RES 6551 Module 2 Descriptive Results Example

Who Answered and What They Reported: Participant Characteristics, Scale Descriptives and Correlations for a Study of Voice in Hospital Revenue Cycle Teams

Student Name

American College of Education

RES6551: Analyzing the Dissertation Research

Module 2 Assignment

Instructor Name

October 19, 2026

Introduction

Chapter 4 begins by describing the people in the study and the variables before testing any relationship among them. Readers need to know who answered before they can judge what the answers mean, and committee members will look at the descriptive tables to decide whether the sample resembles the population the study claims to inform. This paper reports the characteristics of the 221 revenue cycle employees retained after screening, the descriptive statistics and reliability of the three scales, the correlations among study variables and comparisons across health systems, roles and work locations. It ends with what these results mean for the models in Module 3.

Participant Characteristics

Participants came from three health systems, labeled A, B and C to protect their identity. System A contributed 100 participants (45.2%), System B 68 (30.8%) and System C 53 (24.0%), roughly in proportion to the size of their revenue cycle workforces. By role, 72 worked in patient access (32.6%), 54 in coding (24.4%), 48 in denials and follow-up (21.7%) and 47 in billing (21.3%). Slightly more than half, 116 (52.5%), worked on site, and 105 (47.5%) worked remotely or on a hybrid schedule. Tenure in the current role averaged 4.95 years (SD = 3.64) with a median of 3.6 years, reflecting a workforce in which many employees had joined recently and a smaller group had long service. Demographic details such as age and gender were collected in bands and are reported only in aggregate in the chapter's first table, so that no small group within a system can be identified.

Response Rate and Representativeness

Revenue cycle leaders at the three systems reported that the invitation reached about 640 eligible employees. With 241 surveys started, the response rate was about 37.7%, and the usable response rate, based on 221 retained cases, was about 34.5%. Rates were similar across systems, between 32% and 37% usable. Each system's human resources office also provided the role mix of its revenue cycle workforce in aggregate. Compared with those figures, patient access staff were slightly underrepresented in the sample and coders slightly overrepresented, by about four percentage points in each case; billing and denials staff matched closely. Because no system provided individual-level data on nonrespondents, the comparison cannot rule out other differences, such as in attitudes toward supervisors. The sample is therefore described as broadly similar to the population in its role mix, with that limit noted for Chapter 5.

Descriptive Statistics for the Study Variables

On the five-point Global Transformational Leadership scale (Carless et al., 2000), participants rated their supervisors at a mean of 3.48 (SD = 0.81), with scores ranging from 1.43 to 5.00. On the seven-point team psychological safety scale (Edmondson, 1999), the mean was 4.90 (SD = 1.08), ranging from 1.86 to 7.00. On the seven-point voice scale (Van Dyne & LePine, 1998), the mean was 4.61 (SD = 1.08), ranging from 1.33 to 6.83. All three means sit above their scale midpoints, so participants on average described supervisors who showed transformational behaviors at least sometimes, teams that felt moderately safe and a moderate tendency to speak up. Internal consistency held up in the screened sample: both leadership and safety reached alpha = .92, and voice reached .91. The full range of each scale was used, which suggests the measures were not limited by ceilings or floors.

Correlations

All three study variables were positively correlated, and the pattern matched the order the mediation model expects. Transformational leadership correlated most strongly with psychological safety, r = .59, then psychological safety with voice, r = .44, and leadership with voice, r = .35; each was significant at p < .001. With 221 cases, any correlation of about .13 or larger in absolute value is significant at the .05 level, so the control variables were essentially unrelated to the study variables. Tenure correlated .04 with leadership, -.07 with psychological safety and .00 with voice. Remote or hybrid work correlated -.03 with leadership, -.02 with psychological safety and .09 with voice, none significant. These values show the controls are unlikely to change the main estimates much, although they remain in the models as planned.

Comparisons Across Systems, Roles and Locations

One-way analyses of variance compared scale scores across groups. Psychological safety differed by health system, F(2, 218) = 3.69, p = .027, eta squared = .033: System A averaged 5.10 (SD = 1.08), System C 4.84 (SD = 0.98) and System B 4.65 (SD = 1.12). Leadership ratings differed by role, F(3, 217) = 3.58, p = .015, eta squared = .047, with patient access staff rating supervisors highest (M = 3.68) and coders lowest (M = 3.24). Leadership did not differ by system, F(2, 218) = 1.43, p = .24, and voice did not differ by system, role or work location; the largest of these comparisons, remote versus on-site voice, was F(1, 219) = 1.85 with means of 4.71 and 4.52. Psychological safety did not differ by role or location. The two significant differences were small, each accounting for less than 5% of variance.

What the Descriptives Mean for the Models

Three points carry forward to Module 3. First, the correlations are in the range that mediation analysis can examine without collinearity problems, as the variance inflation factors in Module 1 confirmed. Second, because psychological safety differed somewhat by health system, the planned models will be supplemented by a sensitivity analysis that adds system as a pair of indicator variables. This was not specified in Chapter 3 and will be reported as an additional analysis, not a change to the primary plan. Third, because coders rated their supervisors lower than other groups, the sensitivity analysis will also check whether role changes the leadership estimates. The descriptive results do not themselves answer any research question, and interpretation of these differences is reserved for Chapter 5.

Table Plan

Two tables will present these results in Chapter 4. Table 1 lists participant characteristics by health system, with counts and percentages for role and work location and the mean, standard deviation and median for tenure. Table 2 gives the mean, standard deviation, observed range and alpha for each scale, with the correlation matrix below the diagonal, so that a reader can see the descriptive and relational information together. Group comparisons will be reported in the text rather than in a separate table because only two were significant. Each table will carry a note defining the scales' ranges and stating that higher scores mean more of the construct, so that it can be read without the text. Decimal places will be consistent across both tables.

Conclusion

The 221 participants spread across three health systems and four revenue cycle roles, with nearly half working remotely or hybrid. Scale means sat above midpoint, reliability was strong and the three study variables correlated in the expected pattern while the controls did not. Small differences in psychological safety by system and in leadership ratings by role led to a sensitivity analysis that will be reported separately from the planned tests. Module 3 presents the results for each research question. Taken together, the descriptive picture gives no sign of data problems that would prevent the planned models from being run as specified.

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

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