Author
BARANOWSKI, TOM - Children'S Nutrition Research Center (CNRC) |
Submitted to: The Games for Health Journal: Research, Development, and Clinical Applications
Publication Type: Other Publication Acceptance Date: 5/7/2014 Publication Date: 8/1/2014 Citation: Baranowski, T. 2014. Measurement method bias in games for health research. The Games for Health Journal: Research, Development, and Clinical Applications. 3(4):193-194. Interpretive Summary: Technical Abstract: Studies of games for health often use self report measures which are subject to many known and unknown sources of error. When two self report measures are used in the same study, they tend to be moderately to highly correlated. This is not because there is a true moderate to high correlation, but because they share common sources of error, thereby providing misleading findings. In such circumstances, games researchers should use two measures of the same thing but from different sources (e.g. self report and sensor). In this case, the errors will not be correlated, leaving a truer estimate of the relationship of interest. |