This project investigates whether online training programs for managers improve their relationships with employees. Using a randomized controlled trial (RCT), the study assesses the causal impact of managerial training on employee-manager relationship quality and explores how this effect varies by gender (heterogeneity), is influenced by conflict resolution (mediation), and differs across job roles (moderation).
- Estimate the average treatment effect (ATE) of managerial training on employee perceptions.
- Assess gender-based heterogeneity in training effectiveness.
- Examine mediation effects via conflict resolution ability.
- Explore moderation effects based on job role (front vs. back office).
- Design: Randomized Controlled Trial (RCT)
- Data Source: Skilled Managers – Productive Workplaces (SMPW)
- Sample: 24 UK-based organizations across public and private sectors
- Approach: Mixed methods (quantitative + qualitative), Difference-in-Differences (DiD), OLS regression
- Tools: R (with mediation package), DAGs for causal inference
- No significant improvement in manager-employee relationship quality post-training.
- Slight positive effects in males, but slight negative impact in females.
- Conflict resolution skills (q7_s) showed a small mediating effect, though not statistically significant.
- Predictor
q8_s
(perceived improvement after talking with the manager) was the strongest influence on relationship quality.
📦 project-directory/
├── total_effect.R # Average Treatment Effect (ATE)
├── heterogeneity_dropped_na.R # Heterogeneity with dropped NAs
├── heterogeneity_imputed.R # Heterogeneity with imputation
├── mediation.R # Mediation analysis
├── moderation.R # Moderation analysis
├── requirements.txt # List of required R packages
├── output/ # (Optional) Folder for any output files (plots, tables)
└── README.md # This file
- Clone the repository:
git clone https://github.com/tahirkorma/skilled-manager-intervention-analysis.git cd skilled-manager-intervention-analysis
- Install required packages in R:
install.packages(readLines("requirements.txt"))
- Run each script independently in R or RStudio:
source("total_effect.R") source("heterogeneity_dropped_na.R") source("heterogeneity_imputed.R") source("mediation.R") source("moderation.R")
📝 License This research is conducted as part of an academic dissertation and is shared for educational purposes only.
🙏 Acknowledgements Skilled Managers – Productive Workplaces (SMPW)
Supervisors and faculty at Durham University
Bowyer & Urwin (2024) for RCT design and data