The study of pain at rest (PAR) and movement-evoked pain (MEP) in persons with musculoskeletal pain has received substantial attention. Despite strong interest, relatively little attention has been directed to the psychometric development of these constructs. Our purpose was to explore the relationship between PAR and MEP and to examine the prognostic utility of these measures in persons with knee arthroplasty. We used prospectively collected data from persons scheduled for knee arthroplasty who had moderate to high levels of pain catastrophizing. Preoperative latent variables for PAR and MEP were developed and used to determine if they were associated with a binary latent variable of good versus poor pain and function outcome trajectories. Factor correlations were used to determine the extent to which the variance for PAR and MEP was overlapping. PAR and MEP are significant predictors of good versus poor pain and function classes. Odds ratios ranged from 1.21 to 1.64 (p < 0.001) indicating a significant increase in the likelihood of poor outcome. Correlation between PAR and MEP latent variables was high (r = 0.89; 95% CI: 0.86-0.92) indicating substantially overlapping variance. PAR and MEP, as defined in our study, can be used to make prognostic judgements regarding risk of poor postoperative outcome trajectory following knee arthroplasty. However, PAR and MEP showed substantially overlapping variance indicating that measurements of both are not necessary when making prognostic assessments. Preoperative PAR and MEP latent variables, as defined in our study, had prognostic significance for 1 year pain and function outcome trajectories. PAR and MEP latent variables had substantially overlapping variance which suggested that only one is needed to make prognostic judgements. The prognostic significance of PAR and MEP as well as their substantially overlapping variance is new to the field prognostic research in knee arthroplasty.