This course provides an introduction to causal thinking, which is the required thinking for successful completion of causal effect studies. You will learn about the data structures that are required in causal effect analyses and also about types of data and problems which make a causal effect analysis unfeasible.
You will be introduced to one of the most widely used methods for estimating causal effects – the propensity score methods. The methods are frequently used in social sciences and medical research when estimating causal effects and conditional associations, and in impact evaluation studies. The methods are design-based and consist of tools and techniques enabling one to produce a statistically balanced study design in comparative research, i.e., a study design consisting of two or more statistically comparable groups.
This is a two-day course where you are introduced to: (i) how to think causally in statistical data analysis; (ii) the kinds of questions that can be answered with causal effect analysis; (iii) how to perform a causal effect analysis; (iv) how to use propensity score methods in practice; and (v) to what kind of data examples the methods can be applied.
On the first day, we talk about different types of data, their structures and their reliability. We talk about big data versus small data and discuss data fields in different scientific disciplines: education, economics and medicine, while discovering the art of causal thinking. The causal inference theory is introduced.
On the second day, we cover the propensity score methods’ theory and apply the methods to some data set in order to answer a research question. We cover the required steps to perform a propensity score study. The R statistical program is used.
Who should attend
The workshop is suitable for: (i) those who deal with experiments, quasi-experiments, or non-compliance issues when analysing data; (ii) those who deal with impact evaluation studies, evidence based studies, association studies or aiming to answer causal questions quantitatively; and (ii) anyone else who deals with statistical data analyses and would like to expand their statistical knowledge and statistical thinking skills to more advanced levels.
As a prerequisite and to benefit the most out of the workshop, a basic knowledge of statistics, i.e., a completed undergraduate course of statistics and the knowledge of multivariate analysis, should be possessed.