Contributions published at Marketing and Market Research (René Algesheimer)

Recent years have witnessed both a growing number of cross-cultural datasets but also a growing awareness of problems connected with cross-cultural comparisons. In particular, researchers have realized that measurement invariance is a necessary precondition for a meaningful comparison of data across cultures or countries. The literature on measurement invariance is very rich and provides researchers with a variety of approaches which suggest when and how measurement invariance may be tested. This chapter provides an explanation of what measurement invariance is and offers a guide designed to help researchers interested in testing for measurement invariance. We distinguish six main issues that have to be addressed by researchers while testing for measurement invariance. The first issue concerns the level of measurement invariance to be tested (configural, metric or scalar), the second – the type of data used (continuous or ordinal-categorical), the third – choice of rules to evaluate whether measurement invariance is established, the fourth – decision about whether cross-loadings of the items are to be permitted, the fifth – the scope of measurement invariance (full or partial), and the sixth – the accuracy of invariance that needs to be established (i.e. exact or approximate). We describe possibilities to address these issues and formulate recommendations for applied researchers.

This longitudinal study explores the stability and change of values in childhood. Children's values were measured in Poland three times (with one-year intervals) using the Picture Based Values Survey (PBVS-C; Döring, Blauensteiner, Aryus, Drögekamp, & Bilsky, 2010), developed to measure values differentiated according to the circular model of Schwartz (1992). 801 children (divided into 5 cohorts aged 7, 8, 9, 10, and 11 years at the first measurement occasion) completed the PBVS-C three times on a yearly basis. Separate analyses were performed for each cohort using the data of the three measurement occasions. Multidimensional scaling revealed that, in children, Schwartz's (1992) circular structure of values is stable and does not change over time. Although priorities of values displayed moderate stability over time, the means changed between the ages of 7 and 11 years. Specifically, latent growth curve modeling revealed changes in children's values hierarchy as indicated by the decrease in the mean level of conservation values and the increase in the mean level of openness to change values. Self-transcendence and self-enhancement also changed in different directions. As indicated by mean levels over time, self-transcendence first increased in importance, slightly decreased, and finally increased again. In contrast, self-enhancement first decreased in importance, then increased, and finally began to decrease again.

null

null

null

null

null

null

null

null

null

null

This thesis analyzes the influence of intra- and inter-country dynamics on the takeoff of new products. So far, prior studies have only analyzed takeoff on the product-level. This thesis focuses on the firm-level takeoff as a more precise measure for predicting product success across countries. A multi-level hazard model will be used to analyze the dataset, which con- sists of 436 broadband Internet operators in 87 countries. The results reveal that it is crucial for firms to correctly time their market entry in relation to the country-level takeoff. Moreo- ver, influences of inter- and intra-country variables vary across income clusters. Whereas the number of competitors within a country negatively affect firms in high and lower middle in- come groups, it has a positive impact on upper-middle income groups. Cross-country spillo- ver effects of neighboring firm-level takeoffs have a negative influence across all income clusters and only vary in the intensity. Empirical results will be discussed and implications for managers derived.

Search aggregators like Kayak, Orbitz and Expedia add value to consumers by providing a comprehensive list of available options (e.g. all airline routes that serve a given origin-destination pair along with their schedules and price). Consumers often use aggregator websites to find their best match but then book directly on the individual website, for which the aggregator receives no payment. In the absence of an aggregator, a consumer would have to search multiple airline websites and a firm might not get as many transactions. We aim to quantify the value that search aggregators give consumers and individual websites by utilizing a unique feature of the airline industry – Southwest airlines is not part of any aggregator. Consumers with Southwest in their consideration set have to make an extra search to get a comprehensive list: this feature enables us to quantify the value-add that aggregators give consumers as well as individual airline companies.

Valuing customers is key to the success of any firm. It enables marketers to identify and target key customers. Customer lifetime value (CLV) is a 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 a non-contractual business setting, probabilistic customer attrition models such as the Pareto/NBD model are the preferred choice in literature and practice. Their ability to simultaneously forecast both, customer’s actual lifetime and future transactions is unique. However, empirical evidence suggests that standard probabilistic customer attrition models do not outperform basic management heuristics. The question why this is the case has not been answered. A possible explanation is that standard probabilistic customer attrition models neglect seasonality, even though most transactional datasets exhibit such a pattern. Time-varying covariates would allow controlling for seasonality in probabilistic customer attrition models. Recently, such an extension has been presented for discrete non- contractual business settings. However, an implementation of time-varying covariates for the continuous non-contractual business setting is not available. We propose an approach to model time-varying covariates within the Pareto/NBD model. Then, we apply our model as well as the standard implementation of the Pareto/NBD model to four real-world datasets and evaluate their aggregate- and individual-level performance. Our findings have strong implications for both, marketing practice and research. Besides giving detailed recommendations on when to use which modeling approach, we also provide practical advices for estimating these models.

null

null

null

null