Contributions published at Marketing and Market Research (René Algesheimer)

Large-scale surveys such as the Programme for International Student Assessment (PISA), the Teaching and Learning International Survey (TALIS), and the Programme for the International Assessment of Adult Competences (PIAAC) use advanced statistical models to estimate scores of latent traits from multiple observed responses. The comparison of such estimated scores across different groups of respondents is valid to the extent that the same set of estimated parameters holds in each group surveyed. This issue of invariance of parameter estimates is addressed in model fit indices which gauge the likelihood that one set of parameters can be used across all groups. Therefore, the problem of scale invariance across groups of respondents can typically be framed as the question of how well a single model fits the responses of all groups. However, the procedures used to evaluate the fit of these models pose a series of theoretical and practical problems. The most commonly applied procedures to establish invariance of cognitive and non-cognitive scales across countries in large-scale surveys are developed within the framework of confirmatory factor analysis and item response theory. The criteria that are commonly applied to evaluate the fit of such models, such as the decrement of the Comparative Fit Index in confirmatory factor analysis, work normally well in the comparison of a small number of countries or groups, but can perform poorly in large-scale surveys featuring a large number of countries. More specifically, the common criteria often result in the non-rejection of metric invariance; however, the step from metric invariance (i.e. identical factor loadings across countries) to scalar invariance (i.e. identical intercepts, in addition to identical factor loadings) appears to set overly restrictive standards for scalar invariance (i.e. identical intercepts). This report sets out to identify and apply novel procedures to evaluate model fit across a large number of groups, or novel scaling models that are more likely to pass common model fit criteria.

The aim of the study was to examine identity statuses in various life domains and the relationship between identity and well-being. We adopted the three-dimensional model of identity (Crocetti et al., 2008), including: in-depth exploration, commitment, and reconsideration of commitment. Moreover, in accordance with domain-specific approach (Goossens, 2001), we sought to empirically derive identity statuses in various life domains. The participants included 835 emerging adults (Mage = 21.81, SD = 2.33). We examined eight domains previously identified in qualitative research: personality characteristics, past experiences, family, friends and acquaintances, worldview, hobbies and interests, aims and plans for the future, and occupation. To measure three identity processes, we used a modified version of the Utrecht-Management of Identity Commitments Scale (Crocetti et al., 2008) and to measure well-being we used the Mental Health Continuum-Short Form (Keyes, 2013). Results indicate that, although the statuses identified in previous research were, to a large extent, replicated (except moratorium), people were classified in different statuses in different domains; thus, we conclude that talking about statuses should be limited to a given domain. Well-being was the highest in achievement statuses and the lowest in diffusion, but only in two examined domains: personality characteristics and past experience.

Large-scale surveys such as the Programme for International Student Assessment (PISA), the Teaching and Learning International Survey (TALIS), and the Programme for the International Assessment of Adult Competences (PIAAC) use advanced statistical models to estimate scores of latent traits from multiple observed responses. The comparison of such estimated scores across different groups of respondents is valid to the extent that the same set of estimated parameters holds in each group surveyed. This issue of invariance of parameter estimates is addressed in model fit indices which gauge the likelihood that one set of parameters can be used across all groups. Therefore, the problem of scale invariance across groups of respondents can typically be framed as the question of how well a single model fits the responses of all groups. However, the procedures used to evaluate the fit of these models pose a series of theoretical and practical problems. The most commonly applied procedures to establish invariance of cognitive and non-cognitive scales across countries in large-scale surveys are developed within the framework of confirmatory factor analysis and item response theory. The criteria that are commonly applied to evaluate the fit of such models, such as the decrement of the Comparative Fit Index in confirmatory factor analysis, work normally well in the comparison of a small number of countries or groups, but can perform poorly in large-scale surveys featuring a large number of countries. More specifically, the common criteria often result in the non-rejection of metric invariance; however, the step from metric invariance (i.e. identical factor loadings across countries) to scalar invariance (i.e. identical intercepts, in addition to identical factor loadings) appears to set overly restrictive standards for scalar invariance (i.e. identical intercepts). This report sets out to identify and apply novel procedures to evaluate model fit across a large number of groups, or novel scaling models that are more likely to pass common model fit criteria.

Comparisons of means or associations between theoretical constructs of interest in cross-national comparative research assume measurement invariance, that is, that the same constructs are measured in the same way across the various nations under study. While it is intuitive, this assumption needs to be statistically tested. An increasing number of sociological and social psychological studies have been published in the last decade in which the cross-national comparability of various scales such as human values, national identity, attitudes toward democracy, or religiosity, to name but a few, were tested. Many of these studies did not manage to fully achieve measurement invariance. In this study we review, in a nontechnical manner, the methodological literature on measurement invariance testing. We explain what it is, how to test for it, and what to do when measurement invariance across countries is not given in the data. Several approaches have been recently proposed in the literature on how to deal with measurement noninvariance. We illustrate one of these approaches with a large dataset of seven rounds from the European Social Survey (2002–2015) by estimating the most trustworthy means of human values, even when strict measurement invariance is not given in the data. We conclude with a summary and some critical remarks.

null

null

We identified behavioral signatures of the values distinguished in the Schwartz et al. refined value theory (2012). We examined behavioral signatures for two types of values, value states and value traits. We conducted two studies using innovative approaches. Study 1 used retrospective self-reports whereas Study 2 used self-reports in real time. In Study 1 (N = 703), we sought act frequency signatures of the 19 basic value traits that the Portrait Value Questionnaire-Revised (Schwartz, 2017) measures. We examined the frequency of 209 acts from the Oregon Avocational Interest Scales (Goldberg, 2010) for which there were no expectations that values would necessarily influence them. We computed partial correlations between each behavioral act and each value. We discuss the theoretical links to each value of the 10 behavioral acts that correlated most highly with it. Study 2 analyzed 9,416 behavioral acts of 374 participants. We measured value expressions in current behavior, i.e., value states, using experience sampling methodology (ESM). We asked participants 7 times per day for 7 days what they had been doing during the past 15 minutes and how important 9 different values from the Schwartz’s refined value theory were to them during that activity. Because the questions about activities were open-ended, the set of behavioral acts analyzed in Study 2 was theoretically unlimited. To find signatures of values in behavior, we identified the activities during which participants reported the highest level of importance for each value. Both studies revealed meaningful associations between values and daily behavior.

Recent research reports the need for consistent incentives in blockchain-based systems. In this study, we investigate how incentives for a blockchain-based inter-organizational network should be designed to ensure a high quality of data, exchanged and stored within the network. For this, we use two complementary methodological approaches: an Action Design Research approach in combination with agent-based modelling, and demonstrate, through the example of a real-world blockchain project, how such an incentive system may be modelled. The proposed incentive system features a rating mechanism influenced by measures of data correction. We evaluate the incentive system in a simulation to show how effective the system is in terms of sustaining a high quality of data. Thus, the paper contributes to our understanding of incentives in inter- organizational settings and, more broadly, to our understanding of incentive mechanisms in blockchain economy.

A pivotal idea in network science, marketing research, and innovation diffusion theories is that a small group of nodes—called influencers—have the largest impact on social contagion and epidemic processes in networks. Despite the long-standing interest in the influencers identification problem in socioeconomic and biological networks, there is not yet agreement on which is the best identification strategy. State-of-the-art strategies are typically based either on heuristic centrality measures or on analytic arguments that only hold for specific network topologies or peculiar dynamical regimes. Here, we leverage the recently introduced random-walk effective distance—a topological metric that estimates almost perfectly the arrival time of diffusive spreading processes on networks—to introduce a centrality metric which quantifies how close a node is to the other nodes. We show that the new centrality metric significantly outperforms state-of-the-art metrics in detecting the influencers for global contagion processes. Our findings reveal the essential role of the network effective distance for the influencers identification and lead us closer to the optimal solution of the problem.

null

Measurement invariance is a necessary precondition for meaningful cross-country comparisons, and three levels have been differentiated: configural, metric, and scalar. Unfortunately, establishing the most stringent form, i.e., scalar measurement invariance, across groups is difficult. Recently, Muthén and Asparouhov proposed testing for approximate rather than exact measurement invariance as this may be sufficient for meaningful comparisons. Following their strategy, the results of cross-country approximate measurement invariance tests of the PVQ-21 scale to measure values in the European Social Survey (ESS) are presented (N = 274,447 respondents from 15 countries participating in all six rounds). Applying the new approximate method for the test of measurement invariance allows both using more moderate constraints of approximate equality of parameters across groups and exploring the extent of noninvariance. Approximate measurement invariance was established in almost all rounds for two higher-order values: openness to change and self-enhancement. In the case of the two other higher-order values, self-transcendence and conservation, approximate measurement invariance was established across a subset of countries.

Real networks typically studied in various research fields—ecology and economic complexity, for example—often exhibit a nested topology, which means that the neighborhoods of high-degree nodes tend to include the neighborhoods of low-degree nodes. Focusing on nested networks, we study the problem of link prediction in complex networks, which aims at identifying likely candidates for missing links. We find that a new method that takes network nestedness into account outperforms well-established link-prediction methods not only when the input networks are sufficiently nested, but also for networks where the nested structure is imperfect. Our study paves the way to search for optimal methods for link prediction in nested networks, which might be beneficial for World Trade and ecological network analysis

Nestedness refers to the structural property of complex networks that the neighborhood of a given node is a subset of the neighborhoods of better-connected nodes. Following the seminal work by Patterson and Atmar (1986), ecologists have been long interested in revealing the configuration of maximal nestedness of spatial and interaction matrices of ecological communities. In ecology, the BINMATNEST genetic algorithm can be considered as the state-of-the-art approach for this task. On the other hand, the fitness-complexity ranking algorithm has been recently introduced in the economic complexity literature with the original goal to rank countries and products in World Trade export networks. Here, by bringing together quantitative methods from ecology and economic complexity, we show that the fitness-complexity algorithm is highly effective in the nestedness maximization task. More specifically, it generates matrices that are more nested than the optimal ones by BINMATNEST for 61.27% of the analyzed mutualistic networks. Our findings on ecological and World Trade data suggest that beyond its applications in economic complexity, the fitness-complexity algorithm has the potential to become a standard tool in nestedness analysis.

null

null

null

Valuing customers is key to any firm. Customer lifetime value (CLV) is the central metric for valuing customers. It describes the long-term economic value of customers and gives managers an idea of how customers will evolve over time. To model CLVs in continuous non-contractual business settings such as retailers, probabilistic customer attrition models are the preferred choice in literature and practice. Our R package CLVTools provides an efficient and easy to use implementation frameworks for probabilistic customer attrition models. Building up on the learnings of other implementations, we adopt S4 classes to allow constructing rich and rather complex models that nevertheless still are easy to apply for the end user. In addition, the package includes recent model extensions, such as the option to consider contextual factors, that are not available in other packages. This article will focus on both, the theory of the underlying statistical framework as well as about the practical application using real world data.