Citation

Granberg-Rademacker JS. Educ. Psychol. Meas. 2010; 70(1): 74-90.

Copyright

(Copyright © 2010, SAGE Publishing)

DOI

10.1177/0013164409344532

PMID

unavailable

Abstract

The extensive use of survey instruments in the social sciences has long created debate and concern about validity of outcomes, especially among instruments that gather ordinal-level data. Ordinal-level survey measurement of concepts that could be measured at the interval or ratio level produce errors because respondents are forced to truncate or round off their responses to fit a given ordinal scale. This article presents a Markov chain Monte Carlo modeling technique that converts ordinal measurements to interval/ ratio. Simulated data demonstrate the robustness of this technique, and implications of this technique are discussed.