In Canada, 29% of all deaths result from heart disease1. Cardiac rehabilitation (CR) programs are offered to patients to improve the health of the heart going forward2. The present study addresses factors predicting self-reported physical activity in patients participating in a 10-week cardiac rehabilitation at a local hospital, involving both patients and support partners (most often spouses). Partners can help to bolster self-efficacy (SE, i.e., patients' confidence in their own ability to exercise as prescribed) and relationship-inferred self-efficacy (RISE; i.e., patients' perceptions of how confident their partner is in their ability to exercise), which could potentially predict successful outcomes3,4. It was hypothesized that (H1) higher values of RISE and SE would predict more physical activity, (H2) agreement between levels of RISE and SE would be associated with more activity, and (H3) should there be discrepancies between RISE and SE, then more activity would be associated with discrepancies in the direction of RISE scores exceeding SE scores. The data were analyzed using polynomial regression with response surface analysis, first with regression analyses in SPSS, then using an Excel template to visualize patterns of discrepancies5. As hypothesized, higher RISE and SE scores predicted more activity; furthermore, agreement between RISE and SE predicted more activity, over and above the individual levels of the variables. However, contrary to H3, when there was a discrepancy between RISE and SE, more activity was associated with SE being higher than RISE. This research has potential implications for developing CR methods and in improving outcomes for CR patients.