# Data analysis for management decision making MN2196

This course develops a comprehensive framework to evaluate whether the predictions of managerial, social and economic theories are supported by empirical evidence.

This course is an introduction to the quantitative techniques guiding evidence-based managerial decision-making. For instance, are family-managed firms less productive than other firms? Are there biases in job interviews? Are political connections important for corporations? Does performance-pay increase productivity? This course provides the tools to investigate empirically whether these statements are supported by empirical evidence.

### Prerequisites

If you are studying for this course as part of a BSc degree, you must have already passed:

• ST105A Statistics 1 (half course) and
• MT105A Mathematics 1 (half course)

### Topics covered

Main topics of the module include:

• The Linear Regression Model: the causality problem, the population model, sampling processes, estimators and estimates, the ordinary least squares (OLS) model, assumptions and properties of the OLS model. Limitations of linear regression: omitted variable bias, non-random sampling, measurement error, outlying observations, heteroskedasticity.
• Multiple Regression: omitted variable bias problem, the multiple regression model, assumptions and properties of the multiple regression model.
• Inference: estimators as random variables, hypothesis testing (assumptions and methods, t-test and F-test), p-values, confidence intervals.
• Functional Form: dummy variables, conversion of discrete variables into sets of dummy variables, quadratic models, models with interactions, use of dummy variables to explore functional form.
• Exploiting Time Variation: types of samples (cross-sectional, time series, repeated cross-sections and panel data), first-difference models, individual fixed effects models, time fixed effects models, differences-in-differences models.
• Instrumental Variables: instrumental variables as a method to alleviate omitted variable bias, reduced form estimates, two-stage least squares estimates.

### Learning outcomes

If you complete the course successfully, you should be able to:

• Understand the possibilities and challenges in using quantitative data to learn about the world.
• Critically evaluate whether statements and theories are supported by sufficient and/or adequate empirical evidence
• Use the R econometric package to manipulate data, run regressions and interpret the resulting output
• Propose and evaluate simple research designs to untangle causal relations between variables
• Understand how to translate theoretical relations into testable hypotheses.

### Assessment

Unseen written exam (3 hrs).