American Sociological Association

Alternately Contested: A Measurement Analysis of Alternately Worded Items in the National Science Foundation Science Literacy Scale

Alternately worded versions of two controversial indicators of science knowledge were included in the 2012 wave of the General Social Survey. Using confirmatory factor analysis, the author tests whether these alternate items serve as better indicators of uncontested forms of science knowledge and finds that although more respondents give the “correct” response, they remain poor indicators of uncontested science knowledge. Moreover, they retain the confounding influence of religious factors present in the original items, making them no better (in terms of measures of science knowledge unbiased by controversial issues) than the original items. Evidence supports the claim that both sets of items measure the same construct, but with different difficulties.


J. Micah Roos





Starting Page


Ending Page