PURPOSE: This study addressed a gap in the social work knowledge base by defining and operationalizing the affective component of human caring in the context of child welfare practice. Previous research has confirmed the relationship between human caring/related constructs (e.g., altruism, service orientation, sense of mission) and retention among child welfare workers (Ellett, 2000; Ellett & Ellett, 1996; Ellett, Ellett, & Rugutt, 2003; Harrison, 1995; Kern, et al., 1993; Landsman, 2001; Rycraft, 1994). Increased knowledge of the human caring construct has significant practical implications in the resolution of the current child welfare workforce crisis. The purpose of this study was to improve the fit between the only existing measure of human caring for social work, the Human Caring Inventory (HCI)(Ellett & Ellett, 1996), and the theoretical model (Noddings, 1984)from which it was derived (improve face and content validity); improve internal consistency reliability of data; and examine the factor structure, test-retest reliability, and criterion-related validity of the revised version of the HCI.

METHODOLOGY: The population (N=2190) of public child welfare workers in Georgia engaed in direct services was surveyed by mail. Participants completed three measures: demgraphic questionnaire; Revised Human Caring Inventory (RHCI); Intent To Remain Employed (IRE). To examine criterion-related validity, a random sub-sample (n=300) of the population of workers completed the above measures and had his/her supervisor complete two measures: RHCI adapted for supervisory ratings; and Workers' Capabilities measure. To examine test-retest reliability, a purposive sub-sample of workers (n=39) completed the study measures a second time, at a two-week interval. Data analysis procedures included: principal components analyses; reliability analyses (Cronbach alphas and coefficients of stability); bivariate correlations between workers' RHCI scores and (a)supevisory ratings of workers' levels of human caring, and (b) supervisory ratings of Workers' Capabilities; stepwise regression to examine the extent to which the factored subscales of the RHCI explained variation in IRE.

RESULTS: A response rate of 36% resulted in a sample of 786. Human caring is a multidimensional construct. Six factors explained 42% of the item variance: Receptivity; Interpersonal Reward; Professional Commitment; Commitment to Clients; Personal Attachement; Respect for Clients. Alphas for the RHCI subscales ranged from .62 to .83. Stability coefficients ranged from .59 to .91. The correlaiton (r=.52) between workers' self-ratings on the Professional Commitment subscale and supervisory ratings on this subscale was statistically significant. The correlaiton (r=.63)between workers' self-ratings on the Commitment to Clients subscale and superviory ratings of Workers' Capabilites was statistically significant. A regression model including three of the RHCI subscales explained 27% of the variance in IRE.

IMPLICATIONS: Professional commitment should be considered in the pre-employment screening process of child welfare workers; retention of workers might be enhanced if professional commitment was nurtured in the workplace. Findings indicate a need to expand the nomological net and explore the relationship between human caring and other important criterion variables (e.g., workers' collaboration with clients, client outcomes).