Does goal setting matter? The impact of employee-level and organizational-level goal properties on public employees' perception of performance appraisal fairness

Abstract

PurposeThis study examines the relationship between goal properties, both at the employee and organizational-level, and the perceived fairness of the performance appraisal system by federal employees.Design/methodology/approachWe describe the theoretical framework regarding goals and employee perceptions of performance appraisal fairness. We then develop and test four hypotheses, exploring the relationships among variables using five years of the FEVS data. To strengthen the research design, we created an agency-level dataset, by calculating agency-level averages for all the covariates. Instead of examining 500,000 federal employees each year, we are examining 80 federal agencies. Creating a panel dataset at the agency level allows us to make stronger statements about causality than using cross-sectional data.FindingsThis study finds a significant positive relationship between goal setting factors and employees' perceived fairness of performance appraisals: perceived employee-level goal difficulty and perceived organizational-level goal specificity at the agency level. The study results show that certain control variables, such as intrinsic motivation, play important roles in predicting public employees' perceived fairness of performance appraisals. Federal employees who have a higher level of intrinsic motivation show a more positive perception toward performance appraisal fairness. The appropriate use of extrinsic rewards and intrinsic motivation, combined with effective goal setting strategies in public organizations, may enhance public employees' perceived fairness of performance appraisal systems.Research limitations/implicationsThis study used the FEVS, necessitating the reduction of the sample size to agency level averages to create a panel dataset. Also, this study was limited to federal agencies in the United States, so research results may lack generalizability.Originality/valueThis paper fulfills an identified need to avoid cross-sectional research design and leverage longitudinal panel data.