Square but straight: Measurement tool design to improve response task fluency and certainty

The assessment of overall customer satisfaction is an important issue in market research (Haumann et al. 2014, Homburg et al. 2014). After each online purchase, customers are asked to assess the product or service for which they have paid, usually on a five-point rating scales (e.g., Amazon, Trip Advisor). Comparable to bipolar scales, these tools are effective in making a distinction between polarized evaluations (i.e., either strong positivity or strong negativity). However, the literature on methodology reveals serious problems related to the mid-point displayed on these continuums (Kaplan 1972, Thompson et al. 1995). Actually, this mid-point inappropriately aggregates uncertain responses (difficult evaluation) with ambivalent (a combination of moderate to high positivity and negativity) or indifferent (low positivity and negativity) ones, when these different responses have been shown to reflect different attitudes and drive distinct behavioral responses (Yoo 2010, Thornton 2011). The Evaluative Space Grid (hereafter, ESG), developed in psychology by Larsen and colleagues (2009), could help address part of this methodological issue. The ESG comprises a 5×5 grid that measures both the degree of positivity of a stimulus and its degree of negativity within a bidimensional matrix. One dimension of the matrix is dedicated to the measurement of the respondent's degree of negativity ("not at all negative" - "extremely negative"), its degree of positivity ("not at all positive" - "extremely positive"). The combination of the two dimensions allows the respondent to choose which of the grid's 25 cells best describes its evaluation. The ESG has been validated in psychology with respect to unipolar measurement of positivity and negativity (Larsen et al. 2009). Applying the ESG in a service marketing context, Audrezet et al., (forthcoming) have recently demonstrated its relevance to measure overall customer satisfaction. Regarding the practical implementation of ESG, Larsen and colleagues (2009) formulate 2-minute generic instructions before using the ESG, which include the presence of an experimenter to accompany the respondents in their task. However, they do not discuss the extent to which the grid could be easy to use in self-administered surveys. Based on a rapid qualitative study we assume that the ESG could hinder response task fluency, which could negatively affect response certainty (Regier et al. 2014). This, in turn, could potentially be very detrimental to market research as response certainty is a crucial determinant of data and prediction quality (Fazio and Zanna 1978, Antil 1983). The present article builds on previous research to investigate the influence of different formats of the ESG on response task fluency and certainty. To do so, an experiment specifically manipulating the ESG dimension and the presence of verbal labels in the cells was conducted on a sample of 105 undergraduate students. We demonstrate that the use of verbal labels, rather than a reduction in response alternatives, is a promising way to increase response task fluency and, in turn, improve individuals' response certainty. This work advocates for tool design reflection to create responding behavior incentives and reduce survey drop-out rates which is especially challenging within self-administered electronic settings.