Objective: Recent literature suggests that practice effects (PEs) may have utility for early detection of cognitive decline (Duff et al., 2007; Duff et al., 2012). PEs are often attributed to declarative and procedural learning; however, Suchy et al. (2011) recently proposed an alternative model of PE suggesting that effects of task novelty may also influence PE. They describe these novelty effects as transient initial performance decrements in response to task novelty. We examined learning and novelty effects as independent predictors of PE in older adults across a spectrum of cognitive functioning. Method: We examined PEs in 63 older adults ranging from healthy to moderately impaired cognitive status. Participants completed a brief battery of cognitive tests. Learning was measured with a list-learning test. Novelty effect was measured using a computerized motor planning task. PE were calculated as the difference between two trials of the WAIS-IV Coding and Symbol Search subtests administered at 30-minute intervals. Results: Heirarchical regressions revealed that novelty effect predicted PE on WAIS-IV Symbol Search (β = .008, t(62) = 3.338, p = .001) while learning predicted PE on WAIS-IV Coding (β = .116, t(62) = 2.250, p = .028). Conclusion(s): The results indicate that explicit learning and novelty effect are unique predictors of PE. This suggests that PE is a non-unitary construct that is influenced by the specific cognitive constructs tapped by the measure on which PE are observed. This finding has implications for the use of PE as a diagnostic indicator.