Measuring parent engagement in foster care.
KEY WORDS: engagement; family-centered practice; foster care; parents; reunification
According to the most recent federal estimates, there are approximately 510,000 children in foster care in the United States (U.S. Department of Health and Human Services [HHS], 2008).The majority of these children have been removed from their caregivers (normally, parents) because of abuse or neglect, and parents must prove a certain level of competence to reclaim their children from state custody. In fiscal year 2006, approximately 49% of these children had a case goal of reunification (HHS, 2008). This means that in approximately half of foster care cases, child welfare workers were, to some extent, working with parents, children, foster parents, and service providers toward returning foster children to safe and stable home environments with the individuals from whom they were removed.
Today, child welfare agencies widely endorse a family-centered approach to foster care service delivery. This approach is based on the concept that an empowering, strengths-based parent-caseworker relationship is central to maintaining parents' engagement in services and the belief that continued engagement will propel parents toward success (that is, reunification; for example, Maluccio, 1981; Zamosky, Sparks, Hatt, & Sharman, 1993). However, despite the importance of engaging parents, parents of children in foster care are understudied. The current literature offers little in terms of insight into parents' experiences in the foster care process and even less in terms of quantitative analysis of parent engagement. The present study attempts to flU this gap in the research by testing a new measure of parent engagement.
Social work theorists and researchers note that family-centered practice involves "respect for parents, listening to and addressing their concerns, focusing on strengths, and helping them stay emotionally connected to their children" (Petr & Entriken, 1995, p. 528);breaking down the power differential between client and worker (Forrest, 2003); and communicating empathy and concern for all family members, while clarifying expectations and gently asserting authority (DePanfflis, 2000). Caseworkers must be willing to take into account parents' perspectives. They must respect and incorporate parents' knowledge about their own families and develop service plans, in partnership with parents, on the basis of parents' personal goals.
The approach manifests in frontline foster care casework in a variety of ways (for example, Dawson & Berry, 2002; Petras, Massat, & Essex, 2002). Family centered strategies include the following: intensive family reunification services (Child Welfare League of America, 2002; Fraser, Walton, Lewis, Pecora, & Walton, 1996; Gillespie, Byrne, & Workman, 1995; Hess, Folaron, & Jefferson, 1992; Walton, 1998); family-centered practice training for frontline staff (Alpert & Britner, 2005; Forest, 2003); and specific case management practices such as family group conferencing (Annie E. Casey Foundation, 2008; Corby, Millar, & Young, 1996; Waites, Macgowan, Pennell, Carlton-LaNey, & Weil, 2004).
Unfortunately, some barriers exist to family-centered service delivery. One obstacle is the incompatibility of the mandatory nature of foster care with the goal of parent empowerment. The foster care system is inherently punitive. The unwavering presence of state power, characterized by the threat of having their children permanently removed, pervades the foster care process for parents. It is in this environment that caseworkers are expected to communicate empathy and respect for parents, gain their trust, highlight their strengths, and improve their skills and resources.
Other impediments pertain to the practical translation of family-centered principles into daily casework. Foster care caseworkers have identified training-related obstacles (for example, being trained to consider the child as the primary client; being trained to prioritize child safety above all); logistical obstacles (for example, high caseloads, lack of time and resources); and obstacles based on workers' beliefs and expectations of parents (for example, worker bias against parents, workers' skepticism regarding the ability of parents to be rehabilitated) (Alpert & Britner, 2005).
Research on Parents
Qualitative Studies. Despite the fact that family-centered casework is intended to maximize parent engagement, little research exists investigating the extent to which parents feel in receipt of family-centered services. A small group of qualitative studies describes parents as feeling underserved and overlooked by child protection services. Parents in these studies reported hopelessness and rage toward the child welfare system (Haight et al., 2002), inconsistent caseworker responsiveness, feeling left out of decision making (Kapp & Propp, 2002), and feeling vulnerable to and fearful of their caseworkers (Diorio, 1992).
Dumbrill's (2006) study adds to the discourse by describing the relationship between parents' style of engagement and their perceptions of power within the parent-worker relationship. In this study, parents who perceived their workers as wielding power over them were more likely to directly oppose their workers or "play the game" with them (that is, feign cooperation simply to placate workers). In contrast, parents who perceived their workers as taking a "power with" approach were more likely to report a collaborative parent-worker relationship.
Finally, Corby et al. (1996) contributed to our understanding of parent engagement in their study of child protection case conferences. Although most parents in this sample reported being glad they attended conferences, more than half felt their views were not taken into account during these meetings. Nearly a third stated that they did not feel empowered to challenge professionals who made statements that the parents felt were inaccurate, suggesting that these parents were more concerned with appearing compliant than pursuing a more interactive collaboration with their workers.
Quantitative Studies. Some researchers have embarked on quantitative analyses of parents' experiences. For example, among its vast collection of variables, the National Survey of Child and Adolescent Well-Being (NSCAW) (Administration for Children and Families [ACF], 2005) asks parents to report their degree of satisfaction with their child welfare workers. In a recent report, the mean scale score for this group of questions emerged as a 4.59 on an eight-point scale, indicating a "middle level of perceived relationship quality" (ACF, 2005). In another study using NSCAW data, Chapman, Gibbons, Barth, McCrae, and the NSCAW Research Group (2003) found that the following variables showed a significant negative association with parents' perception of the quality of the relationship with their caseworkers: having two or more child welfare workers, longer time periods between visits, not being offered the kind of help needed, and feeling that more services should have been offered.
Using their Parent Satisfaction with Foster Care Services Scale (PSFCSS), Kapp and Vela (2004a, 2004b) found that parents were more likely to be satisfied overall "when they perceived that the social worker "(i) [was] working with them to get their child back, (ii) had clear expectations of them, (iii) prepared them for meetings, (iv) stood up for them in meetings, and (v) respected their cultural background" (Kapp &Vela, 2004b, p. 203). Harris, Poertner, and Joe (2000) have also attempted to measure parent satisfaction using their Parents with Children in Foster Care Satisfaction Scale (PCFCSS).These researchers reported a one-factor instrument with an internal consistency of [alpha] = .97. Though Harris et al. (2000) highlighted the need for instrument development and further experimentation using the PCFCSS, literature searches did not produce subsequent studies using the tool.
Attempts at measuring parent satisfaction with foster care services are admirable and represent important first steps in quantifying the experience of parents of children in care. However, the quantitative studies mentioned earlier raise an important question: Is satisfaction, per se, the most important construct to measure in an effort to get to the heart of parents' experiences in foster care? To be sure, we want parents to be satisfied with the services they receive, but in looking at a casework model in which the ultimate goal is reunification, do we believe that it is parent "satisfaction" that leads to success, drives a parent to persevere through his or her service plan, and captures the importance of collaboration and the transformative goals of the foster care experience? When viewed from this perspective, "satisfaction," alone, may fall short as a useful construct. The present study takes the position that the influential factor in success is something larger than satisfaction. Specifically, we reconceptualize the phenomenon of interest as "engagement" in the service delivery process, or more specifically, in the parent-worker relationship. Reframing the construct in this way improves our ability to capture parents' experiences as recipients of family-centered casework as it is described in the theoretical literature mentioned earlier. Moreover, it allows us to create an enhanced measurement tool, one that frames worker behavior not only as potentially satisfying, but also as potentially engaging.
THE PRESENT STUDY
The objectives of this cross-sectional, quantitative study were threefold: (1) to devise and test a measure of parent engagement; (2) to test whether demographic and case-related variables influence parent engagement, as measured by the new instrument; and (3) to obtain descriptive information about parent engagement to inform the participating agency about its clientele and enhance the existing foster care literature.
The research presented here is part of a larger study that took place in a nonprofit child welfare agency contracted by a large northeastern city to deliver foster care services. Institutional review board approval was obtained from four institutions: the sampled agency, the city's public child welfare agency, the state's public child welfare agency, and our affiliated university.
The agency in question serves approximately 400 children and their families in its foster care program, which is located at two sites in the city. Site A manages roughly 60% of the agency's foster care cases, and site B serves the remaining families. The agency uses a partnership model in assigning caseloads. Specifically, each caseload is managed by a pair of workers: an MSW-level social worker and a bachelor's-level caseworker. Therefore, parents of children in care with the agency normally work with two main caseworkers.
Parents at both sites were recruited for the study. All parents were eligible to participate except for those whose parental rights to the children in care had already been terminated and those who were younger than 18 years of age (because permitting minor parents to participate would require consent from their legal guardians, thus compromising confidentiality).
Parent Engagement. Indicators of parent engagement were created on the basis of the earlier mentioned themes from the theoretical social work literature indicating that family-focused casework promotes parent engagement by making parents feel empowered, supported, respected, and understood. It is presumed that when parents feel in receipt of family-focused services (that is, feel empowered, and so forth), they will be more actively engaged in services and, therefore, more likely to succeed, that is, reunify with their children).
With these constructs in mind, items were generated to tap two dimensions of parents' experience: (1) the degree to which parents perceive their caseworkers to be doing family-focused actions (that is, the degree to which parents perceive their caseworkers to be empowering, and so forth) and (2) the degree to which parents feel empowered, respected, understood, and supported (presumably, as a result of those caseworker actions). Two items ("My caseworkers help me see my child[ren] often" and "I am involved in decisions made about my case") were drawn from a related measure of parent satisfaction with foster care services (see Harris et al., 2000). Items in part 1 include statements such as "My caseworkers involve me in meetings about my case" and "My caseworkers are available when I need them." Examples of items in part 2 are "I have control over whether or not I succeed in the foster care process" and "When I talk with my caseworkers about my personal situation, I feel like they really listen to me." Because of their connection to the theoretical literature, these items have strong face validity. (In light of the partnership model mentioned earlier, it was important to agency administrators that the measure ask parents to reflect on their experience with the team [as opposed to their experience with one caseworker]. Therefore, the directions on the parent questionnaire ask parents to consider both their social worker and case worker when responding to the items.) Parents rate their agreement with these statements using an anchored six-point Likert-type scale ranging from 1 = strongly disagree to 6 = strongly agree. Negatively worded items were included to protect against response sets (Netemeyer, Bearden, & Sharma, 2003) and were reverse coded for analysis.
Demographic Data. The questionnaire packet ended with a sheet asking parents to report demographic information and some case-related information, including children's type of placement (for example, kinship or nonkinship foster home); length of time working with caseworkers; and number of children in care.
Case-related Variables. Additional case-related variables were gathered from electronic case records. Each responding parent's case record was used to code for date of case opening, reason for children's removal, and whether or not services were indicated for the parent pertaining to mental health (examples of elements of the case record coded as "indicated mental health services" included need for cognitive evaluation, enrollment in scheduled individual therapy appointments, and being under psychiatric care), substance abuse, domestic violence, or a (perpetrated) sexual abuse allegation. The demographic and case-related variables were selected to provide some measure of construct validity, as they are highlighted in the literature as factors associated with or influencing parental collaboration and compliance as well as case outcome (for example, HHS, 2008; Jones, 1998; Littell & Tajima, 2000; Terling, 1999; Wells & Guo, 2003).
The first author collected data for approximately 10 weeks at site A and five weeks at site B between September 2006 and January 2007. Prior to the on-site recruitment, parents were sent a letter in the mail explaining that at one of their next visits to the agency they might be asked to participate in a research study. The letter overviewed the protocol for consent and confidentiality.
The researcher was on-site several days a week, stationed in the waiting room to approach parents waiting there (usually for parent-child visits supervised by social work staff). On meeting a parent, she explained that she was conducting a survey to find out more about parents' experiences at the agency. Maximizing parents' privacy in particiapating was crucial for two reasons. First, parents needed to give consent and complete the survey in a confidential environment. Second, a confidential setting was necessary to maximize parents' truthful responding; in light of their vulnerable situation, parents of children in foster care may be reluctant to be completely forthcoming in discussing their caseworkers for fear that a negative report might jeopardize their chances at reunification. Therefore, if, in the waiting room, the parent expressed an interest in participating, the researcher took the parent to a separate room to present the study further. In the separate room, the parent reviewed the consent form and provided written consent.
To ensure that parents reported on their relationships with their specific workers, the researcher asked the parent, "What are the names of your two main workers?" After the parent noted the names of the workers, the researcher said, "Okay, then this survey will ask you to think about your relationship with [Name 1] and [Name 2]." Depending on the parent's preference, he or she either read through the items and completed the survey or the researcher read each item out loud and circled the number that the parent selected. It took parents approximately 10 minutes to complete the questionnaire. During this time, parents were able to ask the researcher for clarity on the meaning of items or on how to use the six-point scale, which they did often.
In all cases, the researcher collected the survey from the parent when he or she was finished and presented the incentive: One chance to win a 30-day public transportation pass ($76 value). On returning the packet, the participant had one chance to select a card blindly from a bag of 10 cards. One card was labeled as the winning card. If the participant selected the winning card, he or she received the prize on the spot.
Spanish Language Procedures. A Spanish-speaking research assistant implemented the procedures with Spanish-speaking parents by using translated materials. The first author trained the assistant on the study procedures and prepared an instruction manual for the assistant to use as a reference. The first author supervised the assistant twice on-site and then sent the assistant to work alone on eight additional days. The assistant worked solely at site A because this location serves the majority of the agency's Spanish-speaking families.
Management of Missing Data
There were very few instances of missing data. Still, to make maximum use of a relatively small sample, a strategy was used to impute data when missing. A preliminary scale analysis of the parent engagement measure showed that the total 31-item scale had good reliability (Cronbach's [alpha] = .93). Therefore, when a score was missing for an individual item on the scale, the researcher placed that respondent's mean scale score in for the missing data. This insertion was necessary in fewer than 10 out of 1,426 instances (31 items x 46 respondents), making an imperceptibly small change to the distribution of the collected data.
Eighty-nine percent of parents approached agreed to participate in the study, yielding a sample size of 47 parents. One participant was excluded from the analyses because of strong evidence of response sets (that is, consistently selecting the extreme value on one side of the scale despite variations in positive and negative wording of items).Therefore, all analyses dealt with a parent sample of 46. Just over 70% of parents were clients at site A. The vast majority were women (80%) and used the English language instruments (94%). Forty-six percent of parents were categorized as black and 46% as Hispanic; the remainder identified either as white (two parents) or as having multiple racial and ethnic origins, including white (two parents).
Approximately half of the parents indicated they were planning as single parents for their children's return; the others noted that they were planning with a spouse or partner. (In several instances, both members of coupled parents were surveyed; as such, the sample of 46 parents represents only 40 cases.) Parents had an average of 2.72 children in foster care (SD = 1.44) and had an average case length of 30 months (SD = 29.15). That mean reflects a small number of especially long cases; the median case length, 19 months, likely represents a more accurate assessment of the length of the average case in this sample.
Some parents indicated working with two main workers (according to the casework team model mentioned earlier); however, others named only one. Therefore, a new variable, "time spent working with longest-running caseworker," was created. Parents reported working with their longest-running worker for an average of 13 months (SD = 13.68, Mdn = 9 months). Fifteen percent of parents represented cases that had been transferred to the current agency from another agency.
Parents' children rived in a variety of foster placements. Twenty-eight percent of parents reported that their children were in kinship homes or divided among kinship homes and other types of placements (for example, nonkinship homes, residential treatment facilities). Seventy-two percent reported that their children were in nonkinship homes or divided among nonkinship homes and residential treatment facilities. Reason for removal was collapsed into three dichotomous (yes/no) variables reflecting whether or not physical abuse, neglect, or a drug-related issue was noted in the case record as the reason for removal. Twenty-eight percent of parents' records noted physical abuse as the reason for removal, 54% noted neglect, and 39% noted a drug-related concern. Similarly, indicated services for parents were coded from the case record by using dichotomous (yes/no) variables. Mental health services were indicated for 65% of parents, substance abuse services were indicated for 65% of parents, and domestic violence services were indicated for 33% of parents. (Total percentages exceed 100% because it is possible for a parent's record to note more than one reason for removal and more than one type of indicated service.) Case records indicated an average of 1.84 types of indicated services (SD = 0.976, Mdn = 2) per parent.
This pilot study represents the first testing of this questionnaire. Means and standard deviations for each item on the parent engagement measure are provided in Table 1. Immediately noticeable was that the reverse-coded items had the highest variability, that is, standard deviations ranging from 1.77 to 2.00 in contrast to the remainder of the items, with standard deviations ranging from. 147 to 1.69 (with one exception, "I trust my caseworkers"; SD = 1.78). This finding was logical in light of the first author's experience in administering the questionnaire. As parents made their way through the survey, many got caught up on the negatively worded items; they struggled to interpret the double negatives (that is, to comprehend the negatively worded item and then reverse their frame of reference on the six-point scale) and needed considerable guidance from the researcher as to how to select the scale numbers that accurately reflected their positions. In the end, because of the questionable validity of the negatively worded items, those items were dropped from further analyses. (The utility of reverse coded items is explored further in the Discussion.)
The remaining 26 items showed good internal consistency (Cronbach's [alpha] = .93). A scale score was created for each participant by taking the mean of the 26 item scores (M = 4.88, SD = 0.856). The sample was divided into quartiles on the basis of this mean score. A t test was then run on each of the 26 items to determine whether the individual items differentiated between the first (that is, lowest scoring or least engaged) and fourth (that is, highest scoring or most engaged) quartiles of responding parents. All of the items differentiated between these two groups except for the following: "I believe I will complete my service plan successfully," "I believe I will reunify with my children," "I have the ability to be a good parent to my children," and "I am making good progress toward reunification" (see Table 1). Because these four items were not useful in differentiating low- and high-scoring parents, they were dropped from the scale as well, reducing the total measure to 22 items. The final instrument continued to show good reliability (Cronbach's [alpha] = .94).The re-calculated mean score for the 22-item measure had a group mean of 4.74 (SD = 0.972). This mean score, henceforth referred to as the "mean engagement" score, was used in all subsequent analyses.
Scale Score Analyses
T tests and one-way analyses of variance (ANOVAs) were run to determine any within-group differences in mean engagement based on the demographic and case-related variables collected. Parents' mean engagement scores did not differ on the basis of site, gender, placement type (kinship or nonkinship home), status as a transferred case (transfer or not a transfer), relationship status with respect to planning (planning as a single parent or planning with a partner or spouse), or the language of the survey materials (Spanish or English). An ANOVA testing mean differences among the three main racial or ethnic groups yielded a significant omnibus statistic [F(2, 45) = 3.707, p = .033], but a Tukey honestly significant difference post hoc test revealed only a marginally significant (p = .056) difference between black parents (n = 21) and those parents who categorized themselves as white or of multiple backgrounds, including white (n = 4), with black parents reporting significantly lower mean engagement scores.
Full-factorial ANOVAs investigated within-group differences on the basis of reason for removal and indicated services for parents. The three dichotomous (yes/no) variables representing the presence of physical abuse, neglect, and drug-related issues in the reason for removal were entered into the model. This test yielded no main effects or interactions. A similar full-factorial ANOVA was run by using the three dichotomous variables (yes/no) indicating the presence of mental health, substance abuse, and domestic violence services in each parent's service plan. Again, the omnibus test was not significant.
Pearson correlations were used to examine the relationships between mean engagement and the continuous demographic and case-related variables. The following were entered into a correlation matrix: mean engagement score, distance from parent's home to the agency site, number of children in foster care, length of time the parent had been working with longest running caseworker, case length, and number of types of indicated services for the parent in the case record (that is, number of "Yes" codes a parent received when coding for the presence of the various types of services). Mean engagement score was significantly related to distance from the parent's home to the agency [r(45) = -.429, p = .003], and time spent working with the longest running caseworker [r(44) = -.309, p = .042]. As expected, case length was significantly and positively correlated with time spent working with longest running worker [r(43) = .656, p < .001]; however, case length was not significantly associated with mean engagement.
Scale Development. Results indicate that the parent engagement measure shows good internal consistency. The retained items exhibit considerable variability based on the six-point scale, and the ultimate scale score shows good variability as well.
In the initial 31-item instrument, two groups of items emerged that had debatable utility. The first group contained the five reverse-coded items. The wide variability in these items was understood in light of the participants' difficulty in interpreting the statements and selecting appropriate responses. The negatively worded items were initially included as a common strategy for protecting against response sets. However, during the actual implementation, the researcher was able to monitor each parent's survey completion in a more vigilant way than initially expected. As such, she was able to answer any questions parents had regarding the items, ensuring that parents understood the statements and maximizing the validity of the tool. In this way, the researcher's ability to assist parents reduced the concern regarding response sets. Given this supervised survey environment, the fact that parents still seemed to struggle with the negatively worded items led to the conclusion that if the researcher is on hand to guide the parent through the process, reverse-coded items, which have the potential to be confusing regardless of the administrator's supervision, may not be necessary.
The second group of questionable items consisted of the following: "I believe I will complete my service plan successfully," "I believe I will reunify with my children," "I have the ability to be a good parent to my children," and "I am making good progress toward reunification. "Though one would expect a highly engaged parent to respond in the affirmative to these statements, these items do not reflect engagement in the parent-worker relationship or with aspects of the parent's personal service plan. In contrast, these items reflect parents' more global sense of personal efficacy and faith in success. Although they did not harm the measure's reliability (that is, removing them changed the scale score's alpha from .93 to .94), the items did not differentiate high- and low-scoring parents and were therefore deemed--at least with respect to this measure of parent engagement--not useful. It is worth considering, however, that these discarded items, or items similar to them, may be useful in future studies of parent engagement. Indeed, they seem to reflect a parent's sense of hope, a construct that, in addition to engagement, may influence parents' completion of their service plans.
Within-Group Differences. The demographic variables that clearly differentiated parents' mean engagement scores were distance from the parent's home to the agency and time spent working with longest running worker.
Distance was negatively related to the dependent variable, suggesting that parents who live farther away from the agency feel less engaged. This finding makes sense, as parents who must travel long distances to attend meetings and visit their children may feel more overwhelmed and isolated than parents for whom the agency is more accessible.
Time spent working with longest running caseworker was also negatively associated with engagement, indicating that the longer the relationship, the lower the engagement. Perhaps when the parent-worker relationship is new, a parent has more hope for what the relationship can provide; in other words, the longer a parent works with a caseworker, the more opportunities that parent has to be disappointed by that worker. Maybe the caseworkers' services decline in quality over time. A study that surveys parents earlier and then again later during the course of their relationships with their workers is needed to test such a hypothesis. Alternatively, it could be that the inverse relationship between length of time with worker and parent engagement simply indicates that parents with the most severe disorders require the longest casework. More detailed data should be collected to decipher this negative correlation. Of note is the finding that although time with worker was significantly correlated with parent engagement, case length was not. For future reference, this finding calls into question the significance of using time with worker as a proxy for case length.
Limitations and Directions for Future Research
Sample Representativeness and Potential Response Bias. The findings of this study should be interpreted with caution, as they may be influenced by several limitations. Findings here may appear differently in a larger, more representative sample. Also, the sample reflects almost exclusively black and Hispanic parents from one agency; therefore, results cannot be generalized to other agencies or to parents from other racial or ethnic backgrounds. Even within the agency in question, the inability to access all of the agency's parents limits the ability to draw conclusions about its clientele as a whole.
Furthermore, a selection bias likely influences the results. Because the study took place at the agency, parents had to be present to be recruited. As a result, the parent sample likely captures the middle section under what might be a curve of reunification progress. On the low end of this curve would be parents who are unknown to the agency, have abandoned their children in foster care, or are not actively engaged in planning. Such parents were not likely to be sitting in the agency's waiting room during the data collection period and were therefore not accessed. It is logical to assume that these parents are not as engaged with their workers (or in the foster care process, generally), are disconnected from the system for one reason or another, and would therefore report lower scores on the parent engagement measure.
Perhaps less obvious, but also responsible for the nonrepresentativeness of the present sample, is the project's inability to access some parents who are faring relatively well. Foster care aims to place children (and, by extension, parents) in the least restrictive environment possible. Therefore, many parents planning for their children's return are permitted to arrange visits in the community (that is, their visits do not have to be supervised at the agency); hence, they do not have to come to the agency as often as parents whose visits must take place on-site. Parents in this situation may have scored higher on the parent engagement measure, but they were also less likely to be recruited because of the procedures used.
Other layers of self-selection and potential bias exist. Parents who were approached but refused to participate may have reported lower engagement as a result of feeling more angry with the agency, overwhelmed, or hopeless than their responding counterparts. At the other end of the spectrum, responding parents may have reported inflated engagement, giving what they believed to be the "right" answers, or reporting an overly positive experience for fear that their negative reactions might get back to their workers and harm their cases.
Sample Size and Measure Development. The small sample size prohibited a useful factor analysis. In future research, a larger sample size will allow for factor analysis and the investigation of possible subscales. It is possible that the instrument consists of several factors, falling under the larger conceptual umbrella of "parent engagement," that are correlated but not identical to one another. It is also possible that the measure is unidimensional; the high alpha shown in these results suggests that this may be the case.
Unit of Analysis. An additional area for measure improvement pertains to the unit of analysis indicated in the items. Because of the participating agency's casework model, the agency preferred that the items on the measure ask parents to reflect on their relationship with both their social workers and their case workers together. Such an approach assumes that a parent feels the same way about each of his or her workers, or that a parent can combine his or her feeling about two different people into one score. A study that asks a parent to select one worker (probably their longest running worker) and to fill out a form on the basis of his or her relationship with that person is likely to capture better the nuances of the parent-worker relationship. An agency using a casework team model, such as the one mentioned here, could ask parents to fill out two forms, one for each worker. This tactic was not used in the present study out of concern that the doubled time commitment would influence parents not to participate.
Statistical Nonindependence. Two elements of statistical nonindependence may factor into the results of the study. First, six of the parents sampled represented couples from the same family. By including both members from these couples in the analyses, we may be weighting the findings from these parents more heavily than findings from other participants. Of note, findings did not indicate any significant difference in engagement scores between parents planning alone and parents planning with a partner; however, this finding is not sufficient to eliminate the fact that responses from members of couples are statistically nonindependent. In addition, parent responses may be the result of worker effects, as many responding parents shared a caseworker. Analyzing worker effects is an important area for future research, as parent engagement may differ depending on the degree to which caseworkers engage in family-centered practice.
Limitations as Catalysts for Measure Development. As in any research report, the aforementioned limitations are presented to inform the reader of factors that might bear on the interpretability of the present findings. However, it is critical that these limitations also be framed as springboards for innovation in studying the experiences of parents of children in foster care. As noted, this population is difficult to access and is only beginning to be studied in-depth. The present study should be viewed as part of this beginning; the study provides preliminary data and offers a fertile starting point for continued, refined research.
Parents of Children in Foster Care as Research Participants
Finally, in addition to yielding informative psychometric, descriptive, and inferential statistics, the process of implementing this project revealed much about parents of children in foster care as potential research participants. There are several possible reasons why parents are understudied in the child protection literature. Parents of children in care may view participation in research as "just one more thing" that they have to do as part of their involvement with the system and, in an effort to do as little as possible related to the agency, not volunteer. Other parents may be interested in participating but decline, feeling unable to manage research participation in addition to the emotional and logistical stressors built into the foster care experience.
But parents may also be understudied due to the emotional challenge of researching this particular population. Assessing parents of children in foster care means that researchers must confront their own feelings and possible biases about parents who have been neglectful or abusive. If researchers fail to take on this challenge, social work researchers and professionals will miss out on hearing parents' essential perspective. More research on parents of children in foster care is critical to a well-rounded understanding of the child welfare system in general; in a service delivery environment characterized by family-centered practice, the omission of parents' perspectives is glaring.
The participants in the present study provide a compelling argument to overcome these concerns and push forward with researching parents. First, these parents were eager to express themselves, evidenced by the fact that most parents approached agreed to participate. Second, participating parents were deliberate and thoughtful in their responses, thinking critically about each measured item and asking for clarification if the meaning of an item was unclear. And third, but certainly not least, the current project showed that most of the parents were able to make use of all points on a six-point Likert-type scale, despite our initial assumptions about this population's limitations in literacy and cognitive skills. These comments are not offered as support for the validity of the instrument; rather, we note them as a reflection on the process of doing this research and our subsequently reinforced opinion that parents of children in foster care are worth studying. This is an important message for foster care researchers; the knowledge that parents of children in foster care can make good research participants should guide scientists as they consider whether they want to explore this group of people. With an open mind toward studying parents of children in foster care, we can endeavor to learn more about this population, which so critically needs our attention.
Original manuscript received August 27, 2007
Final revision received July 17, 2008
Accepted November 12, 2008
Administration for Children and Families. (2005). National Survey of Child and Adolescent Well-Being (NSCAW). CPS sample component wave 1 data analysis report. Retrieved April 17, 2007, from http://www.acf.hhs.gov/programs/ opre/abuse_neglect/nscaw/
Alpert, L.T., & Britner, P. A. (2005). Social workers' attitudes toward parents of children in child protective services: Evaluation of a family-focused casework training program. Journal of Family Social Work, 9, 33-64.
Annie E. Casey Foundation. (2008). The story of Family to Family: The early years 1992-2006. An initiative to improve child welfare systems. Retrieved May 25, 2009, from http://www.aecf.org/~/media/PublicationFiles/ F2F%20Book%201ayout%20DRAFT%209%2012.pdf
Chapman, M.V., Gibbons, C. B., Barth, R. P., McCrae, J. S., & the NSCAW Research Group (2003). Parental views of in-home services: What predicts satisfaction with child welfare workers? Child Welfare, 82, 571-596.
Child Welfare League of America. (2002). Research roundup: Family reunification. Retrieved June 14, 2004, from http://www.cwla.org/programs/r2p/rrnews0203.pdf
Corby, B., Millar, M., &Young, L. (1996). Parental participation in child protection work: Rethinking the rhetoric. British Journal of Social Work, 26, 475-492.
Dawson, K., & Berry, M. (2002). Engaging families in child welfare services: An evidence-based approach to best practice. Child Welfare, 81, 293-317.
DePanfilis, D. (2000). How do I develop a helping alliance with the family? In H. Dubowitz & D. DePanfilis (Eds.), Handbook for child protection practice (pp. 36-40). Thousand Oaks, CA: Sage Publications.
Diorio, W. D. (1992). Parental perceptions of the authority of public child welfare caseworkers. Families in Society, 73, 222-235.
Dumbrill, G. C. (2006). Parental experience of child protection intervention: A qualitative study. Child Abuse & Neglect, 30, 27-37.
Forest, C. (2003). Empowerment skills for family workers: The comprehensive curriculum of the Cornell Family Development Credential. Ithaca, NY: Family Development Press, Corner University.
Fraser, M.W., Walton, E., Lewis, R. E., Pecora, P.J., & Walton, W. K. (1996). An experiment in family reunification: Correlates of outcomes at one-year follow up. Children and Youth Services Review, 18, 335-361.
Gillespie, J. M., Byrne, B., & Workman, L.J. (1995). An intensive reunification program for children in foster care. Child and Adolescent Social Work Journal, 12, 213-228.
Haight, W L., Black, J. E., Mangelsdorf, S., Giorgio, G., Tata, L., Schoppe, S.J., et al. (2002). Making visits better: The perspectives of parents, foster parents, and child welfare workers. Child Welfare, 81, 173-202.
Harris, G., Poertner, J., & Joe, S. (2000).The Parents with Children in Foster Care Satisfaction Scale. Administration in Social Work, 24(2), 15-27.
Hess, P. M., Folaron, G., & Jefferson, A. B. (1992). Effectiveness of family reunification services: An innovative evaluative model. Social Work, 37, 304-311.
Jones, L. (1998). The social and family correlates of successful reunification of children in foster care. Children and Youth Services Review, 20, 305-323.
Kapp, S.A., & Propp, J. (2002). Client satisfaction methods: Input from parents with children in foster care. Child and Adolescent Social Work Journal, 19, 227-245.
Kapp, S.A., &Vela, R. H. (2004a).The Parent Satisfaction with Foster Care Services Scale. Child Welfare, 83, 263-287.
Kapp, S.A., & Vela, R. H. (2004b).The unheard client: Assessing the satisfaction of parents of children in foster care. Child and Family Social Work, 9, 197-206.
Littell, J. H., & Tajima, E.A. (2000). A multilevel model of client participation in intensive family preservation services. Social Service Review, 74, 405-435.
Maluccio, A. N. (1981). Casework with parents of children in foster care. In R A. Sinanoglu, & A. N. Maluccio (Eds.), Parents of children in placement: Perspectives and programs (pp. 15-25). New York: Child Welfare League of America.
Netemeyer, R. G., Bearden, W. O., & Sharma, S. (2003). Scaling procedures. Thousand Oaks, CA: Sage Publications.
Petr, C. G., & Entriken, C. (1995). Service system barriers to reunification. Families in Society, 76, 523-533.
Petras, D. D., Massat, C. R., & Essex, E. L. (2002). Overcoming hopelessness and social isolation: The ENGAGE model for working with neglecting families toward permanence. Child Welfare, 81, 225-248.
Terling, T. (1999).The efficacy of family reunification practices: Reentry rates and correlates of reentry for abused and neglected children reunited with their families. Child Abuse & Neglect, 23, 1359-1370.
U.S. Department of Health and Human Services. (2008). The AFCARS report: Preliminary FY 2006 estimates as of January 2008. Retrieved June 21, 2008, from http://www.acf.hhs.gov/programs/cb/stats_research/afcars/tar/report14.htm
Waites, C., Macgowan, M.J., Pennell, J., Carlton-LaNey, I., & Weil, M. (2004). Increasing the cultural responsiveness of family group conferencing. Social Work, 49, 291-300.
Walton, E. (1998). In-home family-focused reunification: A six-year follow-up of a successful experiment. Social Work Research, 22, 205-214.
Wells, K., & Guo, S. (2003). Mothers' welfare and work income and reunification with children in foster care. Children and Youth Services Review, 25, 203-224.
Zamosky, J., Sparks, J., Hatt, R., & Sharman, J. (1993). Believing in families. In B.A. Pine, R. Warsh, & A. N. Maluccio (Eds.), Together again: Family reunification in foster care (pp. 155-175).Washington, DC: Child Welfare League of America.
Lily T. Alpert, PhD, is senior policy analyst, Children's Rights, New York. Preston A. Britner, PhD, is associate professor, Department of Human Development and Family Studies, University of Connecticut, Storrs, CT. The research reported in this article was conducted by Lily Alpert while she completed her doctoral studies in the Department of Human Development and Family Studies, University of Connecticut, Storrs, CT. Address correspondence to Lily T. Alpert by e-mail: firstname.lastname@example.org or to Preston A. Britner, University of Connecticut, Department of Human Development and Family Studies, U-2058, 348 Mansfield Road, Storrs, CT 06269; e-mail: email@example.com. The authors thank Maureen Mulroy and Ronald Sabatelli for their assistance in guiding this research.
Table 1: Item Analysis: Mean and Standard Deviation (31 Items), Ability to Differentiate between First and Fourth Quartile Respondents (26 Items), and Ultimate Retention Decision (22-Item Scale) Mean Difference between 1st and 4th Quartile Item M SD My caseworkers focus on my strengths. 4.61 1.358 My caseworkers make me feel like an important part of a team. 4.42 1.598 My caseworkers spend too much time focusing on my weaknesses. (REV)# 4.02# 1.891# My caseworkers involve me in meetings about my case. 5.13 1.376 My caseworkers developed a service plan based on my personal goals. 4.94 1.494 My caseworkers encourage me to share my point of view. 4.76 1.463 My caseworkers value the knowledge I have about my own children. 5.15 1.445 My caseworkers value me as a person. 4.91 1.226 My caseworkers care whether or not I reunify with my child(ren). 4.88 1.305 My caseworkers are available when 1 need them. 3.96 1.806 My caseworkers are not reliable. (REV)# 4.13# 1.973# My caseworkers help me when I ask for help. 4.80 1.344 My caseworkers help me to see my child(ren) often. 4.98 1.468 My caseworkers connect me with the services I need. 4.74 1.373 I have a say in creating the goals of my service plan. 4.33 1.687 I believe I will complete my service plan successfully.# 5.57# 0.688# There's nothing I can do to increase the possibility of reunifying with my child(ren). (REV)# 4.67# 1.765# I believe I will reunify with my child(ren).# 5.67# 0.990# I have control over whether or not I succeed in the foster care process. 5.39 0.954 I have the ability to be a good parent to my children.# 5.98# 0.147# I am involved in decisions made about my case. 4.74 1.512 I am making good progress toward reunification.# 5.33# 1.266# I feel disrespected by my caseworkers. (REV)# 4.74# 1.782# When I talk with my caseworkers about my personal situation, I feel like they really listen to me. 4.39 1.680 I feel that my opinion is respected by my caseworkers. 4.46 1.588 I feel respected as a parent by my caseworkers. 4.72 1.601 I feel alone in the foster care system. (REV)# 3.50# 2.008# I trust my caseworkers. 3.93 1.781 I am getting the services I need in order to complete my service plan successfully. 4.87 1.470 I feel connected to my children in care. 5.26 1.389 I can call my caseworkers if I need help. 4.98 1.556 Mean Difference between 1st and 4th Quartile Item t p My caseworkers focus on my strengths. -8.145 <.001 My caseworkers make me feel like an important part of a team. -6.349 <.001 My caseworkers spend too much time focusing on my weaknesses. (REV)# NA NA My caseworkers involve me in meetings about my case. -3.238 .008 My caseworkers developed a service plan based on my personal goals. -2.298 .042 My caseworkers encourage me to share my point of view. -5.006 <.001 My caseworkers value the knowledge I have about my own children. -3.778 .003 My caseworkers value me as a person. -6.436 <.001 My caseworkers care whether or not I reunify with my child(ren). -5.946 <.001 My caseworkers are available when 1 need them. -5.964 <.001 My caseworkers are not reliable. (REV)# NA NA My caseworkers help me when I ask for help. -7.173 <.001 My caseworkers help me to see my child(ren) often. -4.274 .001 My caseworkers connect me with the services I need. -4.528 .001 I have a say in creating the goals of my service plan. -4.280 <.001 I believe I will complete my service plan successfully.# -1.725# .111# There's nothing I can do to increase the possibility of reunifying with my child(ren). (REV)# NA NA I believe I will reunify with my child(ren).# -1.551# .152# I have control over whether or not I succeed in the foster care process. -2.342 .039 I have the ability to be a good parent to my children.# -1.000# .341# I am involved in decisions made about my case. -3.778 .003 I am making good progress toward reunification.# -2.162# .055# I feel disrespected by my caseworkers. (REV)# NA NA When I talk with my caseworkers about my personal situation, I feel like they really listen to me. -11.552 <.001 I feel that my opinion is respected by my caseworkers. -9.626 <.001 I feel respected as a parent by my caseworkers. -4.907 <.001 I feel alone in the foster care system. (REV)# NA NA I trust my caseworkers. 10.054 <.001 I am getting the services I need in order to complete my service plan successfully. -4.274 .001 I feel connected to my children in care. -2.400 .036 I can call my caseworkers if I need help. -5.096 <.001 Note: Shaded items = item not retained; (REV) = reverse-coded item; NA = item not included in this analysis. For all t values, df = 22. Note: Shaded items = item not retained is indicated with #.
|Printer friendly Cite/link Email Feedback|
|Author:||Alpert, Lily T.; Britner, Preston A.|
|Publication:||Social Work Research|
|Date:||Sep 1, 2009|
|Previous Article:||Beyond infrastructural issues: informal barriers to social work research.|
|Next Article:||The available time scale: measuring foster parents' available time to foster.|