Barry Adam E, Chaney Beth, Piazza-Gardner Anna K, Chavarria Enmanuel A
1University of Florida, Gainesville, FL, USA.
Health Educ Behav. 2014 Feb;41(1):12-8. doi: 10.1177/1090198113483139. Epub 2013 Apr 3.
Health education and behavior researchers and practitioners often develop, adapt, or adopt surveys/scales to quantify and measure cognitive, behavioral, emotional, and psychosocial characteristics. To ensure the integrity of data collected from these scales, it is vital that psychometric properties (i.e., validity and reliability) be assessed. The purpose of this investigation was to (a) determine the frequency with which published articles appearing in health education and behavior journals report the psychometric properties of the scales/subscales employed and (b) outline the methods used to determine the reliability and validity of the scores produced. The results reported herein are based on a final sample of 967 published articles, spanning seven prominent health education and behavior journals between 2007 and 2010. Of the 967 articles examined, an exceedingly high percentage failed to report any validity (ranging from 40% to 93%) or reliability (ranging from 35% to 80%) statistics in their articles. For health education/behavior practitioners and researchers to maximize the utility and applicability of their findings, they must evaluate the psychometric properties of the instrument employed, a practice that is currently underrepresented in the literature. By not ensuring the instruments employed in a given study were able to produce accurate and consistent scores, researchers cannot be certain they actually measured the behaviors and/or constructs reported.
健康教育与行为领域的研究人员及从业者常常会开发、改编或采用调查问卷/量表，以量化和测量认知、行为、情感及社会心理特征。为确保从这些量表收集到的数据的完整性，评估心理测量学特性（即效度和信度）至关重要。本调查的目的是：（a）确定发表在健康教育与行为领域期刊上的文章报告所使用量表/子量表心理测量学特性的频率；（b）概述用于确定所产生分数的信度和效度的方法。本文所报告的结果基于967篇已发表文章的最终样本，这些文章来自2007年至2010年间七份著名的健康教育与行为领域期刊。在所审查的967篇文章中，极高比例的文章未在文中报告任何效度（范围从40%至93%）或信度（范围从35%至80%）统计数据。对于健康教育/行为领域的从业者和研究人员而言，若要最大限度地提高其研究结果的效用和适用性，就必须评估所使用工具的心理测量学特性，而目前这一做法在文献中未得到充分体现。由于未确保特定研究中所使用的工具能够产生准确且一致的分数，研究人员无法确定他们实际测量的就是所报告的行为和/或结构。
