A Quasi-Experimental Test of Two Survey Incentive Conditions with Child Welfare Staff

Background and Importance:Child welfare staff are surveyed often, as they are a crucial source of information about all aspects of child welfare policy, practice, and service utilization. Public employee ethics rules, however, often prohibit traditional direct monetary incentives in the recruitment of public child welfare staff in research. Low survey response rates create a worrisome gap in the research base most often utilized to inform the direction of child welfare policy and practice. With little research identifying alternative incentives for improving survey response rates, either generally or among child welfare and social services professionals, researchers lack guidance in how to maximize participation among these overburdened professionals. The goal of this research is to test the use of two types of incentives on response rates, comparing an incentive that tapped respondents’ self-interest and an incentive that tapped their benevolence toward clients in a web-based statewide survey in order to generate new knowledge about effective strategies to improve research participation.

Methods:All client-facing staff (N=2,196) were invited to complete the staff survey as part of a statewide needs assessment. Child welfare staff are assigned to local offices within counties, and staff were informed that the local offices achieving the top ten response rates statewide would receive a prize. The prize offered to staff in each office varied by the experimental condition. Randomization took place at the county level, with half of the state’s counties assigned to the benevolence condition (receipt of new toys and books for the family visitation room) and half to the self-interested condition (receipt of a Keurig coffee machine and a supply of K-cups for the staff). Halfway through the survey period, incentive conditions were balanced such that all remaining potential respondents were told that the top ten responding offices would receive both types of incentives. Group differences were observed with z-scores after the first phase of survey administration and again after incentive conditions were balanced.

Results:After the initial round of incentive conditions, the response rate among staff in the benevolence condition was significantly higher than that of staff in the self-interested condition (z= 3.60, p<.001). After incentive conditions were balanced at the study midpoint, staff in the self-interested condition, newly informed about the incentive tapping benevolence toward clients, had increased response rates such that significant group differences were erased (z= 1.51, p<.13).

Comparisons were also run to determine whether variations in staff response patterns were related to office-level characteristics, including a culture that leans toward family preservation or child protection, worker alliance with families and degree of feedback given to staff.

Conclusions and Implications:Our study suggests that child welfare staff may be more inclined to respond to a survey when they think that their participation will directly benefit their clients. This finding is informative to researchers with limited empirically supported avenues for incentivizing survey participation by public child welfare staff.  Additionally, this information may aid child welfare managers in identifying what motivates child welfare professionals for improving the well-being of clients at large.