TY - JOUR
T1 - Sources of Response Bias in Older Ethnic Minorities
T2 - A Case of Korean American Elderly
AU - Kim, Miyong T.
AU - Lee, Ju Young
AU - Ko, Jisook
AU - Yoon, Hyunwoo
AU - Kim, Kim B.
AU - Jang, Yuri
N1 - Funding Information:
The first data set was collected from 675 KAE living in Florida in 2008; the study was funded by the National Institute of Mental Health (Jang, Chiriboga, Allen, Kwak, and Haley ; Study 1). This study was designed to help reduce mental health disparities among underserved ethnic minority including KAE living in Florida. The second data set was obtained from the Advanced Research Institute in Geriatric Mental Health in order to examine the physical and mental health status of 420 KAE in the New York City metropolitan area (Jang et al. ; Study 2). Data sets used as the third data set were collected with support from the Agency for Healthcare Research and Quality for a study conducted to develop and test an innovative self-help program aimed at improving the control of high blood pressure among KAE living in Maryland (Kim, Han, Kim, and Duong ; Study 3).
Publisher Copyright:
© 2015, Springer Science+Business Media New York.
PY - 2015/9/19
Y1 - 2015/9/19
N2 - The present study was undertaken to investigate potential sources of response bias in empirical research involving older ethnic minorities and to identify prudent strategies to reduce those biases, using Korean American elderly (KAE) as an example. Data were obtained from three independent studies of KAE (N = 1,297; age ≥60) in three states (Florida, New York, and Maryland) from 2000 to 2008. Two common measures, Pearlin’s Mastery Scale and the CES-D scale, were selected for a series of psychometric tests based on classical measurement theory. Survey items were analyzed in depth, using psychometric properties generated from both exploratory factor analysis and confirmatory factor analysis as well as correlational analysis. Two types of potential sources of bias were identified as the most significant contributors to increases in error variances for these psychological instruments. Error variances were most prominent when (1) items were not presented in a manner that was culturally or contextually congruent with respect to the target population and/or (2) the response anchors for items were mixed (e.g., positive vs. negative). The systemic patterns and magnitudes of the biases were also cross-validated for the three studies. The results demonstrate sources and impacts of measurement biases in studies of older ethnic minorities. The identified response biases highlight the need for re-evaluation of current measurement practices, which are based on traditional recommendations that response anchors should be mixed or that the original wording of instruments should be rigidly followed. Specifically, systematic guidelines for accommodating cultural and contextual backgrounds into instrument design are warranted.
AB - The present study was undertaken to investigate potential sources of response bias in empirical research involving older ethnic minorities and to identify prudent strategies to reduce those biases, using Korean American elderly (KAE) as an example. Data were obtained from three independent studies of KAE (N = 1,297; age ≥60) in three states (Florida, New York, and Maryland) from 2000 to 2008. Two common measures, Pearlin’s Mastery Scale and the CES-D scale, were selected for a series of psychometric tests based on classical measurement theory. Survey items were analyzed in depth, using psychometric properties generated from both exploratory factor analysis and confirmatory factor analysis as well as correlational analysis. Two types of potential sources of bias were identified as the most significant contributors to increases in error variances for these psychological instruments. Error variances were most prominent when (1) items were not presented in a manner that was culturally or contextually congruent with respect to the target population and/or (2) the response anchors for items were mixed (e.g., positive vs. negative). The systemic patterns and magnitudes of the biases were also cross-validated for the three studies. The results demonstrate sources and impacts of measurement biases in studies of older ethnic minorities. The identified response biases highlight the need for re-evaluation of current measurement practices, which are based on traditional recommendations that response anchors should be mixed or that the original wording of instruments should be rigidly followed. Specifically, systematic guidelines for accommodating cultural and contextual backgrounds into instrument design are warranted.
KW - Cultural bias
KW - Ethnic minority elderly
KW - Response bias
UR - http://www.scopus.com/inward/record.url?scp=84939468250&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84939468250&partnerID=8YFLogxK
U2 - 10.1007/s10823-015-9266-9
DO - 10.1007/s10823-015-9266-9
M3 - Article
AN - SCOPUS:84939468250
VL - 30
SP - 269
EP - 283
JO - Journal of Cross-Cultural Gerontology
JF - Journal of Cross-Cultural Gerontology
SN - 0169-3816
IS - 3
ER -