Evaluation of the New York City Home Rebuilders Demonstration

IMPLEMENTATION OF THE HOMEREBUILDERS DEMONSTRATION

Chapter 3:
Section 3.3  Interviews with Caseworkers

Caseworkers were interviewed in both experimental and control groups in HD, LF, and MM, and in the HomeRebuilders groups in JC and NYF. Attempts were made to interview caseworkers for the same sample of families as that used for the caretaker interviews, except that it was not possible to interview caseworkers in OTT. As indicated above, the sample was limited to cases in which the goal at the beginning of the experiment was return home and in which there was at least one child under 13 years old.

Caseworkers were interviewed on a total of 407 cases. Some caseworkers were interviewed about more than one case. These cases were distributed among the agencies and experimental groups as indicated in Table 3-20.

Table 3-20.  Caseworker interviews by agency and experimental group

As in the caretaker interviews, caseworkers were asked a number of questions regarding child and caretaker functioning at the end of the experiment and about the services provided to the family. Again, we analyzed differences between experimental and comparison groups within agencies. Separate analyses were not conducted for birth parents, as in the caretaker interviews. Caseworker respondents varied considerably in their relationships to the families. Some were caseworkers who provided services during the experiment, some were successor workers, and others were supervisors. Sometimes respondents did not have personal knowledge of the family and were responding based on case records.

3.3.1  Outcomes

Family and Caretaker Functioning.  Three scales were constructed based on items in the caseworker interview, as follows:

• Household condition, proportion yes on 13 yes-no items on problems (caseworker interview, question 22a);
• Caretaker problems, proportion yes on 21 items (caseworker interview, question 24); and
• Caretaker functioning, average of 9 items, scaled from 0-4, not adequate to very adequate (caseworker interview, question 23).

Two of the nine within-agency comparisons on these scales were significant at .1; in both, the control group was doing better (HD: caretaker problems, LF: caretaker functioning). As with caretakers, caseworkers were asked for their overall assessments of improvement in the families. There were no significant differences within agencies in the responses to this question.

It would be desirable to have records of subsequent allegations of abuse or neglect from the administrative data. Unfortunately, such records were not available. As an alternative, workers were asked whether there had been subsequent allegations and whether they had been substantiated. Table 3-21 shows the results. As can be seen, there were relatively few substantiated allegations. Differences between experimental and comparison groups were small and not significant, due in part to small numbers of families.

Table 3-21.  Caseworker reports of allegations of maltreatment subsequent to the beginning of the experiment (percents of families)

Child Functioning.  Caseworkers were asked two series of questions about child functioning problems. The first asked about the children collectively, as part of a question on problems in the family (this question included the items on caretaker problems discussed above) (caseworker interview, question 24). There were 10 yes-no items in this question (e.g., physical or mental health problems, substance abuse, learning problems, behavior problems), and a scale consisting of the proportion "yes" was formed. There were no significant differences between experimental and control groups within agencies on this scale.

A second set of eight questions was asked about each child in the family (caseworker interview, question 25). These questions were:

• Is the child alert and interested in what is going on?
• Does the child appear to be small in size and light in weight for his/her age?
• Does child appear to be irritable and easily upset?
• Do you think the child's use of language and vocabulary is adequate for his/her age?
• Is the child aggressive toward the caretaker?
• Does the child have problems in school?
• In the last 3 months has the child been truant from school?
• In the last 3 months has the child run away from home?

Whether any child in the family had a problem in any of these eight areas was determined. A count of the number of problem areas for any child in the family was then formed. None of the mean differences between the experimental and control groups were significant on this scale.

Summary.  As with the caretaker interviews, the caseworker interviews provided little evidence for differences between experimental and control groups in child and family functioning outcomes.

3.3.2  Services

Two scales were constructed:

• Number of services provided out of a list of 26 services (caseworker interview, question 26); and
• Number of different kinds of caseworker activities, based on a list of 17 possible activities (caseworker interview, question 27, except r).

In LF, the experimental group had significantly higher means than the control group on these scales (for number of services, p < .01; for kinds of caseworker activities, p < .1). On kinds of caseworker activities, the other two true experimental agencies (HD and MM) had results in the same direction. On number of services, MM was in the same direction while in HD the control group had a higher number of mean services.

Clusters of similar services were also examined. Table 3-22 shows percentages of families provided services within each of these clusters. In LF, experimental group families received each of these clusters of services more often than control group families and the differences were significant for income support and health care or assessment. In HD, experimental group clients more often received drug or alcohol treatment (p < .1).

Table 3-22.  Services provided in certain service categories (percents)

It would have been useful to examine the "match" of services to problems. Unfortunately, matching is hampered by the fact that we do not have data on the problems of families at the outset of the experiment, only at the time of the interview with the caseworker. Workers were also asked how often they visited caretakers. In LF and HD experimental group workers reported more frequent visits (HD: p < .1; LF: p < .003).

The results indicate that in the perception of caseworkers, in LF, experimental group families received more service. There is little evidence for higher levels of service in the experimental groups of the other agencies.