Geddes, Barbara. "How the Cases You Choose Affect the Answers You Get: Selection Bias in Comparative Politics". Political Analysis, Vol.2 (1990): 131-150.

Geddes, Barbara. "How the Cases You Choose Affect the Answers You Get: Selection Bias in Comparative Politics". Political Analysis, Vol.2 (1990): 131-150.

• Selecting cases based on the outcomes of the measured dependent variable, essentially selecting cases which support the connection being hypothesized, is common in political science, including among studies performed by experienced and respected scholars (131-132).
• Failing to properly select countries leads to major issues in comparative politics, namely by ignoring potential factors or situations. If two countries are selected based on their high growth rates and shown to both possess the same factor, but not compared to other countries with this factor, then nothing has actually been proven (132-133).
• There could be a bias where factor x is present in all countries, but because only the cases with high-growth rates are measured and they share factor x, the fact that factor x is not correlated to high-growth is totally ignored (133).
• For example, a number of scholars have claimed that repression of labor was necessary for high economic growth in East Asia, Brazil, and Mexico. However, by only looking at countries with high-growth, and ignoring low-growth countries with labor repression, these scholars have made a claim without any evidence (134-139).
• Making these kinds of selection bias are not only misleading, but also greatly increases the likelihood of getting the wrong answer during the investigation, because some many alternative factors are excluded (149).