Gordon G. Bechtel

Are Semantic Response Scales Equivalent

Three frequently used response formats -- Likert, semantic-differential and single anchor (Stapel) -- are compared via analysis of covariance structures. The cumulative results based on four data sets provided evidence inconsistent with previous research suggesting that these formats are interchangeable. Consistently across studies, the semantic-differential format is most preferred while in most cases the Stapel format is least preferred.

A Dual‐Source Indicator of Consumer Confidence

The objective of this paper is to improve the U.S. leading confidence indicator by combining its amended version with the Conference Boards equally well known Consumer Expectation Index. The resulting dual-source indicator provides a factorial structure in which survey items are nested. The results presented here show that this structure is not well fit by a generalized linear model. Hence a nonlinear model is invoked which provides differential logistic slopes for the survey items. This latter model is based on the work of McCullagh (1980). The effects estimated in this nonlinear structure provide the United States with a potential leading indicator that is richer and more informative than its current Index of Consumer Expectations. The suggested indicator prevents the mixed signals and inconsistencies that can occur when separately reporting results from our two most prominent sources of consumer perception.

The sociotropic aspect of consumer confidence

Abstract The unidimensionality of consumer confidence, a U.S. leading indicator, is investigated in the European Economic Community (EEC). The present findings confirm American and European evidence that the confidence indicators now in use are not unidimensional. They also point to a strong ‘sociotropic’ component of consumer confidence in each of the four largest nations of the EEC. The present results are provided by model-based survey measures which, unlike the status quo indicators, can be quality checked.

Measuring subjective social indicators

Abstract A logistic response model is proposed for measuring social indicators based upon survey rating scales. A close fit of the model to national life-quality data is demonstrated. Moreover, the model is constrained so as to provide an explicit formulation for item averaging with respect to group aggregates. This item redundancy provides measures for social constructs in a manner analogous to traditional procedures at the individual level of measurement. Further applications of this principle to opinion and mental health monitoring are discussed.

Combining perceptual and economic variables in the demand function

Abstract An item response theory is combined with a main-effect demand function for predicting product sales from perception and price. In the sales function perception is measured on a log interval scale, whereas price is measured on its usual ratio scale. The most important main effect in this function represents a composite of unspecified attributes and other sources of demand variation over competing products. The extraction of this effect serves to unbias the perceptual impact and price elasticity estimates by relieving the underspecification common to demand functions. This main effect also reveals product values that may be studied in subsequent investigations. The present approach of concatenating psychological and economic variables in the same demand equation is illustrated with an analysis of data from a high-tech consumer market.