4
Developmental Psychology 1999, Vol. 35, No. 4, 1020-1023 Copyright 1999 by the American Psychological Association, Inc. 0012-1649/99/S3.00 Parental Language and Verbal Responsiveness to Children in Crowded Homes Gary W. Evans and Lorraine E. Maxwell Cornell University Betty Hart University of Kansas This article is a secondary data analysis of the University of Kansas Language Acquisition Project, which intensively studied, on a regular basis, parent and child language from age 6 months to 30 months. The association between residential density and parent-child speech was examined. Parents in crowded homes speak in less complex, sophisticated ways with their children compared with parents in uncrowded homes, and this association is mediated by parental responsiveness. Parents in more crowded homes are less verbally responsive to their children. This in turn accounts for their simpler, less sophisticated speech to their children. This mediational pathway is evident with statistical controls for socioeconomic status. This model may help explain prior findings showing a link between residential crowding and delayed cognitive development. Several field studies and a few laboratory experiments indicate that crowding is associated with negative socioemotional and cognitive outcomes in young children. Negative outcomes have included elevated psychophysiological stress (Aiello, Nicosia, & Thompson, 1979; Evans, Lepore, Shejwal, & Palsane, 1998), greater behavioral problems (Booth & Johnson, 1975; Saegert, 1982), and delayed cognitive development (Gottfried & Gottfried, 1984; Murray, 1974; Saegert, 1982; Wachs & Gruen, 1982; Wedge & Petzing, 1970). The present study builds on these findings by addressing an important question: What proximal processes are invoked by high-density living conditions that could precipitate such adverse developmental outcomes? Proximal processes are reciprocal interactions between the developing organism and per- sons, objects, and symbols in the immediate environment (Bron- fenbrenner & Morris, 1998). The present study presents a repre- sentative sampling, beginning in infancy until Age 3, of typical, everyday parent-child verbal interactions in the homes of families varying in residential density. We explored the hypothesis that parents in more crowded homes are less verbally responsive to their children. We also examined evidence of a potentially injurious correlate of parental unrespon- Gary W. Evans and Lorraine E. Maxwell, Department of Design and Environmental Analysis, Cornell University; Betty Hart, Schiefelbusch Institute for Life Span Studies, University of Kansas. Preparation of this article was partially supported by the U.S. Depart- ment of Agriculture (Hatch Grants NY 327407 and 327408) and the National Institute of Child Health and Human Development (Grant 1 F33 HD08473-01). We thank Todd Risley, coprincipal investigator on the original Kansas Language Acquisition Project, from which this article is a secondary data analysis. We also thank Steve Lepore for critical feedback on an earlier version of the article. Correspondence concerning this article should be addressed to Gary W. Evans, Department of Design and Environmental Analysis, Martha Van Rensselaer Hall, Cornell University, Ithaca, New York 14853-4401. Elec- tronic mail may be sent to [email protected]. siveness: impoverished parent-to-child speech. The latter, which is strongly linked to cognitive development, could further our under- standing of previous developmental studies linking residential crowding to delayed cognitive development. Adults in crowded homes cope with the plethora of unwanted social interaction in their residences by socially withdrawing from one another (Baum, Gatchel, Aiello, & Thompson, 1981; Baum & Valins, 1979). This withdrawal leads to impaired social support among those who live in crowded homes (Evans & Lepore, 1993; Evans, Palsane, Lepore, & Martin, 1989; Lepore, Evans, & Schneider, 1991). Moreover, parents in more crowded homes are rated as less responsive to their children in comparison with those from less crowded homes (Bradley & Caldwell, 1984; Wachs, 1989; Wachs & Camli, 1991). Residential crowding is also posi- tively correlated to the degree of chaos in homes, which, in turn, is negatively associated with parental responsiveness to infants (Matheny, Wachs, Ludwig, & Phillips, 1995). Moreover, the amount of parent-to-infant speech is negatively correlated with residential crowding (Wachs & Camli, 1991). Thus, we have evidence that parents in more crowded homes are less responsive to their children and talk to them less. We do not know, however, whether the quality of parental speech is also related to crowding, nor do we know whether the relations between crowding and parental speech are mediated by parental respon- siveness. Furthermore, none of these prior studies measured parent-to-child speech or responsiveness continuously over ex- tended periods of time. Instead, observers coded these behaviors once at 12 months of age (Matheny et al., 1995; Wachs & Camli, 1991) or at 6-month (Wachs, 1989) or 12-month intervals (Bradley & Caldwell, 1984). Among the presumably unintended consequences of diminished parental verbal responsiveness to children might be impoverished speech to children. Perhaps under more crowded living conditions, parents not only speak to their children less often (Wachs & Camli, 1991) but also speak with reduced complexity and sophistication, in comparison with parents residing under more spacious condi- 1020

Lamp3

Embed Size (px)

DESCRIPTION

lampiran

Citation preview

Page 1: Lamp3

Developmental Psychology1999, Vol. 35, No. 4, 1020-1023

Copyright 1999 by the American Psychological Association, Inc.0012-1649/99/S3.00

Parental Language and Verbal Responsiveness to Childrenin Crowded Homes

Gary W. Evans and Lorraine E. MaxwellCornell University

Betty HartUniversity of Kansas

This article is a secondary data analysis of the University of Kansas Language Acquisition Project, whichintensively studied, on a regular basis, parent and child language from age 6 months to 30 months. Theassociation between residential density and parent-child speech was examined. Parents in crowdedhomes speak in less complex, sophisticated ways with their children compared with parents in uncrowdedhomes, and this association is mediated by parental responsiveness. Parents in more crowded homes areless verbally responsive to their children. This in turn accounts for their simpler, less sophisticated speechto their children. This mediational pathway is evident with statistical controls for socioeconomic status.This model may help explain prior findings showing a link between residential crowding and delayedcognitive development.

Several field studies and a few laboratory experiments indicatethat crowding is associated with negative socioemotional andcognitive outcomes in young children. Negative outcomes haveincluded elevated psychophysiological stress (Aiello, Nicosia, &Thompson, 1979; Evans, Lepore, Shejwal, & Palsane, 1998),greater behavioral problems (Booth & Johnson, 1975; Saegert,1982), and delayed cognitive development (Gottfried & Gottfried,1984; Murray, 1974; Saegert, 1982; Wachs & Gruen, 1982; Wedge& Petzing, 1970). The present study builds on these findings byaddressing an important question: What proximal processes areinvoked by high-density living conditions that could precipitatesuch adverse developmental outcomes? Proximal processes arereciprocal interactions between the developing organism and per-sons, objects, and symbols in the immediate environment (Bron-fenbrenner & Morris, 1998). The present study presents a repre-sentative sampling, beginning in infancy until Age 3, of typical,everyday parent-child verbal interactions in the homes of familiesvarying in residential density.

We explored the hypothesis that parents in more crowded homesare less verbally responsive to their children. We also examinedevidence of a potentially injurious correlate of parental unrespon-

Gary W. Evans and Lorraine E. Maxwell, Department of Design andEnvironmental Analysis, Cornell University; Betty Hart, SchiefelbuschInstitute for Life Span Studies, University of Kansas.

Preparation of this article was partially supported by the U.S. Depart-ment of Agriculture (Hatch Grants NY 327407 and 327408) and theNational Institute of Child Health and Human Development (Grant 1 F33HD08473-01).

We thank Todd Risley, coprincipal investigator on the original KansasLanguage Acquisition Project, from which this article is a secondary dataanalysis. We also thank Steve Lepore for critical feedback on an earlierversion of the article.

Correspondence concerning this article should be addressed to Gary W.Evans, Department of Design and Environmental Analysis, Martha VanRensselaer Hall, Cornell University, Ithaca, New York 14853-4401. Elec-tronic mail may be sent to [email protected].

siveness: impoverished parent-to-child speech. The latter, which isstrongly linked to cognitive development, could further our under-standing of previous developmental studies linking residentialcrowding to delayed cognitive development.

Adults in crowded homes cope with the plethora of unwantedsocial interaction in their residences by socially withdrawing fromone another (Baum, Gatchel, Aiello, & Thompson, 1981; Baum &Valins, 1979). This withdrawal leads to impaired social supportamong those who live in crowded homes (Evans & Lepore, 1993;Evans, Palsane, Lepore, & Martin, 1989; Lepore, Evans, &Schneider, 1991). Moreover, parents in more crowded homes arerated as less responsive to their children in comparison with thosefrom less crowded homes (Bradley & Caldwell, 1984; Wachs,1989; Wachs & Camli, 1991). Residential crowding is also posi-tively correlated to the degree of chaos in homes, which, in turn,is negatively associated with parental responsiveness to infants(Matheny, Wachs, Ludwig, & Phillips, 1995). Moreover, theamount of parent-to-infant speech is negatively correlated withresidential crowding (Wachs & Camli, 1991).

Thus, we have evidence that parents in more crowded homes areless responsive to their children and talk to them less. We do notknow, however, whether the quality of parental speech is alsorelated to crowding, nor do we know whether the relations betweencrowding and parental speech are mediated by parental respon-siveness. Furthermore, none of these prior studies measuredparent-to-child speech or responsiveness continuously over ex-tended periods of time. Instead, observers coded these behaviorsonce at 12 months of age (Matheny et al., 1995; Wachs & Camli,1991) or at 6-month (Wachs, 1989) or 12-month intervals (Bradley& Caldwell, 1984).

Among the presumably unintended consequences of diminishedparental verbal responsiveness to children might be impoverishedspeech to children. Perhaps under more crowded living conditions,parents not only speak to their children less often (Wachs & Camli,1991) but also speak with reduced complexity and sophistication,in comparison with parents residing under more spacious condi-

1020

Page 2: Lamp3

CROWDING AND PARENTAL LANGUAGE 1021

tions. Thus, inhibited parental verbal responsiveness in reaction tocrowding could lead to impoverished verbal communication in thedwellings of crowded families. We were able to evaluate both ofthese hypotheses with a secondary data analysis of a longitudinaldata set on parental language. This data set includes extensive, invivo observations over a 2.5-year period of everyday parent-childinteraction, beginning at age 6 months. In accordance with priorresearch, we expected less parental verbal responsiveness amongfamilies living in crowded versus uncrowded conditions. We werealso able to test whether this increased unresponsiveness might, inturn, account for lower quality, more impoverished parent-childspeech in crowded families.

Method

Participants

Participants were 42 midwestern American families (31% upper, 24%middle, and 43% lower socioeconomic status [SES]), 60% Caucasian,ranging in family size from 1 to 5 children (M = 1.69 children). Familieswere recruited primarily from birth records and referrals from various localagencies in the Kansas City, Kansas, area. The target child was observedbeginning at about 6 months until Age 3. More details on the sample areavailable in Hart and Risley (1995).

Procedure

The primary goal of the original study, from which the present article isa secondary analysis, was to investigate how children learn to talk in thecourse of everyday, typical social interaction in the home. Families wereencouraged to engage in whatever typical behaviors they would normallydo over the course of each 1-hr observation period in the home. At no timedid the observer engage in conversation with the target child. Observationsoccurred monthly over a period of 2.5 years. Evidence in Hart and Risley(1995) indicated that the investigators successfully established sufficienttrust and familiarity with participating families so that unstructured, unob-trusive observation was achieved.

Two variables from Hart and Risley's (1995) set of language measureswere extracted for the present analysis. They were supplemented with anindex of crowding: people per room. A room was defined by the presenceof a floor-to-ceiling wall with a defined entry separating the area from ahall or another room. Because of the natural covariation of crowding withSES and the latter's documented impacts on parent-child verbal interac-tion (Hart & Risley, 1995), SES was used as a statistical control in all ofthe analyses herein. SES was assigned according to census occupationalcodes that represent income and educational attributes of the particular jobin the national labor force (Stevens & Cho, 1985).

The measure of parental language used in the present analysis waslanguage diversity. Language diversity is the sum of different nouns,adjectives, and adverbs addressed to the child per hour. Each of thesecomponents has an interrater reliability of 1.0. All reliability estimateswere based on 190 randomly selected samples. This measure of parentallanguage represents the richness and variety of experiences that parentstalk about with their children. It reflects the complexity and sophisticationof parent-to-child verbalization.

Parental verbal responsiveness is the proportion of all single parentalverbal utterances to the child within 5 s of the target child's behavior notpreceded by parental verbal initiation. Any child behavior that initiated aparental verbal response was counted. These behaviors could be verbal ornonverbal, which, of course, varied with the child's age. A parentalresponse could be nonverbal (e.g., the parent gave the child a requestedobject without saying anything), but this rarely happened. The number ofparent-initiated verbalizations to the child was subtracted from all parental

verbalizations to the child per hour. This number was then divided by thetotal number of parental verbalizations during the 1-hr period. Interraterreliability for each of these two components exceeded .95. Parental verbalresponsiveness is indicative of the relative amount of parent-to-childspeech that is controlled by the child. It exemplifies parental receptivity tothe child's needs and interests. For more information on the validity of thelanguage diversity and parental verbal responsiveness indexes, see Hartand Risley (1995).

Results

To examine the hypothesized association between crowding andparental verbal responsiveness and the latter's association with thequality of parent-to-child verbal communication, we provide de-scriptive data on these variables. We then discuss the results of aseries of regression analyses. As can be seen in Table 1, the patternof data suggests that parents in more crowded homes are both lessresponsive and converse with their children in less complex ways.Note that density is depicted in quartiles in Table 1 for descriptivepurposes only. Inferential analyses that follow maintain the con-tinuous nature of the density measure, people per room.

To investigate the proposed interrelations among density, pa-rental responsiveness, and parent-to-child speech more fully, weused a series of regression equations with statistical control forSES. The importance of controlling for SES is evident in thecorrelation matrix (see Table 2).

Children in more crowded homes were exposed to significantlyless diverse language in comparison with their counterparts inlow-density homes (see Table 3, row 2). This relation held inde-pendent of social class, as shown by the significant increment inthe change in R2 for the density term when entered after the SESterm in the hierarchical regression equation (ARZ = .07). Parentalverbal responsiveness was also a significant predictor of languagediversity in the home, after controlling for SES. Greater verbalresponsiveness to children's speech was positively associated withhigh levels of language diversity uttered by parents to their chil-dren, irrespective of social class.

Table 3 (row 4) is of most direct interest in evaluating whetherthe reduction in parental responsiveness associated with crowdingmight, in turn, explain more impoverished parental speech amongmore crowded families. Examination of the F ratio for the incre-ment in R2 for the density term (A/?2 = .02) shows that density wasno longer a significant predictor of language diversity after thecovariation with parental responsiveness was partialed out of theequation (Table 3, row 4). The previously significant relationbetween density and language diversity (Table 3, row 2) was nolonger significant when the covariation between density and pa-

Table 1Descriptive Information on Density, Parental Responsiveness,and Language Diversity

Variable

Parental responsivenessLanguage diversity

<0.50

0.81139.22

Density (people/room)

0.50-0.59

0.78114.58

0.60-0.71

0.7696.67

>0.71

0.6658.17

Note. Density quartiles are shown for descriptive purposes only. Allanalyses used density as a continuous variable.

Page 3: Lamp3

1022 EVANS, MAXWELL, AND HART

Table 2Correlation Matrix for Variables Included inthe Regression Analyses

Variable 1 2

1. Socioeconomic status2. Density3. Verbal responsiveness4. Language diversity

— - .55* .50**- . 5 1 * *

.68**-.60**

.69**

** p < .001.

rental responsiveness was statistically removed. The relation be-tween density and parental verbal responsiveness, after controllingfor SES, was significant (Afl2 = .08), F(2, 39) = 4.59, p < .03.1

This pattern of results indicates that the negative relation betweenhousehold crowding and language diversity is at least partiallyattributable to reduced parental responsiveness in crowded homes.

Discussion

The diversity and sophistication of parental speech is lower inmore crowded households. This association is statistically inde-pendent of SES. The link between household crowding and paren-tal speech quality, in turn, appears to be mediated by parents'verbal responsiveness to their children. Parental speech is impov-erished in more crowded homes because parents are less respon-sive to their children's behavior.

We have uncovered a plausible mechanism to explain previ-ously documented relations between household crowding and poorcognitive development in infants (Gottfried & Gottfried, 1984;Wachs & Camli, 1991; Wachs & Gruen, 1982) and in elementary-age school children (Murray, 1974; Saegert, 1982; Wedge &Petzing, 1970). The finding that density is positively associatedwith reduced parental verbal responsiveness to children is alsoconsistent with previous observers' ratings of lower parental re-sponsiveness in high- compared with low-density households(Bradley & Caldwell, 1984; Wachs, 1989). Our findings are alsoconsistent with Matheny et al. (1995), who noted that crowdedhomes were more chaotic and that the degree of chaos was relatedto less parental responsiveness to infants.

The present study makes several unique contributions to re-search on crowding and human development. Foremost, these datarepresent the only in vivo, sustained longitudinal analysis of familyinteractions in homes of varying density. Although many investi-gators have examined the developmental correlates of chronicresidential crowding, no research has observed actual family in-teractions over time in crowded homes (Aiello, Thompson, &Baum, 1985; Baum & Paulus, 1987). Second, we show for the firsttime that quality of parent-to-child speech is related to density.Third, we have investigated a mediational pathway showing thatspeech quality appears to degrade in more crowded homes inrelation to diminished parental verbal responsiveness. Diminishedparental verbal responsiveness is consistent with the more generaltheoretical proposition that crowding is harmful because peoplelearn to cope with high residential density by social withdrawal.Unfortunately, social withdrawal becomes an overgeneralizedstrategy that is applied indiscriminately, negatively affecting socialsupport (Baum & Valins, 1979; Baum et al., 1981; Evans &

Lepore, 1993; Evans et al., 1989; Lepore et al., 1991) and possibly,as suggested herein, parents' responsiveness to their children.

Because the conclusions of this study are derived from correla-tional data, it is important that the findings be replicated in aprospective, longitudinal design. Although we have ruled out SESas a plausible rival hypothesis, it is always possible that some otherfactor related to self-selection into housing could be driving theassociations uncovered herein. Although we can rule out SES asindicated by normative income and educational attributes of par-ents' job classifications, it is possible that some other variable orvariables could be driving the present results. Given the strongconnection between parental language style and education, werepeated all of the analyses herein with additional statistical con-trols for the years of education of the mother and the father. Thisdid not change any of our conclusions. Nonetheless, caution iswarranted until our findings can be replicated in a prospective,longitudinal design.

Because the original intent of the data collection for this studywas unrelated to crowding, the range of density sampled here isrelatively restricted (range = 0.30-1.25 people/room). Thus, itwould also be valuable to examine parental language and verbalresponsiveness among families across a wider range of householddensities. This might also further the important objective of iden-tifying at what point density begins to affect children negatively.Inspection of Table 1 suggests that somewhere above 0.71 personsper room, marked drops in parental responsiveness occur withincreasingly serious declines in parental language quality.

One plausible, alternative explanation for why residentialcrowding is correlated with poorer quality parental speech issimply that parents in more crowded homes have less time andenergy to give to each child. As the ratio of the number of childrento the number of adults goes up, parents are less responsive perchild.2 Thus, we created the ratio of number of children pernumber of adults in each of our 42 households to examine thisalternative explanation for our crowding results. The ratio ofnumber of children to parents in the home was not related toparental speech quality. The well-documented, negative cognitivecorrelates of household density (Gottfried & Gottfried, 1984; Mur-ray, 1974; Saegert, 1982; Wachs & Gruen, 1982; Wedge & Petz-ing, 1970) are not attributable to divided parental attention tochildren. We propose that crowding is related to diminished pa-rental verbal responsiveness, which leads to impoverished parent-to-child speech, because families cope with the surfeit of unwantedsocial interaction accompanying crowding by socially withdraw-ing from one another.

In summary, independent of SES, residential crowding is sig-nificantly correlated with parents' verbal responsiveness to their

1 As a partial check on spuriousness, we reversed the order of inclusionfor the hypothetical mediator (parental responsiveness) and density. If oneor more unspecified variables accounted for the relation between densityand parental responsiveness, reversing the order of terms in the fourthregression equation of Table 3 would not have affected the regressionresults. However, reversing the terms in the fourth regression equation ofTable 3 did substantially change the results. Parental responsiveness withthe SES control was significantly related to language diversity (b = .43)after density was partialed out, F(3, 38) = 13.55, p < .001. These resultsdo not support the alternative hypothesis of spuriousness.

2 We acknowledge an anonymous reviewer for this creative idea.

Page 4: Lamp3

CROWDING AND PARENTAL LANGUAGE 1023

Table 3

Regression of Language Diversity Onto Density and Parental Responsiveness

Predictor

Socioeconomic statusDensityResponsivenessDensity with additional control

for responsiveness

TotalR2

.46

.53

.63

.65

F(total R2)

34.65**22.36**34.67**

24.23**

AR2

.46

.07

.17

.02

F(AR2)

34.65**5.86*

18.72**

2.07

df

1,402, 392,39

3, 38

b

1.47**-77.49*215.49**

-42.33

SE of b

0.2531.5549.80

29.42

*p<.05. **p<.001.

children. As in prior studies, we found that parents in morecrowded homes are less responsive to their children. An importantcorrelate of this relative unresponsiveness appears to be less so-phisticated parental speech. Parents in more crowded homes useless diverse language when speaking to their children. This patternof findings may provide insight into why delays in cognitivedevelopment have been observed in several studies among chil-dren from more crowded homes. Although we can rule out SES,parental education, and the sheer amount of parental time andeffort allocated per child as alternative explanations for the rela-tionships we have uncovered, prospective, longitudinal data arenecessary to increase our confidence that the proposed model isboth robust and not confounded by one or more, as yet, unspecifiedvariables.

References

Aiello, J. R., Nicosia, G., & Thompson, D. (1979). Physiological, social,and behavioral consequences of crowding on children and adolescents.Child Development, 50, 195-202.

Aiello, J. R., Thompson, D., & Baum, A. (1985). Children, crowding andcontrol: Effects of environmental stress on social behavior. In J. F.Wohlwill & W. van Vliet (Eds.), Habitats for children (pp. 97-124).Hillsdale, NJ: Erlbaum.

Baum, A., Gatchel, R., Aiello, J., & Thompson, D. (1981). Cognitivemediation of environmental stress. In J. Harvey (Ed.), Cognition, socialbehavior, and the environment (pp. 513-533). Hillsdale, NJ: Erlbaum.

Baum, A., & Paulus, P. B. (1987). Crowding. In D. Stokols & I. Altman(Eds.), Handbook of environmental psychology (pp. 533-570). NewYork: Wiley.

Baum, A., & Valins, S. (1979). Architectural mediation of residentialdensity and control: Crowding and the regulation of social contact. In L.Berkowitz (Ed.), Advances in experimental social psychology (Vol. 12,pp. 131-175). New York: Academic Press.

Booth, A., & Johnson, D. (1975). The effects of crowding on child healthand development. American Behavioral Scientist, 18, 736-749.

Bradley, R., & Caldwell, B. (1984). The HOME inventory and familydemographics. Developmental Psychology, 20, 315-320.

Bronfenbrenner, U., & Morris, P. (1998). The ecology of developmentalprocesses. In W. Damon (Ed.), Handbook of child development (5th ed.,pp. 993-1028). New York: Wiley.

Evans, G. W., & Lepore, S. J. (1993). Household crowding and socialsupport: A quasiexperimental analysis. Journal of Personality and So-cial Psychology, 65, 308-316.

Evans, G. W., Lepore, S. J., Shejwal, B. R., & Palsane, M. N. (1998).Chronic residential crowding and psychological health: An ecologicalperspective. Child Development, 69, 1514-1523.

Evans, G. W., Palsane, M. N., Lepore, S. J., & Martin, J. (1989). Resi-dential density and psychological health: The mediating effects of socialsupport. Journal of Personality and Social Psychology, 57, 994-999.

Gottfried, A. W., & Gottfried, A. E. (1984). Home environment andcognitive development in young children of middle socioeconomicstatus families. In A. W. Gottfried (Ed.), Home environment and earlycognitive development (pp. 57-115). New York: Academic Press.

Hart, B., & Risley, T. R. (1995). Meaningful differences. Baltimore:Brookes.

Lepore, S. J., Evans, G. W., & Schneider, M. (1991). Dynamic role ofsocial support in the link between chronic stress and psychologicaldistress. Journal of Personality and Social Psychology, 61, 899-909.

Matheny, A., Wachs, T. D., Ludwig, J., & Phillips, K. (1995). Bringingorder out of chaos: Psychometric characteristics of the confusion, hub-bub, and order scale. Journal of Applied Developmental Psychology, 16,429-444.

Murray, R. (1974). The influence of crowding on children's behavior. In D.Canter & T. Lee (Eds.), Psychology and the built environment (pp.112-117). New York: Wiley.

Saegert, S. (1982). Environment and children's mental health: Residentialdensity and low income children. In A. Baum & J. E. Singer (Eds.),Handbook of psychology and health (pp. 247-271). Hillsdale, NJ: Erl-baum.

Stevens, G., & Cho, J. (1985). Socioeconomic indexes and the 1980 censusoccupational classification scheme. Social Science Research, 14, 142-168.

Wachs, T. D. (1989). The nature of the physical microenvironment: Anexpanded classification system. Merrill-Palmer Quarterly, 35, 399-419.

Wachs, T. D., & Camli, O. (1991). Do ecological or individual character-istics mediate the influence of the physical environment upon maternalbehavior? Developmental Psychology, 11, 249-264.

Wachs, T. D., & Gruen, G. (1982). Early experience and human develop-ment. New York: Plenum.

Wedge, P., & Petzing, J. (1970). Housing for children. Housing Review, 19,165-166.

Received May 7, 1998Revision received December 7, 1998

Accepted December 8, 1998