Search results

The assumption that a set of observed variables is a function of an underlying common factor plus some error has dominated measurement in marketing and the social sciences in general for decades. This view of measurement comes with assumptions, which, however, are rarely discussed in research. In this article, we question the legitimacy of several of these assumptions, arguing that (1) the common factor model is rarely correct in the population, (2) the common factor does not correspond to the quantity the researcher intends to measure, and (3) the measurement error does not fully capture the uncertainty associated with measurement. Our discussions call for a fundamental rethinking of measurement in the social sciences. Adapting an uncertainty-centric approach to measurement, which has become the norm in in the physical sciences, offers a means to address the limitations of current measurement practice in marketing.

While industrial marketing managers have long been concerned about the representativeness of sample information, few direct measures have been available of how accurately the sample represents the population. Suggests the use of characteristics available from sources external to the survey process as a basis for such assessment. Presents procedures for gathering, analysing and interpreting such surrogate measures.

Purpose – Revisiting Fornell et al.'s (1996) seminal study, this chapter looks at the evidence for observed and unobserved heterogeneity within data underlying the American customer satisfaction index (ACSI) model. Examining data for two specific industries (utilities and hotels) reveals only modest differences. However, we suppose that unobserved heterogeneity critically affects the results. These insights provide the basis for shaping further differentiated ACSI model analyses and more precise interpretations.

Methodology/approach – This study applies the partial least squares (PLS) path modeling method and uses empirical data to estimate and compare the ACSI model results on the aggregate and industry-specific data levels. In addition, the finite mixture PLS path modeling (FIMIX-PLS) method is employed to further examine across industry similarities and within industry differences.

Findings – This research uncovers unobserved heterogeneity that guides forming three segments of customers within each industry. The major segment in each industry represents customers that are fairly loyal (i.e., neither disloyal nor extremely loyal) while the other two smaller segments are not as similar across the two industries. Our study identifies substantial differences across these segments within each industry. An importance-performance map analysis illustrates these differences and provides the basis for managerial implications.

Originality/value of the chapter – The unobserved heterogeneity revealed within industries in a given country (i.e., the United States of America) underlines the need to be open to differences within populations, beyond the observed heterogeneity across distinct groups or cultures, and the need to reconsider reporting requirements in academic research.

This paper aims to respond to John Rossiter's call for a “Marketing measurement revolution” in the current issue of EJM, as well as providing broader comment on Rossiter's C‐OAR‐SE framework, and measurement practice in marketing in general.

The paper is purely theoretical, based on interpretation of measurement theory.

The authors find that much of Rossiter's diagnosis of the problems facing measurement practice in marketing and social science is highly relevant. However, the authors find themselves opposed to the revolution advocated by Rossiter.

The paper presents a comment based on interpretation of measurement theory and observation of practices in marketing and social science. As such, the interpretation is itself open to disagreement.

There are implications for those outside academia who wish to use measures derived from academic work as well as to derive their own measures of key marketing and other social variables.

This paper is one of the few to explicitly respond to the C‐OAR‐SE framework proposed by Rossiter, and presents a number of points critical to good measurement theory and practice, which appear to remain underdeveloped in marketing and social science.

This paper aims to examine the impact of customer orientation, at the levels of both the company and the salesperson, on salesperson performance.

A survey of 283 salespeople provides the database that was analyzed using structural equation modeling.

Prior studies suggest that both company and salesperson customer orientation has a positive effect on performance. The findings of this study suggest that a salesperson's customer orientation completely mediates the relationship between company customer orientation and salesperson performance. Thus, the influence of a company's customer orientation on salesperson performance acts through the customer orientation of the salespeople.

The study reinforces the importance of customer orientation and the role of salespeople in putting customer orientation into practice.

Alongside structural equation modeling (SEM), the complementary technique of partial least squares (PLS) path modeling helps researchers understand relations among sets of observed variables. Like SEM, PLS began with an assumption of homogeneity – one population and one model – but has developed techniques for modeling data from heterogeneous populations, consistent with a marketing emphasis on segmentation. Heterogeneity can be expressed through interactions and nonlinear terms. Additionally, researchers can use multiple group analysis and latent class methods. This chapter reviews these techniques for modeling heterogeneous data in PLS, and illustrates key developments in finite mixture modeling in PLS using the SmartPLS 2.0 package.

“Garbage in, garbage out” is a common expression that academics and practitioners use to emphasize that empirical analysis is only as good as the basis on which it relies. Although the importance of sound data and valid measures has long been acknowledged, it is nevertheless often problematic to follow required quality standards in concrete research situations. Potential sources of error are usually unknown, methods to ensure data quality are unavailable, and existing methods for scale development, index construction, data collection, and data analysis are insufficient or erroneously applied. This is especially true of international marketing research, which often makes great demands on the data and measures used, as well as on the research methodology applied. Against this background, this volume addresses issues pertaining to measurement and research methodology in an international marketing context. Thanks to the efforts of authors and reviewers, we are pleased to present nine articles that deal with cutting-edge topics such as formative measurement, response-bias in cross-cultural research, marketing efficiency measurement, and segmentation methods.

The purpose of this paper is to investigate the possibility of curvilinear patterns of relationships between workplace affective commitment and in‐role performance, organizational citizenship behaviors and burnout. As most commitment theories assume strictly linear relations with these outcomes, demonstrating that these positive associations do not hold above some ceiling point in the commitment continuum is potentially important for research and practice.

The possibility of nonlinear relations was examined in a sample of 273 hospital employees.

The results yielded strong support for the authors' hypotheses. Indeed, most of the relations observed (ten of 15) between affective commitment foci and work outcomes were curvilinear, revealing a ceiling to the positive association between commitment and outcomes. Although these results vary in strength across work outcomes and commitment targets, they reveal that affective commitment has negative associations with employee productivity and psychological health at extreme levels.

Methodologically, these results illustrate the need to systematically explore the true nature of relations among constructs, even in areas where it is assumed to be well known. Practically, these results suggest that, ultimately, moderate levels of commitment may be more beneficial than extremely high levels.