Technical Notes

“Chances Are” Question Technique

This research combines the “Unmatched Count Technique” (UCT) with a straightforward “has this happened to you?” self-report approach. UCT involves showing respondents a statement inventory and asking how many statements are true for them, but not which specific statements. This protects their anonymity whilst allowing us to measure the incidence rate of taboo topics. Comparing the UCT incidence rate to the self-report incidence rate is a measure of a given behaviour’s “embarrassment factor”. Statistical analysis can also be used to examine how these incidence rates vary across the population. For more information see Coutts, E., & Jann, B. (2011). Sensitive Questions in Online Surveys: Experimental Results for the Randomized Response Technique (RRT) and the Unmatched Count Technique (UCT). Sociological Methods & Research, 40(1), 169-193.

“Emotalizer” Database

All emotions stem from feeling negative or positive and passive or adrenalised. But the same “core affect” can be experienced as different emotions. For example, disgust and irritation are different, but they both entail feeling negative and slightly adrenalized. The English language contains 3,000 emotion words, of which 200 are in common use. Because it’s impossible to poll people across all 200 emotions, we’ve developed a proprietary approach for coding respondents’ free text into one of these common emotions (and its associated core affect). Measuring emotions can provide fresh insights into a topic compared to traditional polling. For more information see Russell, J. A., & Barrett, L. F. (1999). Core Affect, Prototypical Emotional Episodes, and Other Things Called Emotion: Dissecting the Elephant. Journal of Personality and Social Psychology, 76(5), 805-819.

“The Worth” Question Format

Willingness-to-pay is a useful common currency for comparing the strength of people’s preferences and extends traditional ‘x% of people agree’ research by measuring the intensity of that agreement. People, however, are better at reacting to a proposed price than producing their own estimate. Accordingly we use a Choice Based Valuation (CBV) approach, which asks people their likelihood of “buying” a given outcome for a given price, but varies the price between respondents. By fitting an ordered probit model to the resulting data, we can estimate the distribution of people’s valuations across the population, including within given segments. For more see Monroe, K. B., & Lee, A. Y. (1999). Remembering versus knowing