Extreme and acquiescence bias in a bi-ethnic population.

Background: Extreme and acquiescence biases are the tendency to give a positive or extreme answer regardless of the 'true' answer. These biases may compromise comparisons of attitudes regarding health between population groups. The aim of the study was to measure the extent of extreme and acquiescence biases and identify factors associated with them in two ethnic groups: Jews and Arabs in Israel.

Methods: A random telephone survey was conducted during 2006, interviewing 2322 Jews and 809 Arabs. Three attitude questions were presented twice with opposite wording to measure extreme and acquiescence biases in these two groups.

Results: Extreme bias ranged from 2 to 14% among Jews and from 6 to 29% among Arabs, depending on the question. Acquiescence bias ranged from 2 to 10% among Jews and 5-19% among Arabs. The less educated respondents gave more extreme biased responses for all items. The older respondents gave more extreme answers for two out of the three questions tested. After adjusting for age and education the odds ratio (OR) of giving more extreme biased answers was higher among Arabs compared with Jews for all three questions [OR = 2.49, confidence interval (CI) = 1.87, 3.31; OR = 2.33, CI = 1.75, 3.10; and OR = 2.94, CI = 1.83-4.71, respectively, for each question].

Conclusions: Levels of response biases are higher in the Arab minority population compared with the majority Jewish population and depended on the subject, age and education.