| Course | RES 6551 Analyzing the Dissertation Research |
|---|---|
| Module | Module 3 |
| Paper type | Results by research question |
| Length | 1,190 words, about 4 pages plus title and reference pages |
| Format | APA 7 student paper |
| School | American College of Education |
| Program | Ed.D. and DBA doctoral core |
| Updated | October 2026 |
Free sample paper for RES 6551 Module 3
Leadership, Safety and Speaking Up: Results for Each Research Question in a Mediation Study of Hospital Revenue Cycle Staff
Student Name
American College of Education
RES6551: Analyzing the Dissertation Research
Module 3 Assignment
Instructor Name
October 26, 2026
Introduction
This paper reports the results that answer the study's three research questions, using the 221 cases prepared in Module 1. Each question is restated, the analysis is identified, the results are given with test statistics, effect sizes and confidence intervals, and the hypothesis decision is stated. All models include tenure and work location as controls, as Chapter 3 specified. Residual diagnostics are reported once, after the models, and two sensitivity analyses close the paper. Interpretation is held back for Chapter 5. Statistics are reported to two or three decimal places, with p values given exactly except when below .001, and unstandardized coefficients are listed first because they keep the scales' original units.
Research Question 1
The first research question asked whether supervisors' transformational leadership is associated with employees' voice behavior after controlling for tenure and work location. Voice, measured with the six-item scale of Van Dyne and LePine (1998), was regressed on leadership, measured with the seven-item global scale (Carless et al., 2000), and the two controls. The model was significant, F(3, 217) = 11.02, p < .001, R-squared = .132. Leadership was positively associated with voice, b = 0.468, SE = 0.084, t(217) = 5.56, p < .001, 95% CI [0.302, 0.634], standardized beta = .35. A one-point higher leadership rating on the five-point scale corresponded to nearly half a point more voice on the seven-point scale. Neither tenure, b = -0.001, p = .94, nor remote or hybrid work, b = 0.216, p = .11, was significantly related to voice. The first null hypothesis, which predicted no association, was therefore rejected. This model also supplies the total effect, c, used in the mediation analysis.
Research Question 2
The second research question asked whether transformational leadership is associated with team psychological safety, controlling for the same variables. Psychological safety, measured with the seven-item team scale (Edmondson, 1999), was regressed on leadership, tenure and work location. The model was significant, F(3, 217) = 39.45, p < .001, R-squared = .353. Leadership was strongly and positively associated with psychological safety, b = 0.791, SE = 0.073, t(217) = 10.81, p < .001, 95% CI [0.647, 0.935], standardized beta = .59. Tenure, b = -0.026, p = .11, and work location, b = -0.012, p = .92, were not significant. The null hypothesis was rejected. This coefficient is the a path of the mediation model.
Research Question 3
The third research question asked whether psychological safety mediates the association between leadership and voice. The model was estimated with ordinary least squares regression and a percentile bootstrap of the indirect effect with 5,000 resamples, the approach recommended for indirect effects because their sampling distributions are often not normal (Preacher & Hayes, 2008; Hayes, 2022). With leadership, psychological safety and the controls predicting voice, the model was significant, F(4, 216) = 15.06, p < .001, R-squared = .218. Psychological safety was positively associated with voice when leadership was held constant, b = 0.362, SE = 0.074, t(216) = 4.87, p < .001, 95% CI [0.215, 0.508]; this is the b path. The direct association between leadership and voice, c', fell to b = 0.182, SE = 0.099, t(216) = 1.83, p = .068, 95% CI [-0.014, 0.378].
The indirect effect of leadership on voice through psychological safety was ab = 0.286, bootstrap SE = 0.066, 95% CI [0.162, 0.419]; because the interval excludes zero, the null hypothesis of no indirect effect was rejected. The indirect effect accounted for about 61% of the total association (0.286 of 0.468).
How the Paths Fit Together
The coefficients are internally consistent: the total effect equals the direct effect plus the indirect effect, 0.468 = 0.182 + 0.286. Because the direct effect was not statistically significant once psychological safety entered the model, the pattern is one in which most of the association between leadership and voice runs through psychological safety in this sample. Following current guidance, the result is described in terms of the size and interval of the indirect effect rather than labeled full or partial mediation, since a nonsignificant direct effect may reflect limited power rather than a true absence of a direct path (Hayes, 2022). Figure 1 in Chapter 4 will show the model with unstandardized coefficients on each path.
Incremental Variance
A model comparison shows how much psychological safety adds to the explanation of voice beyond leadership and the controls. The Research Question 1 model explained 13.2% of the variance in voice. Adding psychological safety raised that to 21.8%, an increase of 8.6 percentage points, F change(1, 216) = 23.75, p < .001. Expressed as Cohen's f-squared for the model in Research Question 1, leadership and the controls together produced f-squared = .15, conventionally a medium effect; for the a path model, where leadership explained 35.3% of the variance in psychological safety with the controls, f-squared = .55, conventionally large. These figures describe the size of the associations in this sample and are reported so that later studies can compare against them; what they mean for practice is taken up in Chapter 5.
Residual Diagnostics
Assumptions were checked on the residuals of the full model. Plotted against fitted values, the residuals formed no curve, which supports linearity. The Durbin-Watson statistic was 2.05, indicating independent errors. The normal Q-Q plot of residuals followed the reference line closely. The Breusch-Pagan test for heteroscedasticity was at the boundary of significance, chi-square(4) = 9.50, p = .050, and the residual plot showed slightly wider spread at lower fitted values. Because of that borderline result, all three models were re-estimated with heteroscedasticity-consistent (HC3) standard errors. No conclusion changed: every coefficient significant in the main models remained significant, and the direct path remained nonsignificant. Variance inflation factors, reported in Module 1, were all below 1.6.
Sensitivity Analyses
Two sensitivity analyses tested whether results depended on decisions made after data collection. The first, added after Module 2 found small differences by system and role, included indicator variables for health system and role in all three models. Estimates changed little: the total effect was b = 0.502, the a path 0.786, the b path 0.387 and the indirect effect 0.304. The direct path rose slightly to b = 0.198, t = 1.96, p = .051, again just short of significance. The second removed the six fast responders kept in Module 1. The a path was 0.772, the b path 0.365, the direct path 0.152, p = .13, and the indirect effect 0.282. In both analyses the pattern and every hypothesis decision were the same as in the main results.
Summary of Hypothesis Decisions
All three null hypotheses were rejected. Transformational leadership was positively associated with voice (RQ1) and with psychological safety (RQ2), and psychological safety carried a significant indirect effect from leadership to voice, about three fifths of the total (RQ3). The direct association between leadership and voice was not significant with psychological safety in the model. Results were robust to unequal variance, to added controls for system and role and to removing fast responders. Module 4 turns these results into APA tables and figures.
References
Carless, S. A., Wearing, A. J., & Mann, L. (2000). A short measure of transformational leadership. Journal of Business and Psychology, 14(3), 389-405. https://doi.org/10.1023/A:1022991115523
Edmondson, A. (1999). Psychological safety and learning behavior in work teams. Administrative Science Quarterly, 44(2), 350-383. https://doi.org/10.2307/2666999
Hayes, A. F. (2022). Introduction to mediation, moderation, and conditional process analysis: A regression-based approach (3rd ed.). Guilford Press.
Preacher, K. J., & Hayes, A. F. (2008). Asymptotic and resampling strategies for assessing and comparing indirect effects in multiple mediator models. Behavior Research Methods, 40(3), 879-891. https://doi.org/10.3758/BRM.40.3.879
Van Dyne, L., & LePine, J. A. (1998). Helping and voice extra-role behaviors: Evidence of construct and predictive validity. Academy of Management Journal, 41(1), 108-119. https://doi.org/10.2307/256902
What the RES 6551 Module 3 instructions ask for
In the third module of RES 6551 you usually report the central results of the dissertation. Take the research questions in order: restate each one, name the analysis, report the results with the statistics your field expects and state the decision on each hypothesis. For quantitative studies, that means test statistics, degrees of freedom, p values, effect sizes and confidence intervals; for qualitative studies, themes with supporting excerpts for each question. Report assumption checks and any robustness or sensitivity analyses, clearly labeled. Resist the urge to explain or discuss findings here. Chapter 4 says what was found; Chapter 5 says what it means. If a planned analysis could not be run as specified, say so here and explain what replaced it.
How this RES 6551 Module 3 example is built
Each research question gets its own section with the restated question, the model, the overall fit, the key coefficient with its interval and standardized value, the controls and the hypothesis decision. The mediation section reports the b path, the direct path and the bootstrapped indirect effect, then a short section confirms that the paths add up and explains why the result is not labeled full mediation. Residual diagnostics follow, including a borderline test for unequal variance and the robust re-estimation it prompted. Two sensitivity analyses report their own estimates, and a final section summarizes every hypothesis decision in one paragraph.
Reading the RES 6551 Module 3 rubric
Results sections are graded on accuracy, completeness and discipline. Faculty check that every research question is answered with the analysis Chapter 3 promised, that statistics are complete and correctly formatted in APA 7, that effect sizes and intervals accompany significance tests and that hypothesis decisions follow correctly from the results. Assumption checks and any departures from the plan should be reported openly. Interpreting results, comparing them with the literature or drawing implications belongs in Chapter 5 and costs points when it appears here. Numbers in text must match tables and figures exactly, so careful cross-checking is part of the grade. Labeling added analyses clearly is part of the same discipline.
RES 6551 Module 3 help from the desk
The results section is where the analysis meets the page, and small errors in degrees of freedom, intervals or hypothesis wording are easy to make. If you have output but are unsure how to report it, or your chair has asked for effect sizes or a clearer mediation write-up, our writers can help. Send your output files, research questions and analysis plan with the prompt, and we will draft results for each question in APA style, with diagnostics and any sensitivity checks labeled. Regression, ANOVA, mediation and thematic results are all within our range. Clear results make the discussion chapter far easier to write. We also check every number against your output.
Write yours, or have the desk draft it
This paper is an original model document written by our desk, not a submitted student paper and not an official American College of Education document. Read it for the moves, then write your own to the instructions in your classroom. If you want one built to your exact prompt and rubric, the first custom sample is free and arrives in 24 to 48 hours.
More RES 6551 and Ed.D. and DBA doctoral core sample papers
- RES 6551 Module 1: Data Preparation and Screening
- RES 6551 Module 2: Descriptive Results
- RES 6551 Module 4: APA Tables and Figures
- RES 6551 Module 5: Complete Chapter 4
- RES 6041 Module 5: Synthesized Literature Review
- LEAD 6001 Module 2: Doctoral Self-Assessment
- LEAD 6133 Module 4: Coaching Plan
- RES 6521 Module 3: Instrumentation Section
RES 6551 Module 3 questions, answered
What does RES6551 Module 3 usually ask for?
The third RES6551 module usually asks you to present the results for each research question in order, with the analysis used, the statistics, effect sizes and the decision on each hypothesis.
How do I report a mediation result in APA style?
Report the a, b, total and direct paths with coefficients, standard errors and confidence intervals, then the indirect effect with its bootstrap confidence interval, and state whether the interval excludes zero.
Should I call my result full or partial mediation?
Many methodologists now advise against those labels. Report the size and interval of the indirect effect and the direct effect, and note that a nonsignificant direct path may reflect limited power.
Where can I find a free RES 6551 Module 3 sample paper?
Look at this page: three research questions on leadership, psychological safety and voice are answered in order with regressions, a 5,000-resample bootstrap and two sensitivity checks.
What if an assumption test is borderline?
Report it and run a robustness check, such as heteroscedasticity-consistent standard errors, then state whether any conclusion changed.