"The extent to which the wording of questions affects the answers obtained depends almost entirely on the degree to which the respondent's mental context is solidly structured." People who lack reliable and consistent frames of reference "are highly suggestible to the implication of phrases, statements, innuendoes, or symbols of any kind that may serve as clues to help them make up their minds." Questions which bluntly state some deviation from an established norm are less likely to receive favorable replies than questions which imply the same deviation but state it more by implication. Where a new and somewhat complicated problem is to be posed about which people have thought little, the free-answer type of question should be used. "The split-ballot technique should be used wherever possible in order to test the stability and consistency of opinion by noting the effect of . . . variation between free and prescribed responses." Responses to many single questions asked by the polls should be compared with responses to other questions which place the same issue in different contingencies. (PsycINFO Database Record (c) 2006 APA, all rights reserved)