Development and Validation of Financial Self-Efficacy Scale for Battered Women

Purpose: Financial literacy training has been emerging as a promising intervention to empower battered women. However, evaluation of this intervention has limited validity due to small sample sizes, high attrition rate, and lack of standardized measurements. This study aims to develop a standardized financial self-efficacy scale for battered women to support future evaluations. Although scholars have extensively discussed the self-efficacy construct and some general self-efficacy measurements have been developed, little is known about financial self-efficacy. General self-efficacy scales have limited predictive and evaluative value since they do not measure specific domains of functioning.

Methods: Thirty items were developed based on the five domains of short-term financial management. The process that was used to validate this scale included: (1) having an expert panel review the construct-related validity of the items; (2) conducting “think-aloud” interviews with five people to check the response-related validity; (3) conducting pilot tests with 35 undergraduate and 15 graduate students; and (4) doing a focus group discussion with the 15 graduate students to get their feedback. After the pilot test, the number of items was reduced to 24. Then, the scale was tested with 91 battered women nationwide. The sample consisted of Caucasian (71%); Hispanic (9%); Asian American (6%); African American (6%); Native American (3%); and Biracial/Multiracial (5%) women. The majority of women had at least a college degree (70%) or some college degree (25%). Data were analyzed using jMetrik software to check the reliability, validity, and items functioning. Participants were recruited through email listservs and domestic violence services networks nationwide. They completed the scale online.

Results: The reliability analysis yielded a coefficient alpha of 0.9247 and Standard Error Measurement (SEM) of 4.8297. All items had item-Total Pearson scores higher than 0.2, and there were no negative item-Total Pearson. Three items regarding “income tax filing” and one on “equal financial decision making in relationship” had outfit indices larger than 2.0, which indicated a poor fit to the model. These three items have been removed from the scale. The item non-parametric curve analysis showed that the rating scale was working well and there were no odd-looking curves. The Wright map (showing the actual item difficulty) and the Construct map (hypotheses about item difficulty) were nearly parallel, indicating good evidence of validity. The person-item map showed that all items were below or at the same level with a person’s ability, a desired outcome given the high level of education in this sample.

Implications: This scale shows promise as a measure of financial self-efficacy for battered women, filling an important gap in our ability to evaluate financial literacy programs for battered women. Results suggested all items correlate well with the overall measurement and have a good ability of discrimination. Statistical analyses also showed good evidence of reliability and validity. Since measurement is sample specific, further studies to test this scale with other populations will be helpful to determine if there is a difference between responses among different groups.