Differential Effects of Type of Children’s Contact with Their Jailed Mothers and Children’s Behavior Problems

M (SD)

1. Age

–

118.05 (20.00)

2. Gender^a

−.09

–

0.46 (0.50)

3. Ethnicity^b

.06

−.02

–

0.59 (0.49)

4. SES^c

−.04

.07

.13

–

4.18 (0.85)

5. Letter-M^d

−.06

.15

−.21*

.17

–

5.57 (3.79)

6. Letter-C^e

−.11

.11

−.11

.14

.39^**

^–

5.88 (3.68)

7. Phone-M

.06

.11

−.07

.12

.31^**

.24^**

^–

7.41 (3.51)

8. Phone-C

.10

.16

.04

−.04

.12

.41^**

.68^**

^–

7.09 (3.84)

9. Visits-M

−.10

.09

−.14

−.03

.08

.17

.27**

.16

–

3.20 (3.63)

10. Visits-C

−.02

.20*

−.07

.10

−.08

.11

.27**

.30**

.74**

–

4.49 (3.93)

11. Visits-CH^f

.18*

.011

−.02

.15

.08

−.10

.20*

.17

.28**

.27**

–

0.07 (0.26)

12. CBCL-IntMg

.13

.09

−.10

.04

.05

.03

−.06

.32**

.17

.07

–

6.91 (6.47)

13. CBCL-IntC

.11

.06

−.22**

.04

−.08

−.11

−.18*

.09

−.03

.06

.37**

–

7.37 (7.15)

14. CDI^h

−.14

.05

−.10

.02

−.10

.06

−.04

.02

.11

.04

.07

.23**

.16

–

8.04 (7.07)

15. MASCⁱ

−.13

.23**

−.04

.08

−.18*

.09

−.05

.09

−.04

.05

−.01

.17

.14

.40**

–

50.43 (17.97)

16. CBCL-ExtMj

−.02

−.06

−.18*

.07

−.02

−.06

.06

−.05

.12

.03

.11

.62**

.23**

.28**

−.02

–

9.91 (9.44)

17. CBCL-ExtC

.06

−.08

−.23**

.10

.05

−.01

.03

−.07

.05

−.10

.09

.29**

.59**

.22**

.07

.52**

–

11.46 (10.44)

18. RBP^k

.17*

−.05

−.08

−.12

−.10

−.03

.02

.12

−.01

−.04

.12

.20*

.23**

.19*

.04

.28**

.27**

–

4.35 (4.10)

19. Bullying^l

.06

−.05

−.01

−.28**

−.07

−.01

.12

.18*

−.01

−.00

−.01

.14

.13

.28**

.03

.25**

.15

.54**

6.21 (2.84)

Note: *p < .05, **p < .01

^aChild gender was coded as 1 = female, 0 = male

^bChild ethnicity was coded as 1 = Black, 0 = Not Black

^c SES was calculated based on mother and caregiver report of occupation and education via the Hollingshead method

^d M = mother reported variable

^eC = caregiver reported variable

^fCH = child-reported variable

^gCBCL-Int = Child Behavior Checklist-Internalizing Behavior Subscale

^hCDI = child depression inventory

ⁱMASC = multidimensional anxiety scale for children

^jCBCL-Ext = Child Behavior Checklist-Externalizing Behavior Subscale

^kRBP = risky behavior protocol

^lBullying = kids in my class at school

Mothers’ and caregivers’ reports of contact between the child and the incarcerated mother are presented in Table 2.2. Although mothers’ and caregivers’ reports were significantly, positively correlated (rs range from .39–.74, ps < .01), paired samples t-tests indicated significant differences between mothers’ and caregivers’ reports of phone contact and visitation. Specifically, mothers reported more phone contact than did caregivers, whereas caregivers reported more visits than did mothers.

Table 2.2

Mother and caregiver reports of child’s contact with the incarcerated mother

Reporter	Mothers	Caregivers	Correlation	t-Value (df)
Mode of contact
Frequency of phone contact	7.41 (3.51)	7.09 (3.84)	.68**	2.38 (120)*
Frequency of mail contact	5.57 (3.76)	5.88 (3.68)	.39**	1.00 (119)
Frequency of visitation	3.20 (3.62)	4.49 (3.93)	.74**	4.84 (123)**

Note: *p < .05, **p < .01

2.3 Primary Analyses

Given that we obtained child, mother, and caregiver report of most variables, we used structural equation modeling (SEM) to disentangle structural effects among latent constructs from correlations resulting from reporter-specific method variance. Analyses were conducted using Amos version 20 (Amos Development Corp., Crawfordville, FL). Covariance matrices were analyzed using full-information maximum likelihood estimation (FIML³). To test our hypotheses, Visits and Alternative Forms of Contact (letters and telephone calls) (as well as lack of contact, see Hypothesis 2) were modeled as simultaneous predictors of internalizing and externalizing behaviors. Residual error terms for observed variables reported by the same reporter (e.g., e1, e3, e5, e9, and e13 in the case of mother-reported variables) were specified to be intercorrelated (these paths are omitted from Fig. 2.1 for clarity of presentation). Child age, gender, ethnicity, and SES were controlled by specifying these variables as exogenous predictors of both outcome variables.