Date of Award

Fall 12-20-2021

Document Type


Degree Name

Doctor of Philosophy in Industrial/Organizational Psychology (PhD)


Industrial/Organizational Psychology

First Advisor/Committee Member

Joey A. Collins, Psy.D.

Second Advisor/Committee Member

Lynette H. Bikos, Ph.D., ABPP

Third Advisor/Committee Member

Helen H. Chung, Ph.D.


workplace coaching, coaching effectiveness, perceptions of supervisor support, occupational self-efficacy, work engagement, multilevel analysis


Objective: Managerial coaching is a nascent area of research; as such there are few models supported by independent inquiry endorsing their effectiveness in impacting employee outcomes. The purpose of the current investigation was to present a concise, practical model of managerial coaching (RAD: relationships, accountability, development) that can be used for teaching and evaluating coaching behaviors. The model was hypothesized to predict ratings of coaching effectiveness (CE), perceptions of supervisor support (PSS), occupational self-efficacy (OSE), and work engagement (WE). Further, each factor of the model was tested in a series of secondary hypotheses to determine which factors were the most influential in predicting each outcome.

Method: Participants consisted of 1477 employees who reported to 439 managers enrolled in managerial coaching workshops belonging to a variety of organizations from over 30 countries. Each employee rated managers on their coaching behaviors, CE, and PSS. They also provided self-ratings about their OSE and WE. This cross-sectional data was used in a series of multi-model regressions using a compositional approach to centering, which allows analysis of individual (L1) and group effects (L2) in addition to cross-level interactions.

Results: The RAD model predicted CE (βL1 = 0.66; βL2 = 0.83), PSS (βL1 = 0.42; βL2 = 0.43; βinteraction = -0.17), OSE (βL1 = 0.18; βL2 = 0.15), and WE (βL1 = 0.45; βL2 = 0.39) with all significance levels at p < .001. The L1 effects support the use of coaching with each direct report; the L2 effects suggest that outcomes show additional improvements when managers coach all their employees rather than just some. The cross-level interaction for PSS indicates that when managers coach all their employees, it can act as a buffer effect for employee perceptions even when managers do not do supportive behaviors for an individual employee. Secondary hypotheses revealed that each factor had differing individual- and group-level effects, suggesting that each factor could be used strategically to intentionally improve the different outcomes.

Conclusions: This study adds to the mounting evidence for the potential effectiveness of managerial coaching across a variety of outcomes. The examination of each factor as a separate predictor provided insights about how managers might leverage strategic coaching, indicating that further research on multi-factored models may consider a similar nested design and statistical approach. Ultimately, the RAD model shows great promise for organizations interested in developing leaders and improving outcomes for employees.


Copyright Status