|Micheli, Nilda - BAYLOR COLLEGE OF MED|
|Jaramillo, Sandra - BAYLOR COLLEGE OF MED|
|Hughes, Sheryl - BAYLOR COLLEGE OF MED|
Submitted to: Journal Of The American Dietetic Association
Publication Type: Abstract Only
Publication Acceptance Date: July 17, 2003
Publication Date: July 20, 2003
Citation: Micheli, N., Jaramillo, S., Hughes, S., Nicklas, T. 2003. Strategies used for enhancing the equivalence of questionnaire items across two cultures. Journal Of The American Dietetic Association. 103(9):A-67. Interpretive Summary: AN INTERPRETIVE SUMMARY IS NOT REQUIRED.
Technical Abstract: Child feeding measures have historically been developed, usually with Caucasians, and translated into different languages and/or field-tested with other ethnicities to discern cultural sensitivity. Three methods were used to develop a child feeding measure that characterized the specific feeding behaviors among low-income Hispanic (H) and African-American (AA) parents of preschool children: 1) observational videotapes; 2) de-centering techniques; and 3) cognitive interviews. From twelve observational videotapes, a variety of feeding style behaviors was selected to develop a preliminary measure of items indigenous to both ethnicities. Using de-centering techniques, the measure was translated into Spanish. Five cognitive interviews were conducted with each ethnicity to assess the meaning and sensitivity of the items. Interviews revealed that between the two ethnicities, some words: 1) appeared offensive; 2) were ambiguous; and 3) differed in meaning. For example, AA disliked the words 'threaten' and 'guilty', which evolved into 'warn' and 'feel bad', respectively. For H, 'clean' was changed to 'finish' in order to clarify the meaning of the item and subsequently translate properly into Spanish. The final questionnaire was administered to 231 caregivers (101 AA; 130 H). Exploratory factor analysis revealed three dimensions reminiscent of Baurnrind's parenting styles. The three-factor solution accounted for 23.5% authoritarian, 10.5%, authoritative, and 7% permissive feeding. Test-retest correlation after 7-14 days was .86, .71, and .56, respectively. When developing behavioral measures targeting minorities, a profound understanding of their language and culture is necessary. Extensive techniques are needed to identity items that can work equally well for both ethnicities.