Académique Documents
Professionnel Documents
Culture Documents
Family Structure, Materialism, and Compulsive Consumption Author(s): Aric Rindfleisch, James E. Burroughs and Frank Denton Source: Journal of Consumer Research, Vol. 23, No. 4 (Mar., 1997), pp. 312-325 Published by: The University of Chicago Press Stable URL: http://www.jstor.org/stable/2489568 . Accessed: 04/11/2013 07:41
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 of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.
The University of Chicago Press and Journal of Consumer Research, Inc. are collaborating with JSTOR to digitize, preserve and extend access to Journal of Consumer Research.
http://www.jstor.org
This content downloaded from 188.26.232.243 on Mon, 4 Nov 2013 07:41:38 AM All use subject to JSTOR Terms and Conditions
Family
Structure,
Materialism, and
Compulsive
Consumption
Over the past30 years,the traditional American family has undergone dramatic structural changes, as evidenced by double-digit increases in the rates of divorce, remarriage, cohabitation, and single-parent families (Cherlin 1992). As of 1991, married couples with children represented only 37 percent of all families and 26 percent of all households in the United States (Ahlburg and De Vita 1992). Demographersestimate that more than half of today's children will spend some time growing up in a single-parent family (Cherlin 1992). In sum, most young Americans are likely to spend part of their childhood in a disrupted-family structure,which we define as the dissolution of a two-parent family due to divorce or separation. To date, only a handful of studies in consumer behavior and marketinghave investigated the implications of these dramatic shifts i.n the structure of the American family. Both consumer researchersand marketersappearto maintain the conception that the intact two-parent family is an adequate representation of most consumers' family life
*Aric Rindfleisch and James E. Burroughs are doctoral candidates in marketing at the University of Wisconsin-Madison, School of Business, 975 University Avenue, Madison, WI 53706. FrankDenton, Ph.D., is editor of the Wisconsin State Journal, 1901 Fish Hatchery Road, Madison, WI 53713, and was also a doctoral candidate at the University of Wisconsin-Madison at the time this research was conducted. The authors thank Edward Fischer, Jeff Inman, David Mick, Robert Miller, Christine Moorman, John Murry, Charles Schaninger, three reviewers, and the editors for their encouragement and helpful comments. Please address all correspondence to the first author.
experience. In their recent literature review, Ahuja and Stinson (1993) found only three studies published in marketing and consumer research journals that investigated the relationship between family structure and "selected marketing related variables" (p. 469). Only one of these studies examined the relationship between family structure and children's consumption behaviors (Kourilsky and Murray 1981). Apart from those discussed in the Ahuja and Stinson review, we were able to find only two other studies investigating the consumption ihnplications of family structure (McAlexander, Schouten, and Roberts 1993; Wilkes 1995). Given that the extant consumer research and marketing literature focuses almost exclusively on the traditional two-parent family, we know very little about how nontraditional family structures, such as single-parent families, might impact consumption patterns and preferences. Recently, a broad base of family sociology research has empirically linked disruptedfamilies to such outcomes as delinquent behavior, teenage pregnancy, substance abuse, poverty, poor health, and academic underachievement (Amato 1993; Cherian 1989; Cherlin 1992; Flewelling and Bauman 1990; Kurdek and Fine 1993; McLanahan and Booth 1989; Thompson, Hanson, and McLanahan 1992; Wu and Martinson 1993). What might be the relationship between family structureand consumer attitudes and behaviors? We address this question by investigating the relationships between family structure and two consumption-related outcomes: materialistic attitudes and compulsive consumption behavior. Materialism and compulsive consumption are appropriateconstructs for inves312
C) 1997 by JOURNALOF CONSUMERRESEARCH, Inc. * Vol. 23 * March 1997 All rightsreserved.0093-5301/97/2304-0005$03.00
This content downloaded from 188.26.232.243 on Mon, 4 Nov 2013 07:41:38 AM All use subject to JSTOR Terms and Conditions
AND COMPULSIVE CONSUMPTION MATERIALISM, STRUCTURE, FAMILY tigation for two reasons. First, both constructs are regarded as importantby-products of our consumer culture (Belk 1985; Faber and O'Guinn 1992; Hirschman 1992; O'Guinn and Faber 1989; Richins and Dawson 1992). Second, and more important,both the consumer behavior and family sociology literatureprovide a strong theoretical basis for suspecting that these constructs may be linked to a disrupted-family structure.
313
CONCEPTUAL FRAMEWORK
The consumer research and marketing literature provides scant information about the relationship between disruptedfamilies and the consumer attitudes and behaviors of individuals reared in nontraditional families. On the other hand, family sociologists have been at the forefront of exploring the relationship between family structure and the behaviors of individuals growing up in nontraditionalfamily forms. We draw on this literaturebase to propose that (1) young adults rearedin disrupted-family structuresexhibit higher levels of materialism and compulsive consumption than young adults reared in intact families, (2) the relationshipbetween family structureand both materialismand compulsive consumption can be mediated by a decrease in family resources and the increase in family stressors associated with a family disruption, and (3) the impact of family disruption on both family resources and stressors is moderated by socioeconomic status.
Although most of the consumer research literatureappears to view materialism as a negative trait, materialistic attitudes also appear to hold some redeeming qualities and normative value. As Yankelovich (1982) notes, materialistic attitudes are held by most American adults, and material objects are a necessity for successfully functioning in modem society. Likewise, Csikszentmihalyi and Rochberg-Halton (1981) and Holt (1995) suggest that some forms of materialism serve an instrumental role, providing individuals with a means of creating valued experiences and enhancing interpersonal relations. For example, McAlexander et al. (1993) observe that recently divorced adults appearto use material objects and acts of consumption as means to cope with the stress produced by this life transition. It seems reasonable to expect that children faced with the disruptionof their family structure might enact similar coping strategies, perhaps to even a greater extent. Both views lead to the same conclusion: Children experiencing the disruption of their families appear especially likely to place greater emphasis on material objects in an effort to adjust to their new roles as members of a disrupted family. Specifically, family disruption may have both direct and indirect influences on the development of materialistic attitudes in children. First, family disruption may influence materialistic attitudes directly by prematurely expanding children's consumption roles and responsibilities. As McLanahan and Booth (1989) observe, children living in economically deprived single-parent households must often assume adult responsibilities such as earning money for the family or taking care of younger siblings. Thus, the development of materialism may be viewed as a by-product of their assumption of these new consumption-related roles. In addition, if divorced parents compete for their child's affection through gift giving, this may lead the child to equate material goods with feeIing loved and wanted and thus provide for an early socialization toward greater materialism. In addition to these potential direct effects, changes in family structure may also influence materialistic values indirectly. As An, Haveman, and Wolfe (1993) observe, young adults who face stressful life changes during childhood or early youth often experience feelings of insecurity, which they try to assuage by claiming "possession" of persons or objects they can control. Moreover, the breakdown in communication that cormmonlyaccompanies divorce appears to have an indirect impact on children's consumer socialization. Children from households characterizedby infrequent communication with parents are less likely to perform socially desirable consumption activities and are often more materialistic than children rearedin families characterizedby frequent and open consumption-oriented communication (Moore and Moschis 1981). In addition, consumer researchershave established that material possessions serve an importantrole as symbolic ties to interpersonalrelationships (Belk 1988; Richins 1994). Thus, children in disrupted households may
This content downloaded from 188.26.232.243 on Mon, 4 Nov 2013 07:41:38 AM All use subject to JSTOR Terms and Conditions
314 come to rely on special possessions to symbolically bridge the physical gap between themselves and an absent parent. Hla: Young adults from disrupted families tend to be more materialistic than young adults from intact families.
RESEARCH OF CONSUMER JOURNAL sumption is by definition harmful and has adverse consequences for most individuals when left unchecked (Hirschman 1992; O'Guinn and Faber 1-989). Hlb: Young adults from disrupted families tend to exhibit greatercompulsive consumption inclinations than young adults from intact families.
This content downloaded from 188.26.232.243 on Mon, 4 Nov 2013 07:41:38 AM All use subject to JSTOR Terms and Conditions
FAMILY STRUCTURE, MATERIALISM, AND COMPULSIVE CONSUMPTION quently or not at all. According to one estimate, 35 percent of these children never see their absentee fathers, and another24 percent see them less than once a month (Seltzer and Bianchi 1988). The combined loss of these multiple forms of tangible and intangible family resources often represents a significant reduction in the total parental "investment" made in children from disrupted families (Amato 1993; Fuchs 1983). This inequality in family resources places children from disruptedfamilies at an inherent disadvantage compared with children from intact families, increasing their risk of antisocial and dysfunctional behaviors and reducing their well-being and life chances. H2a: The relationship between family structureand materialistic attitudes is mediated by young adults' perceived level of family resources. H2b: The relationship between family structureand compulsive consumption is mediated by young adults' perceived level of family resources. The Family-Stressors Hypothesis. For many years, developmental psychologists have stressed the importance of stability in caregivers and resource providers to ensure the healthy development of children. High amounts of instability and change are particularlyharmful to very young children, who often lack the emotional or psychological maturity needed to cope with stressful life changes (Craig 1993). Family researchers consistently note that children who are subject to one or more family disruptions experience a number of particularly stressful events, including parental conflict, movement to a new home, loss of friends and relatives, and changes of adult caregivers (Amato 1993; McLanahan and Booth 1989; Wu and Martinson 1993). For example, family researchers estimate that nearly 40 percent of divorced mothers undergo a change in residence in the first year after their divorce (McLanahan and Booth 1989). The first two to three years following a family disruption are especially stressful for both parents and children. Many researchers find that during this period, financially and emotionally burdened single parents and their children experience a "chaotic lifestyle" markedby erraticmeals and bedtimes, inconsistent discipline patterns, shifting household routines, and tardiness and truancy at school (Cherlin 1992; Martinson and.Wu 1992; McLanahan and Booth 1989). Not surprisingly, parental divorce is one of the most stressful life events as measured on psychological scales (e.g., Sarason, Johnson, and Siegel 1978). According to the family-stressors hypothesis, the instability and change associated with family disruption can result in emotional and behavioral problems that are sometimes severe and long lasting. As Slater and Calhoun (1988) suggest,-"familial conflict, and not the divorce per se, may account for many of the behavioral and psychological problems noted in the literature" (p. 118). If these
315
various stressors are allowed to build up in an individual and are not vented in a healthy manner, they may lead to destructive or compulsive tendencies (An&shensel 1992; Hodges 1990). Compared with offspring living in intact two-parent families, children from mother-only families are more likely to drop out of school, have lower incomes, and engage in drug and alcohol abuse (Cherlin 1992; McLanahan and Booth 1989). Moreover, as a result of the stress surrounding disruption, many children suffer reduced self-esteem, are more distant in their familial relationships, and have difficulties with interpersonal interaction (Holdnack 1992; Young and Parish 1977). In sum, the accumulatedevidence suggests that family stressors serve as a mediating factor between disrupted-family structuresand both materialistic attitudes and compulsive consumption behaviors. H3a: The relationship between family structureand materialistic attitudes is mediated by young adults' perceived level of family stressors. H3b: The relationship between family structureand compulsive consumption is mediated by young adults' perceived level of family stressors.
This content downloaded from 188.26.232.243 on Mon, 4 Nov 2013 07:41:38 AM All use subject to JSTOR Terms and Conditions
316 psychologists consistently find that socioeconomic status is negatively associated with a host of stressful life events (Aneshensel 1992). H4a: The relationship between a disrupted family and family resources is moderated by socioeconomic status. The relationshipbetween disruption and resources will be stronger among subjects from families with lower socioeconomic status than among subjects from families with higher socioeconomic status. H4b: The relationship between a disrupted family and family stressors is moderatedby socioeconomic status. The relationshipbetween disruption and stressors will be stronger among subjects from families with lower socioeconomic status than among subjects from families with higher socioeconomic status. A graphical summaryof the relationships hypothesized in our study is provided in Figure 1.
Family
Hib (+)
Compulsive
Consumption
Family Resources
Disrupted
HU2a ()
Materialism
Ce s
Family
(+)
ModeratingEffects of SocioeconomicStatus Family Resources H4a Disrupted Family SocioEconomic Status H4b Family Stressors
final sample size of 65 respondents from this ad targeting persons from disrupted families. Our second ad (intact sample) generated 83 inquires, and we mailed out 69 questionnaires. Of these 69 potential respondents, 64 returnedsurveys, for a 93 percent response rate. After eliminating questionnaireswith missing data, we were left with a final sample size of 61 respondents from this ad targeting persons from intact families. In sum, our final sample composition from both the mail survey and the two newspaper ads consisted of 261 subjects, 165 from intact families (61 from the second newspaper ad and 104 from the mail survey) and 96 from disruptedfamilies (65 from the first newspaper ad and 31 from the mail survey). To examine the potential impact of using two different sampling approaches on respondent profiles, we conducted separate two-factor (sampling approach X family structure)univariate MANOVAs for each of our five key measures (i.e., family resources, family stressors, material values, compulsive buying, and socioeconomic status). These MANOVAS assessed the degree of explained variance in each of these measures attributedto differences
This content downloaded from 188.26.232.243 on Mon, 4 Nov 2013 07:41:38 AM All use subject to JSTOR Terms and Conditions
FAMILY STRUCTURE, MATERIALISM, AND COMPULSIVE CONSUMPTION in (1) sampling approach, (2) family structure, and (3) the sampling approach X structure interaction. Prior to conducting the MANOVAs, we first examined the means and standard deviations for each key measure for both the intact and disrupted groups across our two sampling approaches. Overall, our respondent profiles appear to be remarkably similar. The only significant method-related difference is the mean level of socioeconomic status among our disrupted subjects (mail survey X = -.55, newspaper ad X = -.17, p ? .03). We believe that this differefice in socioeconomic status largely reflects the fact that newspaper readership is generally associated with higher levels of socioeconomic standing (Bogart 1989). The MANOVA results indicate that (1) the familystructureeffect is significant (at p ? .00 1) for all five key measures, (2) the sampling approach effect is significant (at p ? .05) for only socioeconomic status, and (3) there are no significant interactionsfor any of the five measures. With the effect-size criterion established by Cohen (1977), the family-structure factor has a large effect on family resources (w2 = .18), family stressors (w2 = .47), material values (w02 = .17), andsocioeconomicstatus(co2 = .26) and a medium effect on compulsive buying (w02 = .09). Conversely, the sampling approachhad a nonsignificant effect on family resources, family stressors, material values, and compulsive buying, and only a small effect on socioeconomic status (w02 = .05). In sum, there do not appear to be any important systematic response differences between the mail survey and the newspaper ad. Thus, we conducted all of our analyses on the combined sample. Of our 261 subjects, 95 were male, 166 were female, 91 percent were white, and their mean age was 26.
317
Measurement
The survey instrument was a five-page questionnaire containing multi-item measures of materialism, compulsive consumption, family resources and family stressors, a single-item measure of family structure, a measure of socioeconomic status (SES), and a standardset of demographic variables. The items used in each of our new measures are detailed in the Appendix, and the summary statistics for all measures are reported in Table 1. In this section, we highlight the measurement approach for all of our measures. Materialism and Compulsive Consumption. Materialism was measured using the Richins and Dawson (1992) material values scale, with its three dimensions of centrality, happiness, and success. Compulsive consumption was assessed via the Faber and O'Guinn (1992) clinical screener for compulsive buying. Both measures used fivepoint Likert-type scales and demonstratedadequatelevels of reliability; the material values scale had an overall coefficient alpha of .87, while the clinical screener for compulsive buying had a coefficient alpha of .80. Family Resources and Family Stressors. Both family
resources and family stressors were assessed with fivepoint, Likert-type, multi-item scales developed for this study. Our family-resources measure is an eight-item scale that assessed perceived family resources by asking respondents to rate the amount and quality of various tangible (e.g., food and clothing) and intangible (e.g., guidance and emotional support) resources provided by their parents when they were growing up. Guidance for selection of these items came from the family sociology literature (Cherlin 1992; McLanahan and Booth 1989). We measured family stressors with a 10-item scale, adapted from an established instrument, the Life Experiences Survey (Sarason et al. 1978). Our family-resources and family-stressors scales demonstrated adequate reliability. The overall family-resources scale had a coefficient alpha of .90 (with an alpha of .81 for the tangible resources dimension and an alpha of .92 for the intangible resources dimension), and the family-stressors scale had a coefficient alpha of .76. We assessed our a priori conceptualization of the tangible/ intangible dimensionality of family resources via an exploratory factor analysis with a varimax rotation. This analysis revealed that our family resources scale did indeed have a two-factor solution, with three items (spending money, food, and clothing) loading cleanly on the first factor (tangible resources) with an average loading of .82, and the other five items (time and attention, discipline, life skills and instruction, emotional support and love, and role modeling and guidance) loading cleanly on the second factor (intangible resources) with an average loading of .83. Our family-resources and -stressors scales also exhibit a considerable degree of validity. First of all, our measures were drawn from a careful literaturereview to ensure an acceptable degree of content validity. In addition, we assessed constructvalidity by forcingall 13 individual items across both scales into a single-factor analysis. Ideally, if family resources and family stressors represent conceptually distinct constructs,they should load on separate factors. Using a varimax rotation for a two-factor solution, we found that all of the family-resource items loaded strongly on the first factor with rotated loadings ranging from .61 to .80, and all but one of the familystressor items (major, abruptchange in family's financial status) loaded cleanly on the second factor with rotated factor loadings ranging from .31 to .68. Thus, our measures of family resources and family stressors display reasonable discriminant validity. Finally, we examined the empirical validity of family resources and family stressors by comparing the mean resources and stressors of intact- versus disrupted-family subjects. As noted previously, the family sociology literature consistently finds that children from disrupted-family structures experience greater stress and are provided fewer resources than children from intact families. Therefore, one indicant of the validity of our measures is the degree to which they empirically demonstratethis pattern of stress elevation and resource depletion among our dis-
This content downloaded from 188.26.232.243 on Mon, 4 Nov 2013 07:41:38 AM All use subject to JSTOR Terms and Conditions
318
TABLE I KEY MEASURE STATISTICS Mean 1. 2. 3. 4. 5. 6. Family structure Family resources Family stressors Material values Compulsive buying Socioeconomic status (standardized) NA 3.70 1.51 2.82 2.06 .01 SD NA .88 .39 .88 .65 .67 1 (NA) -.40** .52** .21** .25** -.33** 2
among the measures are on the off-diagonal. NOTE.-NA,not applicable.The coefficientalphafor each measure is on the diagonaland the.intercorrelations at p ? .05. *Statistically significant at p ? .01. **Statistically significant
rupted subjects. A comparison of the mean levels of family resources and stressors across family structurereveals that our disrupted-family subjects report significantly lower levels of resources (disrupted X = 3.24, intact X = 3.97; t = 6.95, p ? .01) and higher levels of stress (disrupted X = 1.78, intact X = 1.36, p c .001) than subjects from intact-family backgrounds. Thus, our measures of family resources and family stressors appear to discriminate quite well on an empirical basis. Socioeconomic Status.. Our measure of respondents' socioeconomic status is based on a combination of three separate subdimensions: (1) perceived family wealth, (2) parentaleducation, and (3) housing status. We combined these variables into an overall measure of SES, by averaging subjects' summed standardizedscores for these measures (Ghiselli 1964). Because the three measures-parental education, housing, and perceived wealth represent three separate subdimensions of the SES construct, coefficient alpha is not calculated (Howell 1987, p. 121). Because of potential problems with subject recall of specific dollar amounts and lack of knowledge about their parent's finances, we employed a measure of perceived relative wealth in place of asking respondents to report their actual level of household income. According to O'Guinn and Wells (1989), subjective measures of economic well-being are useful predictors of consumer phenomena and in some cases may be better predictors than more objective income measures. Furthermore,a pretest showed that respondents have much greater confidence in their ability to accurately respond to questions about perceived wealth than to inquiries about household income at a specific point in time. Parentaleducation was measured by asking subjects to reportboth parents' educational attainment.In contrast to most family studies, which focus on the father's education, we believe that the educational attainment of each parent is likely to influence the effects of a disrupted family, especially-because mothers are awardedchild custody in nearly 90 percent of all divorce settlements (Wilkes 1995). For housing status, respondents were
asked to identify the type of dwelling they lived in at age 18. Because 82 percent of our respondents reported that they resided in single-family homes, we classified housing status as home ownership or nonownership.
This content downloaded from 188.26.232.243 on Mon, 4 Nov 2013 07:41:38 AM All use subject to JSTOR Terms and Conditions
319
Dependent variable Material values: Regression 1: Family resources Regression 2: Material values Regression 3: Material values Compulsive buying: Regression 1: Family resources Regression 2: Compulsive buying Regression 3: Compulsive buying
Coefficient
p-value
Family structure Family structure Family resources Family structure Family structure Family structure Family resources Family structure
the effect of the family structureon materialism or compulsive consumption is weaker in the thirdregression than i-nthe second. A case of perfect mediation is indicated if family structure has no effect when either mediator is included in the model. As can be seen in Tables 3 and 4, the test of the mediating effects of both family resources and family stressors for material values fails in the third regression outlined above, indicating that neither resources (p = -.10, p . 12) nor stressors (p = .05, p - .51) mediate the relationshipbetween family structureand material values. Thus, Hypotheses 2a and 3a are not supported. Conversely, the results of the mediating influence of both family resources and family stressors for compulsive buying do satisfy all of these conditions, indicating that both resources and stressors mediate the relationship between family structure and compulsive buying. However, because the inclusion of family resources reduces the beta coefficient associated with family structure by only 28 percent (from .25 to .18) and the inclusion of family stressors reduces the beta coefficient associated with family structure by only 24 percent (from .25 to .19), the mediating impact of both variables appears to be only partial. Hence, Hypotheses 2b and 3b are partially supported.
NOTE.-Forregression model 1, F = 48.30, p c .0001, R2 = .16. For regressionmodel 2, for material values, F = 12.20, p c .001, F2 = .05; for = .06. Forregressionmodel 3, compulsivebuying,F = 17.57, p c .0001, RF2 for materialvalues, F = 7.40, p c .001, RF2 = .06; for compulsive buying, F = 13.36, p c .0001, R2 = .10.
of moderation is performed by entering the individual predictor variables followed by the product term for their interaction. A moderating effect is confirmed by the presence of a significant two-way interaction. The results of this analysis are presented in Table 5. As recommended by Cronbach(1987), the independentvariables were mean centered in order to reduce the colinearity between the predictors and their product terms. The results indicate a significant moderating effect (X = .35, p - .02) of SES on the relationshipbetween family structureand family resources. Therefore, Hypothesis 4a is supported. However, the results failed to indicate a significant moderatingeffect (p = -.10, p ? .13) for SES on the relationship between family structure and family stressors. Thus, Hypothesis 4b is not supported. To furtherunderstandthe nature of the moderating influence of socioeconomic status on the relationship between a disrupted family and family resources, we conducted additional analyses as recommended by Aiken and West (1991). This approach,termed "simple slope analysis," involves creating the following conditional values for our SES variable: (1) high SES (more than one standard deviation above the SES mean), (2) moderate SES (within one standarddeviation above or below the mean), and (3) low SES (more than one standarddeviation below the mean). We used these conditional SES values to form interactionswith family structureand specified three separate regression models estimating the effects of family structure,the conditional values of SES, and their interaction on family resources. This slope analysis revealed that the effect of the family structureand SES interactionupon
This content downloaded from 188.26.232.243 on Mon, 4 Nov 2013 07:41:38 AM All use subject to JSTOR Terms and Conditions
320
TABLE 4 TEST OF THE MEDIATINGEFFECTS OF FAMILYSTRESSORS Independent variables
Dependent variable Material values: Regression 1: Family stressors Regression 2: Material values Regreqssion 3: Material values Compulsive buying: Regression 1: Family stressors Regression 2: Compulsive buying Regression 3: Compulsive buying
Coefficient
p-value
Family structure Family structure Family stressors Family structure Family structure Family structure Family stressors Family structure
the family-resources NOTE.-For regressionmodel,F = 41.66, p c .0001, R2 = .33. Forthe family-stressors regressionmodel, F = 37.93, p c .0001,
R2 = .31.
NOTE.-Forregression model 1, F = 95.98, p c .0001, R2 = .27. For regressionmodel 2, for material values, F = 12.20, p c .001, R2 = .05; for compulsivebuying,F = 17.57, p < .0001, R2 = .06. Forregressionmodel 3, for material values, F = 6.38, p c .002, R2 = .05; for compulsivebuying,F =11.58, p c .0001, R2 = .08.
family resources is strongest when SES is lowest (,B = 1.29, p s .05) and that the strength and significance of this relationship generally declines as SES increases from moderate (3 = .1 1, p < .09) to high levels (,B= .24, p c .75). In sum, this slope analysis suggests that young adults from more privileged socioeconomic backgrounds are less likely to have their family resources diminished following a family disruption than are young adults from families with lower-SES backgrounds.
DISCUSSION
Review of Our Findings
Direct Effects of Family Structure. Our findings provide considerable evidence that family structureis related to both materialism and compulsive consumption. Young adults whose parents were divorced or separated demonstratedhigher levels of both material values and compulsive buying than those whose parents were not divorced or separated.These direct relationships (see Table 2) support and extend the prior work of both consumer and family researchers. For example, Richins and Dawson (1992) suggest that people with high material values desire higher levels of income and financial security and place lower emphasis on personal relations. Compared with our intact-family sample, our disrupted-family subjects had lower perceived levels of household wealth (disrupted X = 2.47, intact X = 3.04, p ? .0001), which may encourage them to place a high degree of importanceon financial security.
Of the three dimensions in Richins and Dawson's (1992) material values scale, the biggest difference between our intact and disrupted-family samples (see Table 2) is in the dimension of centrality, that is, "the importance of acquisition and possession generally" (Richins and Dawson 1992, p. 309). This suggests that young adults from disruptedbackgroundsmay use material objects as surrogates for absent parents. This finding helps substantiate Belk's (1988) and Schouten's (1991) theoretical assertion that physical objects and acts of consumption can serve as importantreplacements for human contact. As for compulsive consumption, family sociologists have found that children from single-parent families are at higher risk of drug and alcohol abuse, smoking addiction, and other forms of compulsive behavior (Cherlin 1992; Flewelling and Bauman 1990; McLanahan and Booth 1989). Consumer researchers have also theorized a connection between growing up in a disrupted family and a propensity to engage in impulsivi and compulsive consumption (Hirschman 1992; Rook 1987). For example, Rook (1987) specifically suggests that, "the capacity for delaying gratification [is] correlated positively with age, intelligence, social responsibility, and the presence of a father in the home" (p. 190, italics added). Mediating Effects of Family Resources and Stressors. Our findings reveal that the relationship between a disrupted-familystructureand compulsive buying is partially mediated by both family resources and family stressors (see Tables 3 and 4). This is consistent with a growing literature base in family sociology which suggests that the by-products of disruption (such as resource reduction and stress escalation) are responsible for many of the consequences of parental divorce or separation. Both Amato (1993) and Hodges (1990) propose that family disruption leads to an increased level of family stress, which increases the likelihood of adverse consequences for children in these families. Likewise, divorce and separation typically reduce the amount and quality of both tangible and intangible family resources, which may in-
This content downloaded from 188.26.232.243 on Mon, 4 Nov 2013 07:41:38 AM All use subject to JSTOR Terms and Conditions
CONSUMPTION AND COMPULSIVE FAMILY STRUCTURE, MATERIALISM, crease the risk of compulsive consumption and other forms of addictive behavior. Our results suggest that some of the adverse consequences associated with disrupted families might be partially alleviated throughpolicies designed to reduce the stress and increase the resources available to children from such families. We found that, in contrast to the mediating effects for compulsive buying, neither family resources nor family stressorsmediate the relationshipbetween disruptedfamilies and material values. It may be that parental divorce or separation primarily exerts a direct and unmediated effect on young adults' material values. Both Belk (1985) and Wernimont and Fitzpatrick (1972) point out that an individual's level of concern with material objects varies across the life span and in response to various key life events such as entry into the workforce, marriage, and parenthood. It seems reasonable to suggest that family disruptionbelongs to this class of key life events and that this disruptionis capable of directly influencing a person's materialistic attitudes. Another possible explanation is a distinction between attitudes and behaviors. Perhaps attitudes such as material values, which are centrally held, are mediatedby internalfactors such as self-esteem (Richins and Dawson 1992). Psychology research has found evidence that children whose parents divorce or separate often have lower self-esteem, which may lead them to look to material objects to bolster their self-image (Holdnack 1992). To probe more deeply into this issue, we conducted separate tests of mediation for both types of family resources (tangible and intangible) as well as for all three dimensions of material values (centrality, happiness, and success). Intangible resources emerged as a significant partialmediator of the relationship between family structure and material values; the inclusion of intangible resources reduced the beta coefficient associated with family structure by 19 percent. We also found that the family-resources dimension significantly mediates the relationship between family structure and only the happiness subdimension of material values. This mediating effect was also partial, but substantial, as the inclusion of family resources reduces the beta coefficient by 35 percent. These findings suggest that it is the diminution of interpersonalresources such as love and affection, rather than financial resources, that links family disruption and materialism.In addition, the higher levels of material values evident among young adults from disrupted families may represent an instrumental desire to use material objects as surrogates for an absent parent. The fact that both family resources and family stressors act as mediators for compulsive buying and fail as mediators of material values suggests that the two variables (resources and stressors) may represent a single, higherorder construct. They are also significantly correlated (r = -.53, p - .0001). However, as discussed earlier, our measures of family resources and stressors display considerable discriminant validity and load largely on separate factors. In addition, their conceptual distinction is further
321
supportedby an apparentdifference in their antecedents. As our results indicate, the relationship between family structureand family resources is moderated by socioeconomic status, while the relationship between family structure and family stressors is not. As Amato (1993) notes, families that are low in resources are more likely to experience stress. Thus, these two constructs appear to be closely interrelatedyet remain conceptually distinct. This conceptual distinction has importantimplications for public policy. Although the provision of services such as family counseling and crisis interventionmay help relieve some of the stress in a disrupted family, it may do little to alleviate the problems created by a lack of family resources such as income and role modeling. Moderating Effects of Socioeconomic Status. Socioeconomic status moderates the impact of a disrupted-family structureupon family resources. This is largely consistent with previous findings in family sociology that social and economic factors attenuate,to some extent, the negative impacts of family disruption on children and young adults (e.g., Cherlin 1992; McLanahan 1985; McLanahan and Booth 1989). Our subgroupanalysis showed that children of lower SES are most at risk of experiencing a decline in family resources following a disruption.Therefore, public policy makers concerned with the diminution of resources typically associated with a divorce or separation would be well advised to focus their efforts on enhancing the resources of mothers and children from less privileged backgrounds. At first glance, the lack of support for the moderating effects of SES on the relationship between family disruption and stressors seems contraryto family sociology research, which usually finds that SES explains much of the relationship between family structureand various undesirable consequences (Amato 1993; McLanahan 1985). However, most of those studies have investigated behavioral phenomena known to be highly influenced by income and social status, such as completion of high school, teenage pregnancy, and criminal behavior. Perhapsfamily stressors belong to a class of phenomena that are relatively invariantacross socioeconomic strata.For example, many of the items measured in our family-stressors scale (e.g., difficulties establishing or maintainingrelationships with peers, arguments between parents or other family members) appear to be equally likely to afflict young adults of any status level. Thus, higher levels of household income, home ownership, and higher levels of education may do little to shield children from the stressful events that normally accompany family disruption.
This content downloaded from 188.26.232.243 on Mon, 4 Nov 2013 07:41:38 AM All use subject to JSTOR Terms and Conditions
322
JOURNAL OF CONSUMER RESEARCH assessing the degree of racial diversity in the relationship between family disruption and consumer behaviors and attitudes. Processes of Family Disruption. Our study focuses mainly on the structural aspects of family life. Equally adverse consequences can affect children who live in intact households characterized by high levels of tension and marital conflict (Cherlin 1992; Demo 1993). Family structure is in many ways just an indicant of the social dynamics that underlie various family forms. Future research of the underlying processes and situational dynamics across a broad range of family structuresis needed to better understandwhy being reared in a disrupted family appearsto promote specific consumer attitudesand behaviors. In particular, this research needs to determine to what extent these effects are due to the family disruption itself or are merely factors connected with the processes of parental conflict, loss of income, or movement to a poorer neighborhood (Demo 1993). As Kitson and Morgan (1990) note, the family research literature needs to develop more effective means of establishing causality between family disruption and its purported consequences. The issue of family disruptionraises a number of questions related to the frequency, content, and structure of parent-child relations pertaining to consumption. For example, both parental style (Carlson and Grossbart 1988) and parent-childcommunication patterns (Moschis 1987) have been found to be key determinantsof the outcomes of consumer socialization. How does family disruption impact parental style and consumer socialization? What type of consumption-orientedcommunication patternsexist in families across various stages of disruption (i.e., predisruption,divorce or separation, postdisruption)?As an early inquiry into these types of issues, Kurdek and Fine (1993) found that adolescents livingwvitha divorced father or a multiply divorced parent report more permissive parentingstyles than adolescents rearedin intact families. The processes under which a divorce or permanent separationunfolds appear to have importantimplications on its subsequentconsequences. For example, sudden and unexpected separations may result in more emotional problems and lower levels of adjustmentfor both parents and children, compared with separationsthat are expected well in advance (Price-Bonham and Balswick 1980; Wallerstein and Kelly 1980). Likewise, some family sociologists suggest that both the age at initial disruption and the length of time spent in a disrupted family may be related to the ability of children to cope with this stressful life transition(Wu and Martinson 1993). Finally, research suggests that highly conflict-riddenfamily disruptionsare more likely to lead to compulsive and destructive consumption activities than are more amiable family transitions (Hirschman 1992; Faber and O'Guinn 1992). Although detailed investigation of the process of family disruption may be difficult, consumer researchers should
first empirical inquiry into the long-term impact of a disrupted-family structureon the consumption attitudes and behavior of children, this study opens the door for future investigations of how family structure influences other consumption issues. To provide a guidepost for future research into the relationship between family structure and consumer behavior, an agenda of these key issues is outlined below, focusing on the antecedents, processes, and consequences of family disruption as they relate to consumer research. Antecedents of Family Disruption. Family researchers rely on a variety of theoretical perspectives to understand and explain the causes of family disruption (see White 1990). Perhaps the most intriguing theoretical perspective for consumer and marketingresearchersis Becker's (1981) Nobel Prize-winning research on the use of exchange theory to understandmarital formation and dissolution. According to this perspective, divorce is more likely when costs are low (e.g., couples without children) ahd when alternatives are high (e.g., younger couples, higher incomes). Because exchange theory provides a key foundation for much of our literature,consumer and mar-keting researchers appear to be in a unique position to enrich and extend this exchange perspective with regard to family disruption. Bagozzi's (1975, 1978) programmatic research on the determinantsof exchange relationships (i.e., characteristics of social actors, actors' social influence, the role of third parties, and situational contingencies) might be useful in understanding the nature of marital relationships. For example, consumer researchers could use Bagozzi's framework to investigate how the growth of our consumer culture may affect family disruption by reducing our reliance on the family for social position and sense of worth. Like most research on family disruption, our study largely employs a sample of white subjects. Although this sample certainly does not negate the implications of our findings, it may not be a wholly representative portrayal of family life. For example, blacks have historically had higher rates of marital dissolution than whites (Cherlin 1992; White 1990); although only 40 percent of white children will experience disruption before the age of 16, approximately75 percent of all black children will (Bumpass 1984). Furthermore,there are importantracial differences in the components of disruption. As Wilson (1987) observes, "Whereas white women are far more likely to be divorced than separated,black women are more likely to be separated than divorced" (p. 68). Compared with divorce, separation is a much more informal structural arrangementand may be even more stressful due to the uncertainty and indeterminacy of being neither married nor divorced. Given the substantial differences in family structuresbetween black and whites, it seems reasonable to expect that the relationship between family disruption and consumer -behaviormay also differ. A replication of our study among young adults from various ethnic and racial groups would provide an importantcontribution in
This content downloaded from 188.26.232.243 on Mon, 4 Nov 2013 07:41:38 AM All use subject to JSTOR Terms and Conditions
CONSUMPTION AND COMPULSIVE FAMILY MATERIALISM, STRUCTURE, be aware of the family disruption process and how it might affect consumption patterns and priorities. Consequences of Family Disruption. Traditionally, family researchers have regarded family disruption as a form of deviant and undesirable behavior (Copeland and White 1991). However, as half of today's children are expected to experience divorce or separationof their parents, family disruption has "become a standard part of the American family experience" (White 1990, p. 904). Furthermore, although family disruption is associated with a.hqst of undesirable life consequences, most people who grow up in a single-parent family do not experience these negative outcomes (Cherlin 1992). As Amato (1993) observes, "A great deal of variability is present among children of divorce, with some experiencing problems and others adjusting well or even showing improvements in behavior" (p. 23). It is clear that research must extend beyond just structuralissues in understandingthe consumption outcomes associated with family disruption. For example, to what extent do differences in self-esteem, body image, and other characteristics mediate the relationship between family structure and consumption-related consequences? Moreover, researchers are advised to consider both the potentially positive and negative consequences of alternative family structures. We suggest that researchers should examine the extent to which a family disruptionfavorably affects the consumption attitudes and behavior of children who would otherwise live in an intact, yet conflict-ridden household. Sociologists have long noted the need for further research into the adaptive mechanisms that facilitate the adjustmentof family members to divorce and separation (Berman and Turk 1981). According to Kitson and Morgan (1990), researchers should consider how children of divorced families adjust both economically and socially to their new lives. Thus, ratherthan view family members as passive reactors buffeted by divorce and separation, this adjustment perspective encourages researchers to consider family members as agents who actively attempt to reduce the stress and uncertaintyassociated with family disruptionby enacting a variety of coping strategies. Our results raise the possibility that elevated material values and/or compulsive buying might serve an instrumental role in helping these young adults cope with the stress and uncertainty associated with family disruption. These behaviors might even help young adults develop an ability to deal more effectively with subsequent life transitions. Furtherinvestigation into the role of these and other consumption attitudes and activities in helping individuals reduce or alleviate the stress associated with family disruption and other significant role transitions seems to be an especially intriguing avenue for future research.
323
My evaluation of the total support (quantity and quality) provided by my family while I was growing up, breaks down as follows: 1. Spending money 2. Food 3. Clothing 4. Time and attention 5. Discipline 6. Life skills and instruction 7. Emotional support and love 8. Role modeling and guidance Family stressors (adapted from Family Stressors. Sarason et al. [1978]) were measured on a five-point scale (1 = strongly negative and 5 = strongly positive). Considering up through your 18th birthday,please circle the overall extent to which each of the following events impacted your life (positively or negatively) around the time(s) they occurred. Use "No Impact" if the event made no difference or never occurred. 1. Move(s) to a different home or place of residence 2. Difficulties with school work 3. A major, abruptchange in your family's financial status 4. Frequent or lengthy periods in which one or both parents were temporarily absent 5. Difficulties establishing and/or maintaining social relationships with peers 6. The loss (other than death) or separation from family members or loved ones 7. Encounters with juvenile authorities or police 8. Physical abuse by parents or other family members 9. Arguments between parents or other family members (including self) 10. Changes in the membership or composition of your family unit other than the divorce of your parents (e.g., remarriage of your parent(s), birth of your own child, etc.) Socioeconomic Status. measure. Socioeconomic status is a new
1. Please give the highest level of completed education for both your mother and father. 2. At the time of your 18th birthday, what type of dwelling did you live in? 3. In your overall estimation, how financially well off was your household when you were growing up (up to your 18th birthday)?
This content downloaded from 188.26.232.243 on Mon, 4 Nov 2013 07:41:38 AM All use subject to JSTOR Terms and Conditions
324
[Received October 1995. Revised September 1996. Brian Sternthal served as editor and John F. Sherry, Jr., served as associate editor for this article.]
REFERENCES
Ahlburg, Dennis A. and Carol J. De Vita (1992), New Realities of the American Family, Population Bulletin no. 47, Washington, DC: Population Reference Bureau. Ahuja, Roshan. and Kandi Stinson (1993), "Female Headed Single Parent Families: An Exploratory Study of Children's Age in Family Decision Making," Advances in Consumer Research, Vol. 20, ed. Leigh McAlister and Michael L. Rothschild, Provo, UT: Association for Consumer Research, 469-474. Aiken, Leona S. and Stephen G. West (1991), Multiple Regression: Testing and InterpretingInteractions, Newbury Park, CA: Sage. Amato, Paul R. (1993), "Children's Adjustment to Divorce: Theories, Hypotheses, and Empirical Support," Journal of Marriage and the Family, 55 (February), 23-38. An, Chong-Bum, Robert Haveman, and BarbaraWolfe (1993), "Teen Out-of-Wedlock Births and Welfare Receipt: The Role of Childhood Events and Economic Circumstances," Review of Economics and Statistics, 75 (May), 195-208. Aneshensel, Carol S. (1992), "Social Stress: Theory and Research," Annual Review of Sociology, 18, 15-38. Bagozzi, RichardP. (1975), "Marketing as Exchange," Journal of Marketing, 39 (October), 32-39. (1978), "Marketing as Exchange: A Theory of Transactions in the Marketplace," American Behavioral Scientist, 21 (March/April), 535-556. Bandura, Albert and Walter Mischel (1965), "Modification of Self-Imposed Delay of Reward through Exposure to Live and Symbolic Models," Journal of Personality and Social Psychology, 2 (November), 698-705. Baron, Reuben M. and David A. Kenny (1986), "The Moderator-Mediator Variable Distinction in Social Psychological Research: Conceptual, Strategic, and Statistical Considerations," Journal of Personality and Social Psychology, 51 (December), 1173-1182. Becker, Gary S. (1981), A Treatise on the Family, Cambridge, MA: Harvard University Press. Belk, Russell W. (1985), "Materialism: Trait Aspects of Living in the Material World," Journal of Consumer Research, 12 (December), 265-280. (1988), "Possessions and the Extended Self," Journal of Consumer Research, 15 (September), 139-168. Berman, William H. and Dennis C. Turk (1981), "Adaptation
This content downloaded from 188.26.232.243 on Mon, 4 Nov 2013 07:41:38 AM All use subject to JSTOR Terms and Conditions
325
Rook, Dennis (1987), "The Buying Impulse," Journal of Consumer Research, 14 (September), 189-199. Sarason, Irwin G., James H. Johnson, and Judith M. Siegel (1978), "Assessing the Impact of Life Changes: Development of the Life Experiences Survey," Journal of Consulting and Clinical Psychology, 46 (October), 932-946. Schouten, John W. (1991), "Selves in Transition: Symbolic Consumption in Personal Rites of Passage and Identity Reconstruction," Journal of Consumer Research, 17 (March), 412-425. Seltzer, Judith A. and Suzanne M. Bianchi (1988), "Children's Contact with Absent Parents," Journal of Marriage and the Family, 50 (August), 663-677. Slater, Elisa J. and Karen S. Calhoun (1988), "Familial Conflict and Marital Dissolution: Effects on the Social Functioning of College Students," Journal of Social and Clinical Psychology, 6 (Spring), 118-126. Thompson, Elizabeth, Thomas Hanson, and Sara S. McLanahan (1992), "Family Structures and Child Well-Being: Economic Resources vs. ParentSocialization," National Study of Families and Households Working Paper No. 29, University of Wisconsin-Madison, Madison, WI 53706. Wallerstein, Judith S. and Joan Berlin Kelly (1980), Surviving the Breakup: How Children and Parents Cope with Divorce, New York: Basic. Walls, Richard T. and Tennie S. Smith (1970), "Development of Preferences for Delayed Reinforcement in Disadvantaged Children," Journal of Educational Psychology, 62 (June), 118-123. Wernimont, Paul F. and Susan Fitzpatrick (1972), "The Meaning of Money," Journal of Applied Psychology, 56 (June), 218-226. White, Lynn K. (1990), "Determinants of Divorce: A Review of Research in the Eighties," Journal of Marriage and the Family, 52 (November), 904-912. Wilkes, Robert E. (1995), "Redefining Family in America: Characteristics and Expenditures of Single-Parent Families," in Marketing Theory and Applications, Vol. 6, ed. David Stewart and Naufel Vilcassim, Chicago: American Marketing Association, 270-276. Wilson, William Julius (1987), The Truly Disadvantaged, Chicago: University of Chicago Press. Wu, Lawrence L. and Brian C. Martinson (1993), "Family Structure and the Risk of Premarital Births," American Sociological Review, 58 (April), 210-232. Yankelovich, Daniel (1982), New Rules: Searching for SelfFulfillment in a World Turned Upside Down, New York: Bantam New Age. Young, E. R. and T. S. Parish (1977), "Impact of Father Absence during Childhood on the Psychological Adjustment of College Females," Sex Roles, 3, 217-227.
This content downloaded from 188.26.232.243 on Mon, 4 Nov 2013 07:41:38 AM All use subject to JSTOR Terms and Conditions