1.9 Chapter summary

In empirical research, we try to answer research questions by measuring phenomena and constructs. Data analysis establishes the relationships between variables we have measured. Variables can be categorical or quantitative. We can either measure their pre-existing variation, which we do in observational studies, or manipulate them, which we do in experimental studies. Experimental manipulation can be done either within subjects or between subjects. We have to design our studies and our analysis strategy carefully to identify the estimand we are interested in. Confirmatory research is research with hypotheses that can be specified ahead of time. These hypotheses can be turn into predictions about what we will see in the data if the hypothesis is true. Hypotheses may or may not arise from theories. To do data analysis well enough, we need to: develop and pre-register an analysis strategy prior to the research; maintain a reproducible chain of evidence from data collection to final results; and archive our data and code openly.