Educational risks for children experiencing homelessness

Educational risks for children experiencing homelessness

Pergamon Journal of School Psychology,Vol. 35, No. 1, pp. 27-46, 1997 Copyright © 1997Societyfor the Studyof School Psychology Printed in the USA 002...

1MB Sizes 12 Downloads 91 Views

Pergamon

Journal of School Psychology,Vol. 35, No. 1, pp. 27-46, 1997 Copyright © 1997Societyfor the Studyof School Psychology Printed in the USA 0022-4405/97 $17.00 + .00

PII S0022-4405(96)00032-5

Educational Risks for Children Experiencing Homelessness Ann S. Masten, Arturo Sesma, Jr., Rekhet Si-Asar, Catherine Lawrence, Donna Miliotis, and Jacqueline A. Dionne University of Minnesota Educational risks were investigated among 73 children 6 to 11 years old from homeless families staying in a Minneapolis shelter. Families were recruited at the shelter and followed up after they had moved into their own housing. Access to school was not a problem. However, significant school success problems, defined in terms of achievement and classroom behavior, were found among the 59 African American children who were the primary focus of analysis.Moreover, academic and behavioral problems often co-occurred, as did good achievement and good behavior. Results support the feasibility of research with highly mobile families whose children have extremely high risk for educational problems. Implications are discussed for researchers and educators who share the goal of fostering school success among high-risk mobile children. © 1997 Society for the Study of School Psychology Keywords: Homeless, Children, Education, Risk, Resilience.

EDUCATIONAL RISKS FOR CHILDREN EXPERIENCING HOMELESSNESS A c a t a s t r o p h e i n slow m o t i o n is u n f o l d i n g i n the U n i t e d States as the r a n k s o f A m e r i c a n s living in poverty swell with c h i l d r e n a n d y o u n g families ( S h i n n & Gillespie, 1994; W. T. G r a n t F o u n d a t i o n , 1988). T h e costs o f this crisis for d e v e l o p m e n t is d r a m a t i c a l l y illustrated by the p l i g h t o f h o m e l e s s c h i l d r e n . H o m e l e s s n e s s a m o n g families with c h i l d r e n i n c r e a s e d i n the 1980s ( I n s t i t u t e o f M e d i c i n e , 1988) a n d has c o n t i n u e d to i n c r e a s e i n m a n y cities, b o t h i n a b s o l u t e n u m b e r s a n d as a p r o p o r t i o n o f w h o is h o m e l e s s (U.S. C o n f e r e n c e o f Mayors, 1994). Schools are faced with g r o w i n g n u m bers o f very p o o r m o b i l e c h i l d r e n at c o n s i d e r a b l e risk for e d u c a t i o n a l failure who have b e e n o r will b e homeless. T h e p u r p o s e o f this study was to e x a m i n e the e d u c a t i o n a l risks associated with h o m e l e s s n e s s a m o n g e l e m e n t a r y school-aged c h i l d r e n i n a midweste r n city, as p a r t o f a n o n g o i n g effort to identify a n d address t h e n e e d s o f Address correspondence and reprint requests to Ann S. Masten, Institute of Child Development, University of Minnesota, 51 East River Road, Minneapolis, MN 55116. Fax: 612-6246373. Email: [email protected] 27

28

Journal of School Psychology

mobile children. There is good reason to believe that such children would be at grave risk of becoming academic casualties. These children often have multiple risk factors long associated with academic and behavioral problems, including chronic poverty, single-parent families, unemployed parents dependent on welfare, poorly educated parents, poor health and health care, elevated lead levels, mobility with attendant school changes, and histories of high exposure to stressful life events, including family and neighborhood violence (Masten, 1992; McChesney, 1993; Rafferty & Shinn, 1991; Shinn & Weitzman, in press). Children who have been homeless share many of these risk factors with millions of other impoverished children in the United States. They also have had unique problems related to educational access, such as residency requirements, and the social stigma attached to living in a shelter. During the 1980s, national concern about homelessness prompted Congress to enact the Stewart B. McKinney Homeless Assistance Act of 1987 (EL. 100-77, July 22, 1987). Among its provisions, Congress required that "each State educational agency shall assure that each child of a homeless individual and each homeless youth have access to a free, appropriate public education .... " (42 U.S.C. Sec. 11431). Amendments to the original McKinney Act in 1990 called for efforts to facilitate success in school for these children as well as access to school (Helm, 1993; Stronge, 1993). Though barriers remain, this law and associated funds to assist the States in the education of children who are homeless improved awareness of the problem and access to education (Helm, 1993; Stronge, 1993). However, addressing the educational needs of these children presents a far greater challenge. If schools, for example, are expected to provide an "appropriate" education, teachers and policy-makers will need to know a great deal more about the nature of the academic problems these children may have and the most promising means of addressing them. Studies of children challenged by homelessness over the past decade have documented significant educational issues. The Bank Street College of Education (Molnar, 1988) and Advocates for Children (Rafferty & Rollins, 1989; Rafferty, 1991) were pioneering organizations in alerting educators, the public, and policy planners to the educational needs of children homeless in New York City. Their data documented the barriers to education often faced by these children, as well as attendance problems and academic delays. The Stanford studies by the Stanford Center for the Study of Families, Children and Youth (1991) also suggested substantial educational barriers and risks, including problems with transportation and attendance, among children and adolescents from homeless families in northern California. A school-based study in Madison, Wisconsin, found both low academic achievement and high levels of behavior problems as rated by classroom teachers (Ziesemer, Marcoux, & Marwell, 1994). In one of the few studies available utilizing a standardized individual achievement

Masten et al.

29

test, Rescorla, Parker, a n d Stolley (1991) f o u n d reading scores on the Wide Range A c h i e v e m e n t Test for a sample of 43 African American children ages 6 to 12 were m o r e than o n e full standard deviation below the normative mean. T h e scores for these children, who were tested at a Philadelphia shelter, were lower b u t not significantly lower than an i n n e r city comparison g r o u p of African American children, a pattern that has b e e n f o u n d in a n u m b e r of studies c o m p a r i n g children who are homeless and o t h e r children living in poverty on a variety of outcomes. O u r previous studies of children f r o m families homeless in Minneapolis also have indicated serious academic problems. Although o u r initial study of 159 children living in the largest shelter in the region was not focused on educational issues, significant p r o b l e m s were evident (Masten, Miliotis, G r a h a m - B e r m a n n , Ramirez, & N e e m a n n , 1993). Parents a n d children alike r e p o r t e d missing substantial a m o u n t s of school and 38% of these 8to 17-year-old children had already repeated a grade. It was also clear that education was a great c o n c e r n of parents a n d children. Ninety-two p e r c e n t of the children rated school as "very i m p o r t a n t " to them, while parents, when asked in an o p e n - e n d e d question what their children n e e d e d the most, indicated that a g o o d education was the most i m p o r t a n t n e e d after basic survival needs o f housing, food, and clothing have b e e n met. These results, coupled with state and local interest in identifying a n d addressing the educational needs of children c o n t e n d i n g with homelessness led us to begin a series of studies focused on u n d e r s t a n d i n g a n d ultimately facilitating school success a m o n g this g r o u p of high-risk, mobile children. Assessment of children f r o m families that do not have their own h o m e presents a n u m b e r of challenges due to their mobility, lack of transportation, crowded and noisy conditions at shelters, and erratic attendance at school. In addition, very few measures have established validity or proven practical utility with such disadvantaged children. Children who are h o m e less a n d o t h e r highly disadvantaged or mobile children are likely to be u n d e r r e p r e s e n t e d in n o r m i n g samples. Given these difficulties, it was i m p o r t a n t to establish feasible and valid m e t h o d s for research with these children at the same time that we sought to learn m o r e a b o u t their educational needs. We c o n d u c t e d two pilot studies of 8- to 12-year-old children living in shelters for families, utilizing the Peabody Individual Achievement TestRevised (PIAT-R, Markwardt, 1989). In both cases, results indicated scores significantly below average for this test. We also f o u n d that the full PIAT-R required m o r e time than m a n y of these children could focus their attention effectively. Therefore, for the present study, we chose a shorter standardized test that had just b e c o m e available.

30

Journal of School Psychology OVERVIEW OF THE PRESENT STUDY

T h e data r e p o r t e d in this paper were part of a feasibility study to follow families as they moved out of a shelter. A representative sample of 73 children (60 African American, 11 American Indian, 2 E u r o p e a n American) ages 6 to 11 took part, and follow-up information was obtained for 69 children. This r e p o r t focuses primarily on the African American subsample and secondarily on the American Indian children, as there appeared to be distinct differences between these groups. Multiple methods were used to gather information p e r t i n e n t to school success, including individual tests, school records, teacher and parent ratings, and parent interviews. We were able to test 52 of the 73 children after they had left the shelter with the Wechsler Individual Achievement Test Screener ( WIAT-S; Psychological Corporation, 1992). This study addressed two main questions. First, do children who have recently lived in a shelter have substantial academic delays as assessed from multiple perspectives (standardized test scores, teacher ratings, cumulative school records)? Based on o u r pilot studies and other research (Rescorla et al., 1991; Zima, Wells, & Freeman, 1994), we expected that these children would be significantly b e h i n d by age and by grade c o m p a r e d to norms on a standardized achievement test, and that teacher ratings and school records would also reflect a cumulative history o f academic difficulties. Second, is academic achievement related to behavior problems and adaptive functioning in the classroom? Virtually no information was available about the covariance of achievement and behavior problems a m o n g children who were or had been homeless. However, there was good reason to expect classroom behavior problems as well as achievement problems in the same children. Both types o f problems have been r e p o r t e d separately for samples of children living in shelters (Masten, 1992; Masten et al. 1993; Rafferty & Shinn, 1991; Zima et al. 1994; Ziesemer et al., 1994). Moreover, both types of problems are likely to be influenced by many of the same risk processes that may occur in other very disadvantaged children, and one type of problem also may contribute to the other (Masten & Coatsworth, 1995; McLoyd, 1990). We expected teachers to r e p o r t problems in externalizing and internalizing symptoms as well as adaptive functioning and academic p e r f o r m a n c e on the Teacher's Report Form. We also expected independently assessed achievement test scores to be significantly related to teacher ratings of behavior in the classroom, such that children with problems would have worse achievement. Concomitantly, we expected that children who had satisfactory achievement test p e r f o r m a n c e (grade appropriate) would have fewer behavior problems than their low-achieving peers. In other words, we expected a small n u m b e r o f children to manifest educational resilience (Masten, 1994).

Masten et al.

31

METHOD Participants Families were initially recruited in the Winter of 1993 while they were living in a shelter for families in Minneapolis that houses by far the largest numbers of children of any shelter in the region and is, to our knowledge, the largest facility of its kind in the u p p e r Midwest. Previous studies had indicated that families in this shelter were typical of families who are homeless and living in shelters in the Twin Cities metropolitan area (Mueller & Friedrich, 1990; Masten et al., 1993). During a 3-month period, parents with children 6 to 11 years of age were invited to participate, using multiple recruitment strategies that have proven to be successful in the past and had p r o d u c e d representative samples. These included signs, written invitations in mailboxes, and word-of-mouth a m o n g families. Families constantly come and go in this shelter, where the average stay is only about 3 weeks. Therefore, during the first 2 weeks, families with children the appropriate age were randomly selected for invitation. The next month, every family with children in the appropriate age range were invited to participate. Since there were many more African American families and the age of children in shelters often is skewed (reflecting the greater vulnerability of young families to homelessness), these children were randomly recruited until three age blocks (20 per group) were filled. In the final weeks, recruitment focused on 10-11 year olds. The resulting sample included 73 children, 37 boys and 36 girls. Sixty were African American, 11 were American Indian, and 2 were European American. Given the small n u m b e r of White families in the sample, data analysis focused on the larger subsamples of minority children, emphasizing the largest group, comprised o f African American children. During the middle m o n t h of full sampling, 96% of eligible African American families participated. T h o u g h small, the American Indian group appeared to represent a distinct subgroup of families in shelter, and their participation rate was also excellent, with over 90% of eligible Indian families who stayed in the shelter during the entire recruitment period participating. Demographic characteristics o f the African American and American Indian families are provided in Table 1. American Indian families had higher rates of two-parent households and higher education levels than the other group, though power was insufficient to test these differences with confidence (even so, chi-square p-values were close to .05). These apparent differences could be r a n d o m variation. However, given the fact that these two characteristics are widely reported correlates of academic achievement and adjustment (Masten & Coatsworth, 1995; Garmezy & Masten, 1994) and the possibility of unknown cultural influences on homelessness pathways, risks and test performance, we analyzed the data for these two groups separately.

32

Journal of School Psychology

Table 1 Demographic Characteristics of African American and American Indian Homeless Children

Children Average Age Boys Girls Parent Average Age Married Single-headedhousehold Average number of children in the shelter Average total number of children Average income during the previous month Range of income Education Some higher education High school degree or GED Currenflyunemployed AFDCrecipient Homeless before

African American (n = 59)

American Indian (n = 11)

8.4 30 29

8.5 4 7

30.5 13 (22%) 46 (78%) 3.4 3.9 $ 579 $0-1500

31 6 (55%) 5 (45%) 2.5 3.5 $ 463 $0-961

18 19 57 49 28

(31%) (32%) (97%) (83%) (48%)

6 4 9 7 8

(55%) (36%) (82%) (64%) (73%)

T h e profile of risk indicated a m o n g the African American families is typical of families living in shelters in u r b a n areas a n d very similar to d e m o g r a p h i c characteristics we a n d others have f o u n d previously in this m e t r o p o l i t a n area (Masten et al., 1993; Mueller & Friedrich, 1990). T h e majority of households have o n e p a r e n t who is d e p e n d e n t on public assistance for their p r i m a r y source of income. Educational a t t a i n m e n t is often low. Family incomes are well below the poverty level andjoblessness is very high, although most parents have held j o b s at s o m e t i m e in the past. T h e reasons given for homelessness are p r e d o m i n a n t l y economic, with relationship p r o b l e m s also p r o m i n e n t . Most families have b e e n homeless for a short period of time before c o m i n g to the shelter, although 48% of the African American parents a n d 72% of the American Indian mothers r e p o r t e d they had b e e n homeless before. Approximately 2 m o n t h s after initial r e c r u i t m e n t and testing at the shelter, families a n d schools were contacted for follow-up assessments. At least partial information, including school records, teacher ratings, p a r e n t interviews, or child assessments, was obtained for 69 o f the original 73 families. O n e African American child was excluded f r o m analysis due to invalid test data a n d a r e p o r t e d developmental disorder (Autism). T h e mobility of these families, the difficulties in contacting them, and the school absences of their children resulted in attrition and additional missing data. For example, key analyses described below focus o n 40 of the

Masten et al.

33

59 African American and 10 of the 11 American Indian children with valid a c h i e v e m e n t test scores. School records were obtained for 53 of the African American a n d 9 of the American Indian children. T h e subsample of African American children with valid achievement test data at follow-up did not differ significantly f r o m the rest o f their subgroup on p a r e n t education level, p a r e n t reports of the target child's learning p r o b l e m s and academic p e r f o r m a n c e , or the child's test scores on two intellectual aptitude tests, all of which were assessed at the time of r e c r u i t m e n t in the shelter. By parent's r e p o r t in the shelter, the six children for w h o m we were unable to locate school records had significantly worse school p e r f o r m a n c e (t = 2.29, df= 57, p < .05) than the other 53 children in this g r o u p but did not differ on the o t h e r variables indicated above.

Measures Individual achievement test. T h e Wechsler Individual Achievement Test Screener (WIAT- S', Psychological Corporation, 1992) was used to estimate academic achievement. A short f o r m of the full WIAT test, the Screener provides n o r m e d a c h i e v e m e n t scores for Basic Reading, Mathematical Reasoning, a n d Spelling, as well as a composite score. Like the full WIAT, it was c o - n o r m e d with the Wechsler Intelligence Scale for Children - Third Edition. Psychometric data provided in the manual suggest that the Screener has excellent reliability a n d validity as a quick index o f achievement. T h e m a n u a l indicates that administration takes a b o u t 10 to 18 minutes, but children in this study usually required 20 to 25 minutes for this test.

Achievement ratings based on school records. Cumulative school records were requested for all children (with parental permission). To condense the information f r o m a diverse set of records, two judges blind to o t h e r information a b o u t each child (including ethnicity) independently rated these records on a 5-point scale: 1 = well below average, 2 = below average, 3 = average, 4 = above average, 5 = well above average. Judges first rated i n f o r m a t i o n based only on the final grading period for the current school year subsequent to leaving the shelter (Spring) and then r e n d e r e d a cumulative j u d g m e n t based on the entire record, giving m o r e weight to p e r f o r m a n c e during the c u r r e n t school year. Reliability was d e t e r m i n e d for i n d e p e n d e n t j u d g m e n t s a n d then disagreements were resolved for a consensus j u d g m e n t . Judges had rated records f r o m a previous study with high reliability. In this study, for 55 school records of African American and American Indian children with Spring data, judges agreed exactly on Spring ratings for 80%, within one level for 100%, a n d the intraclass correlation was .86. For 60 records with sufficient information for cumu-

34

Journal of School Psychology

lative ratings, exact a g r e e m e n t was obtained for 75%, within one level for 100%, and the intraclass correlation was .79. For cumulative ratings, the means for the African American and American Indian samples were 2.04 (SD = .71) and 2.67 (SD = .71), respectively. Ratings were average or above for 17% o f the African American children and 56% o f the Indian children.

Teacher ratings of achievement. Teachers completed the Teacher's Report Form (Achenbach, 1991), described below, which includes a standardized scale o f Academic Performance. They also completed a Brief Teacher Rating Form developed as a pilot instrument for ascertaining teacher impressions o f how children were doing in the classroom. Two items pertained to p e r f o r m a n c e , a rating of "academic potential" and a rating of "academic progress" over the past month. Each was rated on a 5-point scale from well below average to well above average. T h e exact wording was as follows: "Based on your observations, how would you rate the academic potential of this child?" and " H o w much academic progress has this child made over the past m o n t h ? "

Parent reports of achievement, family history and school access. Parents at the shelter were interviewed about family history and the target child. This interview included structured questions about school achievement and learning problems which were read to parents. " H o w well does [Name] usually do in school?" had five structured choices: from "excellent, mostly As" to "failing, mostly F's." Parents also were asked, "Does [Name] have any learning problems in school?" and yes or no was recorded. Additionally, parents were asked whether they felt welcome at their child's school and if they had e n c o u n t e r e d any of a list of potential barriers to school access. Barriers included lack of a p e r m a n e n t address, immunization records, school records, transportation, and a birth certificate. Intellectual functioning. Two brief tests were chosen to estimate intellectual functioning during the shelter stay, the Raven Coloured Progressive Matrices (Raven, Raven, & Court, 1991) and the Vocabulary subtest o f the WISC-III (Wechsler, 1991). T h e Raven has been used internationally as an index o f general nonverbal reasoning ability and may be somewhat less culturally influenced than o t h e r tests o f intellectual functioning (Sattler, 1992). Raw scores were converted to percentile ranks by 1986 U.S. norms. T h e means for the 59 African American children and 10 American Indian children with valid scores were 34th and 55th percentiles, respectively. Verbal ability was estimated with the Vocabulary subtest of the WISC-III, which is viewed as the single best measure o f "g" (general mental ability) on the WISCand also is relatively robust with respect to emotional distress (Satfler, 1992). Standardized scores on this scale averaged 6.27 (SD = 2.60)

Masten et al.

35

for the African American sample o f 59 children and 6.00 (1.63) for 10 Indian children.

Teacher ratings of adjustment.

Classroom teachers completed the Teacher's Report Form of the Child Behavior Checklist (Achenbach, 1991), a widely used measure of behavioral and emotional problems and general classr o o m adaptation with good reliability and validity. It provided four broad indices of adjustment as well as the rating of Academic Performance: Total Problems, Internalizing and Externalizing problems, and Adaptive Functioning. T h e last is a four-item composite pertaining to working hard, behaving appropriately, learning, and being happy. Standardized T scores (for age and gender) were c o m p u t e d by means of the computerized scoring program.

Procedures Families were recruited at the shelter as described above. Parents were invited to take part in the study and after it was described fully to them, informed consent was obtained to relocate families, to request school records and teacher ratings, and to invite children to participate. Parents completed several questionnaires and an interview o f 30 minutes to 1 h o u r and received an h o n o r a r i u m of $20 for their time. If families had more than one child eligible for the study, one child's name was randomly selected. T h e study was described to the target child whose assent was obtained. Children completed a "Kid Survey" set of questionnaires in the shelter that served as a warm-up task (not the focus of this study) and then were administered the Raven and WISC-IIIVocabulary tests. A gift valued at about $5 was given to children as an honorarium. One to 2 months after families had left the shelter, families and schools were contacted to obtain follow-up data. Tracking was complex, as some families had already moved twice and others had left the state. Schools were first contacted by mail, then followed-up as necessary by telephone to request teacher participation and school records. For schools in the metropolitan area, permission was obtained to visit and test the child at school whenever feasible. Due to repeated absenteeism, and the end of the school year, six children were assessed at h o m e and five children came to the University for follow-up testing. The assessment team consisted of five graduate students and one u n d e r g r a d u a t e honors student (4 of whom were minorities).

RESULTS Results are presented separately for the African American and American Indian subgroups o f this sample, as they were distinctively different in risk level as well as cultural background. Analyses focused on the larger sample

36

Journal of School Psychology Table 2 Descriptive Data for W/AT-S Subscale Scores

Co mp osite Age Grade Reading Age Grade Math R e a s o n i n g Age Grade

Spelling Age Grade

M

SD

t~

83.32 81.18

10.83 10.40

9.74"** 11.45"**

83.38 82.23

8.96 9.45

11.73"** 11.89"**

89.80 88.93

10.21 11.18

6.32*** 6.26***

83.85 83.03

10.11 10.63

10.10"** 10.10"**

N = 40. "Age" and "grade" refer to age- and grade-referenced normed scores on the W/AT-S. a One-sample /-test comparing homeless sample means to standardized test means of 100 (SD = 15) on the WIAT-S. ***p < .001.

of African American children. Data on a c h i e v e m e n t are presented first, followed by results on classroom behavior a n d the linkages f o u n d between a c h i e v e m e n t a n d classroom adjustment. Given expected restrictions in range a n d the small sample size, we faced considerable risk of Type II as well as Type I errors. Therefore, when testing an hypothesis required a single statistical test, we set alpha at .05. W h e n testing an hypothesis required multiple correlations or t-tests, alpha was set m o r e conservatively at .01. (Interpretations o f the findings are based on these criteria, however, information on multiple alpha levels are provided for the reader.) Statistical tests for the composite score on the WIAT-S a n d the Total Problem Score on the TRF were considered to be the " o m n i b u s " test for a given hypothesis as these scores were the most reliable and comprehensive.

African American Sample Descriptive data for the standardized individual achievement test are presented in Table 2. Scores were significantly below normative levels for age a n d for grade, with 80% o f the scores falling in the b o t t o m quartile. T h e r e were no significant differences in WIAT scores by sex or by age. Teacher ratings of Academic P e r f o r m a n c e on the TRFwere also significantly below the n o r m s for this scale (see Table 4). Twenty p e r c e n t of the children had r e p e a t e d a grade. Intercorrelations of achievement test scores, teacher reports o f academic p e r f o r m a n c e , the estimates o f intellectual functioning and p a r e n t reports

t~ "-4

.67*** .23 -.45"*

8. Teacher Rating: Academic Potential

9. Teacher Rating: Academic Progress

10. Parent Rating: Learning Problems

.17

.38*

.36*

.63***

.58***

.63***

3

.28

.60***

.59***

.60***

.50**

4

.35*

.54***

.67***

.83***

5

.26

.50**

.27

.19

.32*

.11

WIAT~

.24

.24

.37*

scores.

.12

.28

.03

-.40** -.39** -.44** -.35*

.21

.72***

.58***

.61"**

.58***

.77***

.61"**

2

N = 33-40. W1AT~ scores are grade-referenced. Table includes only those children with *p < .05. **p < .01. ***p< .001.

.28

.59***

7. TRF Academic Performance

13. Raven Score

.66***

6. Spring Record Rating

.45*

.60***

5. Cumulative Record Rating

12. WISC-III Vocabulary Score

.91"**

4. WIAT-S Spelling

.31"

.79***

11. Parent Rating: School Grades

.93***

3. WIAT-S Math

1. WIAT-S Composite

2. WIAT-S Reading

1

.46**

.66***

7

.11

.22

.15

.27

.14

.17

-.55*** -.36*

.38*

.63***

.71"**

6

Table 3 Intercorrelations of Achievement and Related Scores

.25

.31"

.38*

-.37*

.45**

8

.20

-.10

.08

--.39"

9

-.10

.00

.30

10

-.13

- .04

11

.42**

12

38

Journal of School Psychology

of learning p r o b l e m s are presented in Table 3. WIAT scores were strongly related to i n d e p e n d e n t a c h i e v e m e n t ratings based on school records as well as the c u r r e n t classroom teachers' j u d g m e n t s of Academic Perform a n c e on the TRF and their estimates of academic potential on a single 5-point rating scale. O n the Vocabulary subtest of the WISC-III (an index of verbal aptitude that was administered while the children were living at the shelter), scores also were moderately related to a c h i e v e m e n t scores, particularly for reading. Teacher reports a n d school records provided information a b o u t special education a n d C h a p t e r 1 services for m a n y of the participants. Eight of the 48 (17%) African American children for w h o m data were available were currently receiving special education and 21 (44%) were receiving Chapter 1 services. For b o t h types of services, children receiving t h e m h a d significantly lower achievement test scores than children not receiving these services. For special education services, there was m o r e than a full standard deviation difference between the two groups in WIAT Screener composite scores: M = 68.86 for 7 children receiving special education a n d M = 84.70 for the o t h e r 27 children for w h o m data on special education and WIAT test scores were available; t(32) = 4.17, p < .001. For four children with WIAT composite scores m o r e than two standard deviations below grade norms, three were receiving special education and the o t h e r C h a p t e r 1 services. Teacher reports o f behavior problems a n d adaptive functioning are s u m m a r i z e d in Table 4. Average scores on all m a j o r dimensions of behavioral symptoms were significantly elevated above the n o r m s for this test. Adaptive functioning scores were also significantly lower. Additionally, the p r o p o r t i o n s of children falling in the clinical range on these scores were significantly higher than expected values based on the n o r m i n g sample. T h e "clinical r a n g e " for the TRFnot only reflects deviance within the large samples of n o n r e f e r r e d children that provided the normative context for the test; the cutoff scores for the clinical range also reflect the scores f o u n d to discriminate best between n o n r e f e r r e d children and children referred for special education or mental health services (Achenbach, 1991). Intercorrelations of behavior as viewed by teachers and achievement scores are presented in Table 5. Classroom behavior p r o b l e m s were significantly related to a n u m b e r of i n d e p e n d e n t achievement indicators, particularly to reading a n d spelling scores, as well as to ratings of school records a n d the teacher's c u r r e n t perceptions of academic p e r f o r m a n c e . To further explore the relation of classroom behavior and achievement in terms of resilience, we c o m p a r e d children who h a d satisfactory versus low levels of a c h i e v e m e n t (defined by WIAT-S scores) on Total Problems as rated by teachers on the TRE. Children who scored above the b o t t o m quartile on the WIAT-S (by grade norms) were defined as academically successful for this analysis. O f the 35 children who had valid WIAT-S and

Masten et al.

39

Table 4 Descriptive Data for the Scores o n the Teacher Report Form o f the Child Behavior Checklist

Total Problems Boys Girls Internalizing Boys Girls Externalizing Boys Girls Adaptive F u n c t i o n i n g Boys Girls Academic Performance Boys Girls

M

SD

59.50 59.16 59.87 55.81 55.28 56.39 61.10 61.08 61.13 41.68 41.43 41.95 38.77 39.48 38.00

11.95 13.21 10.70 9.23 10.66 7.56 12.33 12.72 12.18 7.41 8.32 6.45 5.48 5.77 5.17

ta

Percent in Clinical Range

Chi-Square b

56%

52.41"**

38%

14.23"**

52 %

37.79***

48%

29.42***

73%

115.70"**

3.43"* 4.33*** 2.38** 3.86*** 4.24*** 4.11"** 4.88*** 5.72*** 9.29*** 11.04"**

N = 48 (25 boys, 23 girls), except Adaptive Functioning, N = 44 (23 boys, 21 girls). One-sample t-tests within gender, comparing homeless to normative sample means reported in Table 2-2 and Table 34 of the TRFmanual (Achenbach, 1991). b Chi-square tests comparing proportion of homeless in the clinical range to expected values provided in Table 6-3 of the TRFmanual (Achenbach, 1991). **p < .01. ***p < .001.

TRFscores, only 7 scored in this range on the achievement test, indicating adequate academic achievement. This academically successful g r o u p was c o m p a r e d to the low-achieving g r o u p on the TRFTotal Problem score. As illustrated in Figure 1, the successful g r o u p of seven had a T R F m e a n score (50.6) at the normative population m e a n (50) for this test, in striking contrast to the m e a n (62.6) for the low-achieving g r o u p (N= 28), which was well into the clinical range a n d m o r e than a standard deviation higher (t-test for m e a n difference = 2.54, df= 33, p --- .016).

American Indian Subsample A c h i e v e m e n t data were consistent with o u r tentative conclusion based on d e m o g r a p h i c i n f o r m a t i o n that this small American Indian g r o u p were relatively lower in risk. Only one child had r e p e a t e d a grade. N o n e received special education and three received C h a p t e r 1. O f the 10 children tested on the WIAT, seven scored at age a n d grade level. T h e correlation of cumulative school record ratings a n d WIAT composite scores was .75, significant (p < .05) even for a sample of eight. Teacher ratings o f behavior

40

Journal of School Psychology Table 5 Intercorrelations Between Measures of Achievement and Teacher Reported Behavior Problems

Teacher Report Form Scores Achievement Scores WIAT-S Scores Composite Basic Reading Math Reasoning Spelling Teacher Report Form Academic Performance School Records Spring Ratings Cumulative Ratings Brief Teacher Ratings Academic Potential Academic Progress

Total Problems Internalizing Externalizing

Adaptive Functioning

-.49"* -.51"* -.21 -.51"*

-.43"* -.41" - . 19 -.45"*

-.39" -.43** - . 16 -.42"

.43* .40* .26 .49*

-.51"**

-.48***

-.35*

.45*

-.46** -.50***

-.40** -.49***

-.34* -.39**

.55*** .51"**

-.43** -.27

-.33* -.08

-.35* -.28

.53*** .26

N = 31-48. WIAT-S scores refer to grade-referenced norms. *p < .05. **p < .01. ***p < 001.

were obtained for nine of these children and four (44%) had scores for Total, Externalizing, and Internalizing problems in the clinical range, with mean scores for all nine children falling at 53.7, 55.1 and 55.9, respectively. Adaptive functioning averaged 46.00 (two children in clinical range) and Academic Performance averaged 45.56 (three in clinical range). Parent Reports on School Access Parents were asked if they had encountered specific problems in getting their children to school. Very few of the 73 parents answered yes to this list of barriers. For example, no parents reported problems with immunization records or birth certificates; one reported a problem due to no permanent address; several reported problems related to transportation or school records. All of the parents reported feeling welcome at their child's school (64 parents answered). DISCUSSION Results of this study add to the growing evidence that children faced with homelessness often have serious educational problems that extend beyond the barriers to education that can derive from residential instability. From

Masten et al.

41

65

ss-Low WIAT-S (N=28)

O

60

[.,

ss-Higher WIAT-S (N=7) ve Sample

55 o

gh m



50

,4at

O

45 Figure 1. Means for Total Problem Score on the TRFfor low and high achievingchildren who have been homelesscompared to the TRFnormativesample. "Low" W/AT-Sscores are in the bottom quartile (by grade norms) and "higher" scores are above that level. multiple perspectives, many of the African American children recently homeless in this study were substantially behind in academic skills and their teachers often r e p o r t e d classroom adjustment problems. Early reports o f educational problems a m o n g children who were homeless suggested high levels o f grade retention, absenteeism, and lower scores on group-administered achievement tests (Rafferty & Shinn, 1991). Our findings with individual achievement testing, consistent with Rescorla et al.'s (1991) data, corroborate this picture of significant education problems. Data on the behavioral and emotional problems of children enduring homelessness have also consistently indicated high rates of significant problems (Masten et al., 1993; Rafferty & Shinn, 1991). These results have relied on parent and child reports. Our results suggest that teachers also perceive serious adjustment problems among a substantial p r o p o r t i o n of these children, consistent with findings from the Madison study (Ziesemer et al., 1994). O u r results also indicate that academic and psychological problems covary, even when assessed by i n d e p e n d e n t methods, such that a substantial n u m b e r of the children would be expected to have behavioral and academic problems. On the other hand, it is important to consider the implications of children who appeared to be holding their own academically despite extremely difficult life situations. T h e r e were children in this study who were doing reasonably well as indexed by scores on the standardized achievement test and they had few classroom behavior problems. T h e r e

42

Journal of School Psychology

were children from both the samples of African American and American Indian children who were doing well from the perspective of parents, records, and teachers. Concomitantly, it was clear that children participating in this study were not h o m o g e n e o u s in background, risks or resources. Understanding how these children succeed in school could guide policy and p r o g r a m m i n g designed to boost the adaptation of other high risk children. Even a m o n g these children with very high cumulative risk levels, there may be important variations in proximal environment and net risk. Some children may have more resources, particularly in the form of involved adults who support educational attainment even in the midst of family crises or chronic poverty. Further studies from this project are addressing the protective effects o f good parenting on the school success of children who have experienced homelessness. Results for the WISC-III verbal I Q score in the present study, indicating an association with academic achievement, are consistent with the possibility that higher achieving children have more cognitive resources. In addition to attachment relationships with caring and c o m p e t e n t adults, intellectual skills are probably the most widely r e p o r t e d protective factor in the resilience literature (Masten, 1994). Educational resilience a m o n g mobile, high risk children is a critically important p h e n o m e n o n that needs further examination (Wang & Gordon, 1994). O u r experience with varying strategies of assessment (e.g., in shelters and in schools, during and after homelessness, utilizing different measures) have convinced us that meaningful research is feasible with mobile, high-risk children. Both the PIAT-R and WIAT-S tests have shown encouraging correlational patterns suggesting validity as indices of academic achievement. The WIAT-S appeared to have considerable potential as a brief individual measure of achievement in these highly disadvantaged children. It had moderate to strong associations with academic history, parent and teacher perceptions o f p e r f o r m a n c e and learning problems, and a verbal test o f intellectual functioning. We also f o u n d that a simple teacher rating o f academic potential was a good index of past and concurrent academic performance. With the permission o f parents, teachers and school personnel were very helpful in providing school records or information, even across states by mail. Still, there are many challenges, particularly for longitudinal studies, in tracking families who move frequently and often do not have telephones or leave forwarding addresses. Yet these families believe education is very important and they are highly responsive to research focused on the educational needs of their children. Clearly more work is n e e d e d before any definitive conclusions would be appropriate. Our studies and the whole body o f available research are limited in scope. This study had a n u m b e r of limitations related to small sample size and the difficulties o f assessment and longitudinal data collection with this population.

Masten et al.

43

Nonetheless, the educational risks observed a m o n g children who are currently or have b e e n recently homeless are clear and urgent. While additional research is necessary, schools can use the knowledge that has accumulated to develop and evaluate programs to foster educational success in these children. Many such efforts have been implemented, although little evaluation appears to have been done (Eddowes, 1993; Powers & Jaklitsch, 1993; Yon, Mickelson, & Carlton-LaNey, 1993; Zeldin & Bogart, 1990). Evidence is m o u n t i n g that children who are or have been homeless have considerable school-related problems, even when barriers to educational access have been addressed. Further, there is little reason to believe that their problems are unique or largely the result of homelessness per se. Rather, homelessness appears to be a marker of very high cumulative educational risk levels likely to be shared by other children living in extreme poverty. Addressing the educational needs of these children is a p r o f o u n d challenge of great complexity. Schools cannot be expected to address the societal level economic and sociocuhural problems that may underlie poverty and related risks. However, there are multiple strategies that could be considered as a part of a "cumulative protection" approach for high risk students (Masten & Wright, in press; Wang & Gordon, 1994; Yoshikawa, 1994). Children experiencing homelessness have basic survival needs with respect to health care, food, shelter, and clothing, as well as academic and behavioral/emotional problems. They also may have peer problems related to appearance and social stigma. Urban homelessness disproportionately affects minority children, as does poverty more generally (Huston, McLoyd, & Garcia Coil, 1994). Hence, racism and prejudice may add additional burdens to the risks faced by these children. Programs to unobtrusively boost the nutrition, hygiene, and appearance of these children at school may n e e d to be considered along with programs to boost school stability, attendance, reading skills, perceived belonging, and homeschool connections. Teachers need to be prepared for the multifaceted needs of these children and the challenges of building rapport with parents who may have no telephone or transportation and who may be preoccupied with survival needs of their families. J o h n s o n (1992) has described a variety of educational support services that could address the multiple and diverse needs of children who are homeless. Stronge (1993) has suggested that a " c o n t i n u u m of educational services" may be n e e d e d to address a c o n t i n u u m of educational needs in these children. Children living in severe and chronic deprivation would have different needs than children temporarily homeless due to acute and short-lived situations. Schools have the potential to provide developmental havens of safety, stability, and care for children living in poverty whose lives are complicated

44

Journal of School Psychology

by homelessness or residential instability. Teachers, physical and mental health-care providers, nutritionists, administrators, policy-makers, and researchers all have roles to play in responding to the educational needs of children in families surviving homelessness. It will not be easy; fostering lasting achievement in disadvantaged children has proven to be a formidable task (Slavin, Karweit, & Madden, 1989; Wang & Gordon, 1994). However, perhaps we can learn how to facilitate educational resilience from the families who overcome adversity, guided by children like those in this study who were succeeding in school in the midst of all the difficulties surrounding homelessness. Action-oriented research is n e e d e d to identify, implement, and test protective processes in the school context, for homeless as well as for other educationally disadvantaged students.

ACKNOWLEDGMENTS We are deeply grateful to the families who made this study possible through their willingness to help others during a difficult time in their lives and to our collaborators at People Serving People, in the Minneapolis Public Schools, and at the Institute of Child Development. A special thanks to Carla Albers, Phyllis Elliott, J o r d a n Hart, Gerald Hickman, J o n Hubbard, Christine Hurley, Margo Hurrle, Bob Jibben, Jill Neremberg, Bob Pickarski, and Bill Price. This study has been financially supported by the Center for Research on Developmental Disabilities f u n d e d by the National Institute of Child Health and H u m a n Development (P30 HD24051) and the Center for Urban and Regional Affairs at the H u b e r t H. H u m p h r e y Institute, University o f Minnesota.

REFERENCES Achenbach, T. M. (1991). Manual for the Teacher'sReport Form and 1991 Profile. Burlington, VT: University of Vermont Department of Psychiatry. Eddowes, E. A. (1993). Education of younger homeless children in urban settings. Education and Urban Society, 25, 381-393. Garmezy, N., & Masten, A. S. (1994). Chronic adversities. In M. Rutter, E. Taylor, & L. Hersov (Eds.), Child and adolescent psychiatry: Modern approaches (pp. 191-208). London: Blackwell Scientific Publications. Helm, V. M. (1993). Legal Rights to education of homeless children and youth. Education and Urban Society, 25, 323-339. Huston, A. C., McLoyd, V. C., & Garcia Coll, C. (1994). Children and poverty: Issues in contemporary research. Child Development, 65, 275-282. Institute of Medicine, National Academy of Sciences. (1988). Homelessness, health and human needs. Washington DC: National Academy Press. Johnson,J. E,Jr. (1992). Educational support services for homeless children and youth. InJ. H. Stronge (Ed.), Educating homelesschildren and adolescents:Evaluating polisy and pract/ce (pp. 153-176). Newbury Park CA: Sage.

Masten et al.

45

Markwardt, E C. (1989). Peabody IndividualAchievement Test-RevisedManual, Circle Pines, MN: American Guidance Service. Masten, A. S. (1992). Homeless children in the United States: Mark of a nation at risk. Current Directions in Psychological Science, 1, 41-44. Masten, A. S. (1994). Resilience in individual development: Successful adaptation despite risk and adversity. In M. Wang & E. Gordon (Eds.), Risk and resilience in inner city Amehca: Challenges and prospects (pp. 3-25). Hillsdale, NJ: Erlbaum. Masten, A. S., & Coatsworth, J. D. (1995). Competence, resilience, and psychopathology. In D. Cicchetti & D.J. Cohen (Eds.), Developmentalpsychopathology. Vol. 2: Risk, disordeg, and adaptation (pp. 715-752). New York: John Wiley. Masten, A. S., Miliotis, D., Graham-Bermann, S. A., Ramirez, M., & Neemann,J. (1993). Children in homeless families: Risks to mental health and development. Journal of Consulting and Clinical Psychology, 61, 335-343. Masten, A. S., & Wright, M. O'D. (in press). Cumulative risk and protection models of child maltreatment. In B. B. R. Rossman & M. S. Rosenberg (Eds.), Multiple victimization of children: Conceptual, developmental, research and treatment issues. Binghamton, NY: Haworth Press. McChesney, K.Y. (1993). Homeless families since 1980. Education and Urban Society, 25, 361-380. McLoyd, V. C. (1990). The impact of economic hardship on Black families and children: Psychological distress, parenting, and socioeconomic development. Child Development, 61, 311-346. Molnar, J. (1988). Home is where the heart is: The crisis of homeless children and families in New York City. New York: Bank Street College of Education. Mueller, D., & Friedrich, J. (1990). Homeless families in the Twin Cities: Profile from two studies. WRC Report: Issues in Human Services Research and Policy, 1, 7-9. Powers, J. L., &Jaklitsch, B. (1993). Reaching the hard to reach: Educating homeless adolescents in urban settings. Education and Urban Society, 25, 394-409. Psychological Corporation. (1992). Wechsler Individual Achievement Test Screener. San Antonio: Author. Rafferty, Y. (1991). And miles to go . . . Barriers to academic achievement and innovative strategies for the delivery of educational services to homeless children. New York: Advocates for Children. Rafferty, Y., & Rollins, N. (1989). Learning in limbo: The educational depivation of homeless children. New York: Advocates for Children. Rafferty, Y., & Shinn, M. (1991). The impact of homelessness on children. American Psychologist, 11, 1170--1179. Raven, J., Raven, J. C., & Court, J. H. (1991). Manual for Raven's Progressive Matrices and Vocabulary Scales. Oxford: Oxford Psychologists Press. Rescorla, L., Parker, R., & Stolley, E (1991). Ability, achievement, and adjustment in homeless children. Ametcan Journal of Orthopsychiatry, 61, 210-220. Sattler, J. M. (1992). Assessment of children (3rd ed.). San Diego: Author. Shinn, M., & Gillespie, C. (1994). The roles of housing and poverty in the origins of homelessness. American Behavioral Scientist, 37, 505-521. Shinn, M., & Weitzman, B. C. (1996). Homeless families are different. InJ. Baumohl (Ed.), Homelessness in America: A reference book. (pp. 109-122). Phoenix: Oryx. Slavin, R. E., Karweit, N. L., & Madden, N. A. (Eds.). (1989). Effective programs for students at risk. Needham Heights, M_A:Allyn and Bacon. Stanford Center for the Study of Families, Children and Youth. (1991). The Stanford studies of homeless families, children and youth. Palo Alto: Author. Stronge, J. H. (1993). From access to success: Public policy for educating urban homeless students. Education and Urban Society, 25, 340-360.

46

Journal of School Psychology

U.S. Conference of Mayors. (1994). A status report on hunger and homelessness in America's dties. Washington DC: Author. W. T. Grant Foundation Commission on Work, Family and Citizenship. (1988). The forgotten half."Pathways to successfor America's youth and young families. Washington DC: Author. Wang, M. C., & Gordon, E. W. (1994). Educational resilience in inner-c#y America: Challenges and prospects. Hillsdale, NJ: Erlbaum. Wechsler, D. (1991). Wechsler Intelligence Scale for Children-Third Edition. San Antonio: Psychological Corporation. Yon, M. G., Mickelson, R. A., & Carlton-LaNey, I. (1993). A Child's Place: Developing interagency collaboration on behalf of homeless children. Education and Urban Society, 25, 410-423. Yoshikawa, H. (1994). Prevention as cumulative protection: Effects of early family support and education on chronic delinquency and its risks. Psychological Bulletin, 115, 28-54. Zeldin, S., & Bogart, J. (1990). Education and community support for homeless children and youth: Profiles of 15 innovative and promising approaches. Washington DC: U.S. Department of Education. Ziesemer, C., Marcoux, L., & Marwell, B. E. (1994). Homeless children: Are they different from other low-income children? Social Work, 39, 658--668. Zima, B. T., Wells, K. B., & Freeman, H. E. (1994). Sheltered homeless children in Los Angeles County. American Journal of Public Health, 84, 206--264.