contrasts appear significant; the pattern for sons as well as daughters is substantially similar.

These results suggest considerable support for the symbolic interactionists and attribu- tion theorists who argue that one's perception of another's behavior is far more important than is the other's actual behavior or attitude in influence processes. Indeed, within the family, the actual opinions of parents appear to have little direct bearing on children's ori- entations, except as the actual orientations are perceived and reinterpreted by the children. It is not what parents think, but what their children think they think, that predicts their offsprings' attitudes.

REFERENCES

Acock, A. C., and V. L. Bengtson 1977 "Opinions within the family: What parents

think and what children think they think." Pa- per presented at the annual meeting of the Pa- cific Sociological Association (March).

1978 "On the relative influence of mothers and fathers: A covariance analysis of political and religious socialization." Journal of Marriage and the Family 40 (August):519-530.

Acock, A. C., and M. DeFleur 1972 "The contingent consistency model of the atti-

tude behavior relationship." American Socio- logical Review 37 (December):714-726.

Aldous, J., and R. Hill 1965 "Social cohesion, lineage type, and intergener-

ational transmission." Social Forces 43 (May): 471-482.

Bandura, A. 1969 "Social-learning theory of identificatory pro-

cesses." Pp. 213-262 in D. A. Goslin (Ed.), Handbook of Socialization Theory and Re- search. Chicago:Rand McNally.

Bengtson, V. L. 1971 "Inter-age perceptions and the generation

gap." The Gerontologist 11 (October):85-90. 1975 "Generation and family effects in value sociali-

zation." American Sociological Review 40 (June):358-371.

Bengtson, V. L., and K. D. Black 1973 "Intergenerational relations and continuities in

socialization." Pp. 207-294 in P. Baltes and W. Schaie (Eds.), Life-Span Developmental Psy- chology: Personality and Socialization. New York:Academic Press.

Bengtson, V. L., and N. Cutler 1976 "Generations and intergenerational relations:

Perspectives on age groups and social change." Pp. 130-159 in R. Binstock and E. Shanas (Eds.), Handbook of Aging and the Social Sci- ences. New York:Van Nostrand Rinehold Com- pany.

Bengtson, V. L., and J. A. Kuypers 1971 "Generational differences and the develop-

mental stake." Aging and Human Develop- ment 2 (November):249-260.

August 1980 JOURNAL OF MARRIAGE AND THE FAMILY 513

This content downloaded from 82.40.156.237 on Sun, 3 Aug 2014 11:48:52 AMAll use subject to JSTOR Terms and Conditions

Page 15: Socialization and Attribution Processes: Actual versus Perceived Similarity among Parents and Youth

Bengtson, V. L., and L. Troll 1978 "Youth and their parents: Feedback and inter-

generational influence in socialization." Pp. 106-130 in R. M. Lerner and G. B. Spanier (Eds.), Child Influences on Marital and Family Interaction: A Life-Span Perspective. New York:Academic Press.

Berger, P. L., and T. Luckmann 1961 The Social Construction of Reality. New York:

Doubleday. Bettleheim, B.

1965 "The problem of generations." Pp. 76-109 in E. Erickson (Ed.), The Challenge of Youth. New York:Anchor.

Bohrnstedt, G. W., and T. M. Carter 1971 "Robustness in regression analysis." Pp. 118

146 in Herbert L. Costner (Ed.), Sociological Methodology. San Francisco:Jossey-Bass, Inc.

Cohen, J., and P. Cohen 1975 Applied Multiple Regression. New York:Wiley.

Connell, R. W. 1972 "Political socialization in the American family:

The evidence re-examined." Public Opinion Quarterly 36 (Fall):321-333.

Cutler, N. E. 1977a "Generational analysis and political socializa-

tion." Pp. 294-326 in S. A. Renshon (Ed.), Handbook of Political Socialization Theory and Research. New York:The Free Press.

1977b "Generational approaches to political sociali- zation." Youth and Society 8 (December): 175- 207.

Davis, K. 1940 "The sociology of parent-youth conflict."

American Sociological Review 5 (October):523- 534.

Doob, L. W. 1947 "The behavior of attitudes." Psychological Re-

view 54 (May):135-156. Flacks, R.

1971 Youth and Social Change. Chicago:Markhan. Friedman, L. N., A. R. Gold, and R. C. Christie

1972 "Dissecting the generation gap: Intergener- ational and intrafamilial similarities and differ- ences." Public Opinion Quarterly 36 (Fall): 334-346.

Gallagher, B. J. 1974 "An empirical analysis of attitude differences

between three kin-related generations." Youth and Society 4 (March):327-349.

Hagestad, G. O. 1977 "Role change in adulthood: The transition to

the empty nest." Unpublished manuscript, Commission on Human Development, Univer- sity of Chicago.

Hill, R., N. Foote, J. Aldous, R. Carlson, and R. Mac Donald

1970 Family Development in Three Generations. Cambridge: Schenkman.

Hovland, C. I., D. J. Harvey, and M. Sherif 1957 "Assimilation and contrast effects in reactions

to communication and attitude change." Jour- nal of Abnormal and Social Psychology 55 (March):244-252.

Hyman, H. 1969 Political Socialization. Glencoe:The Free Press.

Jennings, M., and K. Langdon 1969 "Mothers versus fathers: The formation of

political orientations among young Ameri- cans." Journal of Politics 31 (May):329-358.

Jones, E. E., and V. A. Harris 1967 The attribution of attitudes." Journal of Ex-

perimental Social Psychology 3 (January): 1-24. Kandel, D., and G. Lesser

1972 Youth in Two Worlds. San Francisco:Jossey- Bass, Inc.

Kelley, H. H. 1973 "The process of causal attribution." American

Psychologist 28 (February):107-128. Kinch, J. W.

1963 "A formalized theory of the self-concept." American Journal of Sociology 68 (January): 481-486.

Lerner, R. M., and J. R. Knapp 1975 "Actual and perceived intrafamilial attitudes of

late adolescents and their parents." Journal of Youth and Adolescence 4 (March):17-36.

Mannheim, K. 1952 "The problem of generations." Pp. 276-322 in

D. Kecskemeti (Ed.), Essays on the Sociology of Knowledge (rev. ed.). London:Routledge and Kegan Paul (originally published, 1928).

Niemi, R. G. 1973 "Political socialization." Pp. 117-138 in J. N.

Knutson (Ed.), Handbook of Political Psychol- ogy. San Francisco:Jossey-Bass, Inc.

Payne, S., D. A. Summers, and T. Stewart 1973 "Value differences across three generations."

Sociometry 36 (March):20-30. Robinson, W. S.

1957 "The statistical measurement of agreement." American Sociological Review 22 (February): 17-25.

Shaw, M. E., and P. R. Costanzo 1970 Theories of Social Psychology. New York:

McGraw-Hill. Sherif, C. W., M. Sherif, and R. E. Nebergall

1965 Attitude Change: The Social Judgment-Involve- ment Approach. Philadelphia:W. B. Saunders.

Sherif, M., and C. I. Hovland 1961 Social Judgment: Assimilation and Contrast

Effects in Communication and Attitude Change. New Haven:Yale University Press.

Shibutani, T. 1955 "Reference groups as perspectives." American

Journal of Sociology 60 (February):562-569. Stryker, S.

1965 "Relationships of married offspring and par- ents: A test of Mead's theory." American Jour- nal of Sociology 62 (November):308-319.

Tedin, K. L. 1974 "The influence of parents on the political atti-

tudes of adolescents." American Political Sci- ence Review 68 (December):1579-1592.

Thomas, L. E. 1974 "Generational discontinuity in beliefs: An ex-

ploration of the generation gap." Journal of Social Issues 30 (3):1-22.

514 JOURNAL OF MARRIAGE AND THE FAMILY August 1980

This content downloaded from 82.40.156.237 on Sun, 3 Aug 2014 11:48:52 AMAll use subject to JSTOR Terms and Conditions

Page 16: Socialization and Attribution Processes: Actual versus Perceived Similarity among Parents and Youth

Thomas, W. I. 1972 "The definition of the situation." Pp. 331-336

in Jerome G. Mannis and B. M. Meltzer (Eds.), Symbolic Interaction (rev. ed.). Boston:Allyn and Bacon. (Originally published, 1931.)

Troll, L., and V. L. Bengtson 1978 "Generations in the family." Pp. 126-161 in W.

Burr, R. Hill, I. Reiss, and I. Nye (Eds.), Handbook of Contemporary Theory in Family Research. New York:The Free Press.

Troll, L., B. L. Neugarten, and R. J. Kraines 1969 "Similarities in values and other personality

characteristics in college students and their parents." Merrill-Palmer Quarterly 15 (Octo- ber):323-336.

Yankelovitch, D. 1972 The Changing Values on Campus. New York:

Simon and Schuster.

Women and Marriage in Kpelle Society Caroline H. Bledsoe. This book examines the ways in which economic and political opportunities for men and women influence marriage patterns among the Kpelle of Liberia. It focuses particularly on women's use of domestic and marital institutions to acquire the wealth and status necessary to gain independence for themselves, yet maintain control over other

people. It also examines how urbanization and rapid social

change affect the strategies of Kpelle women and men, notably in relation to the traditional "wealth-in-people" system, which binds people to their superiors in ties of marriage, clientship, and filial obligation. Illustrated. $16.5 o

I'

;:\ h Stanford University Press

August 1980 JOURNAL OF MARRIAGE AND THE FAMILY 515

This content downloaded from 82.40.156.237 on Sun, 3 Aug 2014 11:48:52 AMAll use subject to JSTOR Terms and Conditions