Cluster Analysis: Leadership and Community Policing Studies
This paper presents two re-evaluated analyses using cluster analysis methods applied to separate datasets. The first analysis revisits Dr. Eveland's multiple regression work on leadership effectiveness, identifying communication, planning/execution, and teamwork as the most significant predictors of overall satisfaction with a manager as a leader, with communication emerging as the dominant factor. The second analysis examines a community policing study, reviewing reliability tests and factor analyses on six scales related to training and problem-solving practices. Results indicate that community policing training and problem-solving correlate with crime rates, while factor analysis confirms that survey items form a sufficiently strong relationship to support further statistical analysis.
- Introduction to Cluster Analysis: Overview of paper's dual analytical purpose
- Leadership Effectiveness Analysis: Regression results identifying top leadership predictors
- Predictors by Reporting Group: Predictor importance varies by respondent group
- Community Policing Study Overview: Training and problem-solving correlated with crime rates
- Reliability and Factor Analysis of Community Policing Scales: Item-total correlations and factor analysis sufficiency
- Conclusions and Implications: Training implications from group-level differences
✍️ How to write this paper — guide, tools & examples ▾
What makes this paper effective
- Integrates quantitative results — including regression tables, beta coefficients, and Cronbach's alpha — to support its analytical conclusions, making claims data-driven rather than speculative.
- Clearly distinguishes between two separate datasets and analytical purposes, helping the reader follow parallel lines of inquiry without confusion.
- Uses subgroup analysis (self, direct reports, peers, supervisors) to add nuance to the leadership findings, showing how predictor importance varies by perspective.
Key academic technique demonstrated
The paper demonstrates stepwise multiple regression as a method for refining predictor selection. By starting with all predictors and then narrowing to the most statistically significant ones, the author shows how stepwise regression corrects for issues caused by varying means and variances among variables, ultimately producing a more interpretable and parsimonious model.
Structure breakdown
The paper opens with a brief framing of its dual purpose before dividing into two main analytical sections. The leadership section leads with an overview of the regression model, presents a coefficient table, discusses individual item-level predictors, and then breaks findings down by reporting group. The community policing section follows, covering study design, validity considerations, outcome results, an item-total correlation table, and KMO/BTS factor analysis justification. The conclusion is implicit within each section rather than a standalone segment.
Introduction to Cluster Analysis
Using two different datasets, this paper aims to improve upon two analyses related to leadership effectiveness and community policing studies. The paper examines reliability tests and factor analyses on several of the six scales used in the Community Policing study, and re-evaluates the analysis originally conducted by Dr. Eveland.
Leadership Effectiveness Analysis
From the report conducted by Dr. Eveland, several multiple regression analyses were initiated to help determine whether any relationship existed among the nine scales and the dependent variable, Q75: overall satisfaction with the manager as a leader. When plotted on a graph, the variable was moderately skewed toward higher values.
A simple multiple regression of all predictors showed a statistically significant relationship between predictors and the dependent variable (F = 19.33, df = 9, 462, p < 0.001, R² = 0.27). Nevertheless, beta coefficients for most individual predictors were not significant; only Communication emerged as a significant predictor, with a beta of 0.33. Due to varying means and variances among the predictors, a stepwise regression was used to correct the results (Eveland, 2008). This regression analysis produced a significant result with three predictors (F = 56.0, df = 3, 468, p < 0.001, R² = 0.26). The significant predictors and their coefficients are outlined below.
As the analysis above shows, communication is the most important of the three predictors. Its beta of 0.33 is approximately three times as large as Planning/Execution (b = 0.13) and Teamwork (b = 0.12). Stepwise regressions of all items from each scale on overall satisfaction (Q75) also identified the best individual item predictors.
For Communication, Q8 (genuinely listens to people) had a beta of 0.23; Q17 (shares information readily) peaked at 0.24; Q53 (encourages and accepts feedback from others) was 0.20; and Q61 (asks questions to ensure effective two-way communication) recorded 0.16.
For Planning/Execution, Q2 (spends quality time on planning) and Q63 (consistently follows through on his/her commitments) recorded betas of 0.15 and 0.63, respectively.
For Teamwork, Q5 (demonstrates a personal commitment to teamwork) was 0.19; Q14 (effectively leads teams to achieve business results) was 0.26; and Q50 (appreciates diverse viewpoints) was 0.25.
Predictors by Reporting Group
A final analysis repeated the stepwise regression separately for each of the four reporting groups. The analyses are appropriate given the significant differences in mean overall satisfaction scores across groups, as shown in the table below.
References
Eveland, J. D. (2008). Reliability analysis of Leadership Effectiveness Inventory data from Megasoft's "NORDIC" dataset.
Travis, J. (1998). Community policing in action: Lessons from an observational study. National Institute of Justice.
Always verify citation format against your institution’s current style guide requirements.