Refine
Year of publication
Document Type
- Doctoral thesis (70)
Institute
Language
- English (70) (remove)
Keywords
- Logistik (5)
- Management (5)
- Corporate Social Responsibility (4)
- Einzelhandel (4)
- Supply Chain Management (4)
- Kundenmanagement (3)
- Nachhaltigkeit (3)
- Operations management (3)
- Prognoseverfahren (3)
- Retail operations (3)
Abstract Paper 1:
Cybervetting has become common practice in personnel decision-making processes of organizations. While it represents a quick and inexpensive way of obtaining additional information on employees and applicants, it gives rise to a variety of legal and ethical concerns. To limit companies’ access to personal information, a right to be forgotten has been introduced by the European jurisprudence. By discussing the notion of forgetting from the perspective of French hermeneutic philosopher Paul Ricoeur, the present article demonstrates that both, companies and employees, would be harmed if access to online information on applicants and current employees would be denied. Consistent with a Humanistic Management approach that promotes human dignity and flourishing in the workplace, this article proposes guidance for the responsible handling of unpleasant online memories in personnel decision-making processes, thereby following Ricoeur’s notion of forgetting as “kept in reserve”. Enabling applicants and employees to take a qualified stand on their past is more beneficial to both sides than a right to be forgotten that is questionable in several respects.
Abstract Paper 2:
When hiring new employees, managers often conduct online background checks on applicants. This practice has led many applicants to reach out to legal and technological solutions that render unpleasant online memories forgotten. With this article, we question the effectiveness and social value of approaches aimed at permanently erasing information from the web. Furthermore, we provide practitioners with guidance toward responsible handling of applicant information from the web. In this conceptual article, we review the literature on organizational memory and interpret it through a philosophical lens by turning to the works of the French philosopher Paul Ricoeur and the French sociologist Maurice Halbwachs. We conclude that legally or technologically enforced forgetting on the web fails to provide true protection from a memory in a hiring situation, as remembering in an organization is a complex social process that hardly lets a memory disappear completely. As a result, legal and technological approaches that aim at erasing unpleasant memories from the web represent only an illusion of protection for job applicants. Alternatively, we propose a selection process that makes the erasure of unpleasant memories from the web unnecessary. Previous literature on organizational memory has already criticized a purely mechanistic view of memory in organizations, by which memories are treated as objects that can be retrieved and deleted on demand. To the best of our knowledge, this is the first article that establishes a connection between organizational memory and the use of online information in selection.
Abstract Paper 3:
Companies have started using social media for screening applicants in the selection process. Thereby, they enter a low-cost source of information on applicants, which potentially allows them to hire the right person on the job and avoid irresponsible employee behavior and negligent hiring lawsuits. However, a number of ethical issues are associated with this practice, which give rise to the question of the fairness of social media screening. This article aims to provide an assessment of the procedural justice of social media screening and to articulate recommendations for a fairer use of social media in the selection process. To achieve this, a systematic literature review of research articles pertaining to social media screening has been conducted. Thereby, the benefits and ethical issues relating to social media screening, as well as recommendations for its use have been extracted and discussed against Leventhal’s (1980) rules of procedural justice. It turns out that without clear guidelines for recruiters, social media screening cannot be considered procedurally fair, as it opens up way too many opportunities for infringements on privacy, unfair discrimination, and adverse selection based on inaccurate information. However, it is possible to enhance the fairness of this practice by establishing clear policies and procedures to standardize the process.
Abstract Paper 4:
Social media platforms have presented individuals with sheer endless possibilities for networking, creating and exchanging information. Yet, overwhelming social demands, privacy concerns, and even a lack of access to new technologies have led many users to discontinue social media usage. Simultaneously, companies screen and select applicants on social media. As previous research suggests that employers may view missing social media information with suspicion, we ask if non-users of social media face disadvantages in hiring: Are they discriminated against? In this study, we employ a 2 x 2 experimental survey design to verify whether absence from social media may result in discriminatory behavior towards job applicants. We conduct the study with two samples: The first comprises business students, and the second consists of more senior members of the workforce. Although this study does not confirm discrimination against non-users of social media in the selection process, it adds to the literature in two respects: Firstly, it shows differences in call-back between highly and less qualified candidates as a result of social media information. Secondly, it suggests unintentional, systemic rater biases. Against these findings, we formulate recommendations for applicants and employers and give recommendations for future research in this field.
Specifying or deriving customers' needs into detailed requirements becomes essential for holistic customer-centricity in automotive systems engineering, relying heavily on simulation models. However, with the increase of system complexity, customer diversity, and difficulty of customer data acquisition, it becomes challenging to specify model inputs that represent individual customer usage behavior across the whole product lifecycle. To let these challenges be tackled, this dissertation addresses the problem of customer usage profiling to support decision-making in the context of automotive systems engineering. First, two fundamental data engineering procedures are investigated, including data aggregation for feature reduction and sampling for representing the customer fleets with a few selected reference customers. After data preprocessing, a method for preparing the model inputs from aggregate customer data, i.e., usage profiling, is developed. Usage profiling applies meta-heuristics, synthesizing sufficiently representative from aggregate fleet data. Furthermore, a decision support system is developed to deploy the aggregation, sampling, and usage profiling into model-based automotive system engineering processes. As both data and various models are connected, digital twinning is applied. Using real-world fleet data, an evaluation case study for determining lifetime requirements indicates that the methodology is plausible and capable of consolidating customer-centricity into automotive systems engineering processes.
In this dissertation, I investigate how innate decision-making tools, such as culture and curiosity, affect individual economic education and innovation decisions. The cumulative dissertation consists of the following four chapters:
Chapter 1: Culture in Individual Economic Decision-Making—A Conceptual Contribution
Chapter 2: Individualism and the Formation of Human Capital (based on joint work with Sven Resnjanskij, Jens Ruhose, and Simon Wiederhold)
Chapter 3: Individualism, Creativity, and Innovation
Chapter 4: Play Stupid Games, Win Stupid Prizes? Casino-Based Evidence on Exploration Strategies and Curiosity in a High-Complexity Game (based on joint work with Alexander Patt)
Through a combination of methods (theoretical as well as structural and reduced-form empirical work) and data sources (ranging from large-scale international data to experimentally generated data), each chapter answers a distinct but related research question:
While the first chapter contributes to our conceptual understanding of culture in an economic context, chapters two and three are dedicated to investigating economic effects of individualism. Individualism is a polarizing dimension of culture with its emphasis on personal achievement and personal responsibility. Although individualism is an established driver of long-run economic growth, the underlying individual-level mechanisms have not been well understood. To close this research gap, I study the role of individualism in individual education and innovation decisions. Using data from an international skill assessment, the second chapter shows that individualism strongly and positively affects human capital investments. Individualism matters more for human capital formation than any other dimension of culture found in the economic literature and should, thus, be included more rigorously in the education production function. Complementarily, the third chapter reveals that individualism is positively related to individual innovativeness – which is captured by a wide range of measures from selection into creative-innovative occupations and knowledge diffusion on the job to creativity traits and patenting.
Focusing on the exploration-exploitation trade-off that is prevalent in innovation decisions, the fourth chapter uses slot machine games to investigate exploration behavior in a natural yet controlled setting. The experimental results document that a significant group of players engage in curiosity-based exploration related to the game’s entropy. The different player types identified in the data showcase the variety of strategies economic agents use to navigate a high-uncertainty high-complexity decision environment.
To embrace diversity in thoughts and approaches economically, an understanding of the specific role of cultural tools in economic decision making is imperative.
In this thesis, I evaluate different tax policy reforms or tax policies in general with respect to the three pillars of sustainability: the environmental, the economic, and the social pillar. The cumulative dissertation consists of the following four chapters:
Chapter 1: Towards Green Driving - Income Tax Incentives for Plug-in Hybrids (based on joint work with Henning Giese)
Chapter 2: Corporate Taxation and Firm Performance (based on joint work with Dominika Langenmayr, Valeria Merlo, and Georg Wamser)
Chapter 3: Tax Avoidance Using Hybrid Financial Instruments Among European Countries
Chapter 4: Avoiding Taxes: Banks’ Use of Internal Debt (based on joint work with Franz Reiter and Dominika Langenmayr)
The first chapter relates to the environmental dimension of sustainability. More specifically, I evaluate a specific German tax reform that aimed at reducing greenhouse gas emissions in the transportation sector – a key driver of emissions – by fostering green driving. This essay empirically investigates whether a preferential tax treatment for hybrid plug-in company cars was effective in fostering green driving and whether it was cost-efficient.
The second chapter relates to the economic dimension of sustainability. A subgoal towards more sustainability in this dimension is decent economic growth. Chapter 2 of this thesis contributes to the question of how tax policy can foster firm performance which is an important driver for economic growth. This essay empirically investigates the relationship between corporate taxation and firm performance taking into account firm heterogeneity. As tax policy is an essential instrument for reaching political goals like boosting the economy in crises, this essay may provide valuable insights for policymakers.
The third and fourth chapters relate to the social dimension of sustainability. A subgoal towards more sustainability in this dimension is increased international cooperation. With respect to taxation, the global challenge of multinational firms engaging in tax avoidance made international cooperation in tax policy essential in the last decade. Chapter 3 investigates a specific channel of tax avoidance, i.e., tax avoidance with hybrid financial instruments. This legal analysis reveals tax avoidance opportunities among selected European countries and evaluates countermeasures.
Chapter 4 investigates a different channel of tax avoidance, i.e., intragroup debt finance, specifically for the banking sector. I use administrative data to empirically analyse whether and to what extent multinational banks engage in tax avoidance.
"Present Bias in Choices over Food and Money: Evidence from a Framed Field Experiment" - this paper presents results from a field experiment investigating dynamically inconsistent time preferences over real food choices and commitment take-up in a natural environment. We implement a longitudinal consumption experiment with a continuous convex budget: College students repeatedly choose and consume lunch menus at their college canteen. We not only allow for a full continuum of healthy or unhealthy foods; we also explicitly design the consumption stage to comply with the consume-on-receipt assumption: utility is created right after handing out the food. The paper also focuses on the implications for effective policy design: It sheds light on consumers' tendency to utilize different types of self-control devices to commit to personal consumption plans. We further compare inconsistent behavior between convex food and money choices to investigate the applicability of monetary reward studies to natural behavior. We find no correlation between time parameters in food vs. money tasks. Utility estimates from food choices suggest that dynamically consistent individuals substitute internal self-control with commitment take-up. Non-committing individuals tend to be present-biased and naive about their inconsistency.
"Dynamic Inconsistencies and Food Waste: Assessing Food Waste from a Behavioral Economics Perspective" - this paper investigates the link between dynamically inconsistent time preferences and individual food consumption and waste behavior. Food waste is conceptualized as unintended consequence of consumption choices along the food consumption chain: Because inconsistent individuals postpone the consumption of healthier food at home, storage time is prolonged and food more likely to be wasted. Capitalizing on a rich data set from a nationally representative survey, the paper constructs targeted measures of time-related food consumption and waste behaviors. In line with the theory, the paper finds that more present-biased individuals waste more food. The mechanism can be empirically supported: Although dynamically inconsistent individuals plan their food consumption, they deviate from their consumption plans more compared to consistent individuals. Deviation behavior is significantly associated with food waste. The paper points to the importance of considering dynamic inconsistencies at different stages of the food consumption process to foster the intended effects of food policies (increase in healthy nutrition) and diminish unintended consequences (increase in food waste).
"Economic Behavior under Containment: How do People Respond to Covid-19 Restrictions?" - this paper investigates the effects of policy interventions on individual decision-making and preferences by looking at a period of unprecedentedly strict regulation: the Covid-19 pandemic. Focusing on Germany, we measure regulatory strictness by calculating a policy stringency index at the federal state level. Capitalizing on exogenous variation in predetermined election cycles across states, we instrument policy stringency with the distance to the next election. Results support a positive relation between weeks to next election and state policy stringency. Leveraging a nationally representative short panel, the paper exploits this exogenous variation in policy stringency across states and over time to assess behavioral changes in different economic domains. The paper finds that government restrictions increase working from home and the provision of childcare at home, while it reduces the number of grocery shopping trips. Containment policies also affect risk preferences and fear of Covid-19: individuals become more risk-averse and afraid of Covid-19 as a consequence of stricter policies. The results are robust against various different model specifications.
Exposure to more than one language influences children's language development (Kehoe & Kannathasan, 2021). It has been shown that particularly the relative amount of language input bilingual children receive influences their language development. Bilingual children can never dedicate the same amount of time to either one of their languages as their monolingual peers because their everyday linguistic environment is divided between two languages. This may ultimately lead to typically developing bilingual children with linguistic profiles similar to monolingual children with a language developmental disorder (DLD) if they have reduced contact to one of their languages (Paradis, 2010). For this reason, it is particularly difficult for clinicians (i.e., Speech-Language Pathologists, SLPs) to distinguish between actual language impairment and typical deviations from the (monolingual) norm due to limited language exposure. In Germany, almost half of all children in speech therapy have a bilingual or multilingual language background (Lüke & Ritterfeld, 2011). Despite the high proportion of bilingual children in speech therapy, the assessment and treatment of DLD in this population is still predominantly monoculturally and monolingually oriented (Scharff Rethfeldt, 2016). Previous studies have identified nonword repetition tasks (NWRTs) as promising tools for the clinical differentiation between bilingual typically developing children and children with DLD because they are arguably relatively free from the influence of children's prior language experience (Schwob et al., 2021). However, some researchers have raised the question of whether bilingual children might still be disadvantaged by language-specific characteristics of nonwords (e.g., Chiat, 2015).
In three separate studies, this dissertation will (1) demonstrate the influence of bilingual language exposure on speech sound processing, an essential building block of language development, (2) evaluate how DLD is diagnosed in bilingual children in Germany and which factors contribute to positive attitudes towards bilingualism in SLPs, and (3) investigate whether bilingual children's nonword repetition performance of language-specific versus unspecific (i.e., more language-neutral) nonwords is subject to the influence of their individual language exposure.
In the first study (see chapter 2), the perception and production of Voice Onset Time (VOT) in bilingual Italian-German, monolingual German, and monolingual Italian children was investigated using a combination of neurophysiological and behavioral methods. Forty five-year-old children (16 monolingual German, 24 bilingual Italian-German) participated in the electroencephalographic (EEG) study, in which automatic speech sound processing of Italian- and German-specific VOT contrasts was investigated. To compare the participants language-specific VOT production, in addition to the 16 monolingual German and 24 bilingual Italian-German children, a control group of five monolingual Italian-speaking children were tested with a picture naming task in which the production of language-specific word-initial VOT was recorded based on ten words per target sound. Acoustic measurements were taken in Praat (Boersma & Weenink, 2013). Children's language background was assessed using a parental language background questionnaire. In the second study (see Chapter 3), 66 SLPs from Germany participated in an online survey investigating their attitudes towards bilingualism and their approaches in their daily clinical practice. Finally, in the third study (see Chapter 4), a new NWRT was developed that included Italian- and German-specific as well as language-unspecific nonwords. The NWRT was implemented in an online platform and administered to a group of 26 bilingual Italian-German TD children. Children's repetition performance in the different nonword subcategories was analyzed in the context of their linguistic background (recorded using the same questionnaire as in Study 1).
Study 1 showed that Italian-German bilingual five-year-old children and the monolingual German control group behaved similarly in processing language-specific VOT stimuli. However, differences between the two groups likely related to language exposure were evident in the processing of a German "long lag" stimulus. Surprisingly, the bilingual group had no advantage in processing the Italian "voicing lead" stimuli. This suggests that the bilingual five-year-olds do not (yet) process the Italian contrasts automatically (Yu et al. 2019). Acoustic measures of productive VOT for Italian and German words were consistent with this interpretation. Specifically, bilingual children showed mastery of the German but variability with respect to the Italian VOT contrast. Evidence for crosslinguistic influence (cf. Paradis & Genesee, 1996) was observed. Study 2 showed that there seems to be an imbalance between SLPs' knowledge of the specific requirements for providing their services to bilingual children and their actual application in clinical practice, indicating sufficient awareness but still a relatively high presence of monolingual approaches in the context of speech therapy. Presumably, this is due to a lack of available materials and/or resources. Finally, Study 3 showed that Italian-German bilingual children performed equally well on Italian-specific, German-specific, and language-unspecific nonwords, regardless of their individual language exposure and language dominance patterns, supporting the view that even language-specific NWRTs can be used to clinically assess the language performance of bilingual children with relatively little exposure to either of their languages. However, this result has yet to be replicated and thus confirmed with a group of children with DLD.
In summary, these three studies highlight the gap between scientific research and clinical practice and demonstrate the relevance of clinicians' participation in research projects to (1) improve knowledge transfer between the two fields and (2) promote evidence-based practices in the context of Speech-Language Pathology.
In children exposed to any traumatic experience, a conditional prevalence rate of 15.9% was estimated for posttraumatic stress disorder (PTSD; Alisic et al., 2014). In order to achieve best treatment outcomes, selecting evidence-based treatments is necessary (Durlak & DuPre, 2008). For pediatric PTSD, international guidelines recommend the use of manualized trauma-focused cognitive behavioral therapy (TF-CBT; Forbes, Bisson, Monson, & Berliner, 2020; National Institute for Health and Care Excellence, 2018; Phoenix Australia Centre for Posttraumatic Mental Health, 2013) and many effectiveness studies and initiatives (e.g. Goldbeck, Muche, Sachser, Tutus, & Rosner, 2016; Jensen et al., 2014; National Child Traumatic Stress Network, 2012) disseminated the TF-CBT manual of Cohen, Mannarino and Deblinger (2006, 2017; specific TF-CBT). However, the uptake of evidence-based treatments such as specific TF-CBT into practice is strongly affected by service providers’ attitudes toward evidence-based practice (EBP; Aarons, 2004; Aarons, Cafri, Lugo, & Sawitzky, 2012; Aarons, Sommerfeld, & Walrath-Greene, 2009; Moullin, Dickson, Stadnick, Rabin, & Aarons, 2019). These attitudes can be measured with the Evidence-Based Practice Attitudes Scale-36 (EBPAS-36; Rye, Torres, Friborg, Skre, & Aarons, 2017). In Germany, several effectiveness trials are underway further disseminating specific TF-CBT (Rosner, Barke, et al., 2020; Rosner, Sachser, et al., 2020) but a reliable and validated German version of the EBPAS-36 is missing. In addition, in terms of pooled effect sizes (ES), only one review evaluated specific TF-CBT 10 years ago (Cary & McMillen, 2012). Furthermore, little is known about the stability of its treatment effects, the concordance between children and adolescents’ and their caregivers’ ratings regarding the youths’ symptoms in this context as well as the influence of the heterogeneity of instruments used by individual studies.
Therefore, the aims of this dissertation were fourfold: Publication 1 pooled ES for specific TF-CBT regarding posttraumatic stress symptoms (PTSS) and secondary outcomes to provide an update of the literature. Large improvements were found across all outcomes and results were favorable for specific TF-CBT compared to all control conditions at post-treatment. Publication 2 pooled ES for specific TF-CBT at 12-month follow-up for the same outcomes to evaluate the stability of treatment effects and investigated concordance between participants and caregiver ratings. Again, large improvements were evident across all outcomes and TF-CBT was superior to active treatments and treatment as usual at 12-month follow-up. No significant differences were detected between participant and caregiver ratings with high reliability across almost all outcomes and assessment points. Publication 3 produced a validated German translation of the EBPAS-36 (EBPAS-36D) to facilitate EBP implementation research in Germany. Good psychometric properties and validity of the EBPAS-36D were confirmed. As additional content, this dissertation provides an unpublished report on the pooled ES of publication 1 differentiated by instruments used to assess the outcomes. ES did not differ by the instrument used to assess the respective outcome.
The conclusions that can be drawn from the studies as well as their limitations are discussed in the final section. While the EBPAS-36D was successfully validated and specific TF-CBT was strongly supported, several issues were identified that need further research attention.
Background:
Pollen allergies are widespread around the world, making the study of pollen levels in and around cities a task of high importance. This is also related to the fact that an increasing number of people is living in cities and urban areas are expanding. In addition, environmental changes in the context of climate change are expected to alter plant characteristics that affect the amount of pollen in the air. In order to assess the allergy risk, information on pollen that people are exposed to is needed. Monitoring of aeroallergens in cities is often conducted using a single trap. This trap can provide information on background pollen concentration, however, it does not reflect temporal and spatial variations of pollen at street level of the heterogeneous urban environment. The amount of pollen in the air is the result of a number of processes, one of which is pollen production. Pollen production is expected to be altered by climate change and is influenced by environmental factors such as temperature or air pollutants. To further investigate variations of allergenic pollen in cities, this thesis focuses on these central research questions:
- Are there differences in the amount of airborne pollen between urban and rural locations and do pollen loads vary at different spatial and temporal scales?
- What is the influence of land use and management on pollen loads?
- Is there a difference in pollen production of allergenic species between urban and rural locations and can the influence of environmental factors be detected?
Data and Methods:
Data on airborne pollen were gathered in two study areas with various pollen traps. In Ingolstadt, Germany, background pollen concentrations of Poaceae were measured in 2019 and 2020. Additionally, a network consisting of twelve gravimetric pollen traps was set up during the grass pollen seasons of those years. For an additional sampling campaign, portable volumetric traps were used to measure pollen concentration at street level. Investigations were also conducted in Sydney, Australia, where concentrations of Poaceae, Myrtaceae and Cupressaceae pollen and Alternaria spores were measured during three pollen seasons from 2017 to 2020. In the summer of 2019/2020, additional ten gravimetric traps were set up. For both study areas, pollen season characteristics and temporal and spatial variations of pollen concentrations and pollen deposition were analysed in context with land use, land management and meteorological parameters.
Furthermore, pollen production of the allergenic species Betula pendula, Plantago lanceolata and Dactylis glomerata was investigated in the Ingolstadt study area along an urbanisation gradient, which covered different types of land use. In total, 24 individuals of B. pendula, 82 individuals of P. lanceolata and 54 individuals of D. glomerata were analysed. Pollen production was compared between urban and rural locations and the influence of air temperature and the air pollutants nitrogen dioxide (NO2) and ozone (O3) was assessed. Air temperature was measured in the field and pollutant concentrations were computed using a land use regression model.
Results and Discussion:
In Ingolstadt, grass pollen concentrations and deposition were generally higher at rural than at urban locations as documented in peak values and seasonal totals. Diurnal variations however, did not vary between urban and rural locations. Grass pollen levels were linked to local vegetation and land use but also land management, as grass cutting was reducing pollen levels in the surroundings.
In Sydney, pollen season characteristics varied between the seasons and we found significant correlations between daily pollen and spore concentrations and meteorological parameters. There were significant positive correlations with air temperature for all analysed pollen and spore types. Correlations with humidity (Myrtaceae, Cupressaceae, Alternaria) and precipitation (Myrtaceae, Cupressaceae) were negative. Spatial variations of pollen deposition were observed, but there were no correlations with land use. This could be attributable to the drought before and during this sampling campaign as well as the temporal setting of the sampling campaign.
The results from both study areas suggest that local vegetation is the main influence on the pollen amount at street level. However, when the season’s intensity decreased, other influences such as resuspension or pollen transport become more important. Analyses of pollen production revealed statistically significant differences between urban and rural locations for B. pendula and P. lanceolata. Pollen production decreased with temperature and urbanisation for all species, however, pollen production varied at small spatial distances within species. For increasing air pollutant levels, decreases were observed for B. pendula and P. lanceolata, but increases for D. glomerata. These results suggest that environmental influences seem to be species-specific.
The results from the studies contribute to the knowledge on temporal and spatial variations of airborne pollen in cities, as well as on pollen production of allergenic species. They highlight the importance of pollen monitoring at street level and at different locations in cities to assess the amount of pollen that people are exposed to. Further research in this area should investigate the influence of land management, which was identified to influence the amount of pollen in the air, and which could be a mitigation strategy for allergy-affected people. In addition, pollen production should be further examined with different kinds of experiments, as it is the major influence of ambient pollen concentration in the air.
This cumulative dissertation investigates the newly emerged phenomenon of customer success management (CSM). CSM has received a significant amount of attention from academics and practitioners alike. However, the body of academic literature is still in its infancy, especially when comparing CSM’s relevance to the field of relationship marketing.
Manuscript 1: Zakrzewski, K., Krause, V., Pfaff, C. & Seidenstricker, S. (2023). Is customer success management a new relationship marketing practice? A review, definition, and future research agenda based on practice theory.
Manuscript 2: Zakrzewski, K., Krause, V., Eberl, D. & Rangarajan, D. (2023). Dynamic & proactive segmentation of post-sale customer relationships in B2B subscription settings: The customer health score.
Manuscript 3: Zakrzewski, K. (2023). Customer success management - closing a capability gap in the customer-focused structure toward customer centricity in B2B service contexts.
The first manuscript adopts a new approach to the investigation of CSM, providing a practice-based definition that distinguishes CSM from other relationship marketing constructs. By utilizing the 3 Ps framework of practice theory (praxis, practices, and practitioners), the definition of CSM as a strategic function was refined. In addition, the second manuscript utilized praxis to develop the customer health score, which can be used as an indicator for the dynamic and proactive segmentation of customers. The third manuscript builds on this functional definition to show how CSM’s organizational embeddedness can help overcome barriers to customer centricity by enabling adaptive organizational capability.
In forecast combination, multiple predictions are linearly combined through the assignment of weights to individual forecast models or forecasters. Various approaches exist for defining the weights, which typically involve determining the number of models to be used for combination (selection), choosing an appropriate weighting function for the forecast scenario (weighting), and using regularization techniques to adjust the calculated weights (shrinkage). The papers listed address the integration of the three approaches into holistic data analytical models.
The first paper develops a two-stage model in which weights are first calculated based on the in-sample error covariances to minimize the error on the available data by combining the individual models. Based on the selection status of a model, the weight of an individual model is then linearly shrunk either toward the mean or toward zero. The selection status is thereby derived apriori from information criteria, where Contribution 1 introduces the selection based on the model’s in-sample accuracy and performance robustness under uncertainty.
Contribution 2 modifies the two-stage model to shrink the forecasters’ weights non-proportionally to the mean or zero. Further, a new information criterion based on forward feature selection is proposed that iteratively selects the forecaster that is expected to achieve the largest increase in accuracy when combined.
Contribution 3 extends the iteration-based information criterion presented in the second contribution to consider diversity gains in addition to accuracy gains when selecting and combining forecasters.
A one-stage model for simultaneous weighting, shrinking, and selection is finally built and evaluated on simulated data in Contribution 4. Instead of requiring a prior selection criterion, the model itself learns which forecasters to shrink to the mean or to zero, while relying on a new sampling procedure to tune the model.