Search results

The purpose of this chapter is to provide researchers a summary of some of the latest developments in item response theory (IRT), and to help these groups realize that psychometric tools can now be used for theory testing in addition to the traditional role of improving construct measurement. The author first reviews some of the fundamental tenets of classical test theory to contrast with IRT. He then describes recent advances in goodness-of-fit tests that have helped turn IRT into a model-testing tool. Finally, the author reviews several new test models that provide new flexibilities, summarizing summarize several examples of research that has used these new models in organizational research. At the end of this review, the author provides suggestions to help researchers better use these new IRT tools. Although there have been significant advances in IRT in the past decade, there has not been a systematic review of these developments. This review places those developments in context to provide readers a real appreciation of these breakthroughs.

The present chapter addresses a topic that is of growing interest – namely, the exploration of alternative item response theory (IRT) models for noncognitive assessment. Previous research in the assessment of trait emotional intelligence (or “trait emotional self-efficacy”) has been limited to traditional psychometric techniques (e.g., classical test theory) under the notion of a dominance response processes describing the relationship between individuals' latent characteristics and individuals' response selection. The present study, presents the first unfolding IRT modeling effort in the general field of emotional intelligence (EI). We applied the Generalized Graded Unfolding Model (GGUM) in order to evaluate the response process and the item properties on the short form of the trait emotional intelligence questionnaire (TEIQue-SF). A sample of 866 participants completed the English version of the TEIQue-SF. Results suggests that the GGUM has an adequate fit to the data. Furthermore, inspection of the test information and standard error functions revealed that the TEIQue-SF is accurate for low and middle scores on the construct; however several items had low discrimination parameters. Implications for the benefits of unfolding models in the assessment of trait EI are discussed.

Purpose – The purpose is to present item response theory (IRT) as a promising approach to deal with theoretical and methodological challenges which may not be adequately addressed with linear factor analysis (LFA) techniques.

Approach – We address this limitations and present IRT's approach to counteracting these difficulties. We further present two IRT models in greater detail and make a theoretical comparison between these models and LFA techniques. Then, we illustrate how IRT is applied in practice by analyzing a scale of trust by an IRT and an LFA approach and comparing their results.

Practical implications – Scale properties may vary among countries because some items may function differently in different contexts. Moreover, a scale's reliability depends on the value of the latent variable being assessed, which implies that the scale may discriminate better among certain ranges of the latent trait than among others. The theory we hereby present offers a more adequate approach to dealing with these issues than extant LFA methods.

Originality/value – This chapter presents a methodology of recognized value in other fields of knowledge into the field of strategy and international business, thereby advancing the methodologies available for doing research in these domains.

The multidimensional structure of boredom poses unique measurement challenges related to scale length and statistical modeling. We systematically address these concerns in two studies. In Study 1, we use item response theory to shorten the 29-item Multidimensional State Boredom Scale (MSBS) (Fahlman et al., 2013). In Study 2, we use structural equation modeling to compare two theoretically consistent multidimensional structures of boredom (superordinate and multivariate) with the most commonly used, yet theoretically inconsistent, structure in boredom research (unidimensional parallel model). Our findings provide support for modeling boredom as multidimensional and demonstrate the impact of model selection on effect sizes and significance.

Measures are comparable if and only if measurement equivalence has been demonstrated. Although comparability and equivalence of measures are sometimes used interchangeably, we advocate a subtle but important difference in meaning. Comparability implies that measures from one group can be compared with measures from another group. It is a property of the measures, which is given or not. In particular, comparability presumes valid measures within each group compared. Measurement equivalence, by contrast, refers to the way measures are derived and estimated. It is intrinsically tied to the underlying theory of measurement. Thus, measurement equivalence cannot be dealt with in isolation. Its assessment has to be incorporated into the theoretical framework of measurement. Measurement equivalence is closely connected to construct validity for it refers to the way manifest indicators are related to the latent variable, within a particular culture and across different cultures. From this it follows that equivalence cannot, or should not, be treated as a separate issue but as a constitutive element of validity. A discussion of measurement equivalence without addressing validity would be incomplete.

For many of the current leadership theories, models, and approaches, the answer to the question posed in the title, “Is leadership more than ‘I like my boss’?,” is “no,” as there appears to be a hierarchy of leadership concepts with Liking of the leader as the primary dimension or general factor foundation. There are then secondary dimensions or specific sub-factors of liking of Relationship Leadership and Task Leadership; and subsequently, tertiary dimensions or actual sub-sub-factors that comprise the numerous leadership views as well as their operationalizations (e.g., via surveys). There are, however, some leadership views that go beyond simply liking of the leader and liking of relationship leadership and task leadership. For these, which involve explicit levels of analysis formulations, often beyond the leader, or are multi-level in nature, the answer to the title question is “yes.” We clarify and discuss these various “no” and “yes” leadership views and implications of our work for future research and personnel and human resources management practice.

Economic globalization is leading large companies to focus on international strategic management. Nowadays, the assets referred to as “corporate intangibles,” such as corporate reputation, are becoming increasingly important because they are considered a key factor for the viability of an organization, and companies therefore need to incorporate them into their scorecards for management. The problem is that their measurement is subjective and latent. These two characteristics impede direct international comparison and require demonstrating the accuracy of comparison via a minimum of two tests – construct equivalence and metric equivalence. As regards corporate reputation, construct equivalence was verified by Naomi Gardberg (2006). However, the subsequent studies did not address metric equivalence. Based on the results of a survey provided by the Reputation Institute (n = 5,950, 50 firms evaluated in 17 countries in the Americas, Europe, Asia and Australia), the degree of RepTrak metric equivalence has been tested, using two different methodologies, multigroup analysis (structural equation model), and a new technique from 2016, the Measurement Invariance of Composite Model procedure from the Partial Least Square Path Modeling family. As one would expect from other cross-cultural studies, reputation metrics do not meet the full metric equivalence, which is why they require standardization processes to ensure international comparability. Both methodologies have identified the same correction parameters, which have allowed validation of the mean and variance of response style by country.