As long as surveys are constructed by human beings and administered to human beings, we will never totally eliminate order or wording effects. You have to put your questions in a certain order and use particular words to communicate meaning. These are bound to influence respondents. Nonetheless, we can minimize the impact of context factors by taking precautions when designing the survey. More generally, there are many things researchers can do to improve the quality of attitude questions (Sudman & Bradburn, 1982). Next time you are asked to develop a self-report survey, you might consider these suggestions:

1. Use words that all respondents can comprehend.