Spread the Word . . .
-
Tag Archives: ratings
When Questionnaire Length Threatens Data Quality
When Questionnaire Length Threatens Data Quality
BACKGROUND: When surveys go on too long, respondents can get tired of answering them. They may quit part-way through or just become lazy and speed through without thinking about what is being asked. We discard data from both “drop-outs” and “streamliners,” respectively, but this adds time and money to data collection because many more quality respondents will be needed to make up the difference. The longer the questionnaire, the more drop-outs and streamliners (a.k.a. “straightliners”) result.
RESEARCH METHOD: We analyzed data collected from three questionnaires to measure the incidence of dropping-out and streamlining. Drop-out rate was defined as the percentage of respondents who didn’t complete the questionnaire. Streamline rate was defined as the percentage of respondents who engaged in any of the following:
Research ON Research: Rating 2 Brands at Once
A Time Saver or A Waste of Data?
Background:
There may be time saved by asking a respondent to move through an attribute battery in an online survey j
ust once and rate two brands – rather than rate a single brand, item-by-item, only to go through the battery again to rate a second brand. However, could this slightly more time consuming method yield richer data?
Research Method:
Using standard security screening and a demographically representative online panel, 1,250 respondents, at least partly responsible for household shopping decisions, answered questions about sports drinks. About equal numbers of respondents rated a pair of brands one at a time (n=580) or both at
once (n=586) on 28 attributes. For each attribute, the average gap between ratings when brands were rated one at a time was tested for a significant difference against when brands were rated both at once.
