Overlapping category ranges can mislabel subjects and distort study conclusions

Evidence

I realize that the classification problems go back to the original studies. Some authors would call weekly attendance “moderate,” and others “high.” The reviewer is stuck with their published categories. But the problem still remains. Another thing bothers me about these ranges: they are highly unequal. Usually when you break up a continuum of responses into a few categories, you follow meaningful breaks in the criterion or else try to get the sample evenly distributed in the categories. — Bob Altemeyer, The Authoritarian Specter, 4178 (#expands, via linked claim)

Categorization

• XREF: connects to: Unequal category ranges distort published research results — Bob Altemeyer, The Authoritarian Specter, 4178; Equal-percentile cutoffs produce deeply unequal measured categories — Gary Goertz, Social Science Concepts and Measurement, 899; Negating a middle category yields a comparison group too heterogeneous to interpret — Gary Goertz, Social Science Concepts and Measurement, 1164 (+3 more)

Broader Connections

• Unequal category ranges distort published research results #expands (0.92) — Bob Altemeyer, The Authoritarian Specter, 4178