Abstract
Previous research has found that the use of no opinion filters not only increases the proportion of “don't know” responses but they may also affect substantive response distributions. The latter effect was assumed to occur because the filter removed persons (floaters) from the substantive response distributions who had attitudes different from nonfloaters. This study serendipitously finds that DK filters may affect the substantive responses of nonfloaters, causing them to express different opinions on the filtered question form than they would on the standard form. It also appears that the filter lowers the attitude strength of the nonfloaters. An explanation is offered for these unanticipated findings and an agenda for future research is suggested.
Get full access to this article
View all access options for this article.