Handling Missing Values in Information Systems Research: A Review of Methods and Assumptions

2023 | Information Systems Research | Citations: 0

Authors: Peng, Jiaxu; Hahn, Jungpil; Huang, Ke-Wei

Abstract: In today’s big data environment, missing values continue to be a problem that ha ... Expand

Abstract: In today’s big data environment, missing values continue to be a problem that harms data quality. The bias caused by missing values raises the highest concern, as it cannot be eliminated simply by increasing the sample size. Although the statistics literature has developed approaches to handling missing values and formulated assumptions regarding when these approaches generate valid statistical inferences, these prescriptions have yet to be broadly accepted by many social science disciplines, including the information systems (IS) discipline. By reviewing recently published empirical research in information systems, we find that missing values are indeed an important and pervasive problem. We believe that a review of missing value theory is necessary and timely for the IS community to understand the nature of missing values and to promote more rigorous research practice when missing values are often unavoidable. In addition, the missing not at random (MNAR) mechanism brings about challenges in parameter estimation. We contribute to research practice by proposing and demonstrating the superior performance of a Monte Carlo likelihood approach in correcting bias in parameter estimation. We conclude by suggesting that research validity can be enhanced through a reasoned adoption of the missing value handling method and structured missing value reporting practices.History: Giri Kumar Tayi, Senior Editor; Pei-Yu Chen, Associate Editor.Funding: This work was supported by the Singapore Ministry of Education [Grants R-253-000-136-114, R-253-000-149-490, and R-253-000-158-114] the National Natural Science Foundation of China [Approval No. 72201288], and the Program for Innovation Research in Central University of Finance and Economics. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not reflect the views of the Singapore Ministry of Education, the Singapore Government or the Central University of Finance and Economics.Supplemental Material: The online appendices are available at https://doi.org/10.1287/isre.2022.1104. Collapse

Topics: missing data research and development electronic commerce logistics management mindfulness

Methods: simulation regression analysis method robustness check literature sample data imputation

Is it more than allocating funds? Exploring the effect of enterprise crowdfunding on employee engagement

2022 | HICSS | Citations: 1

Authors: Benz, Carina

Abstract: Enterprise crowdfunding (ECF) has evolved as a novel form to foster innovation a ... Expand

Abstract: Enterprise crowdfunding (ECF) has evolved as a novel form to foster innovation and collaboration inside organizations. Research has so far focused on functional aspects related to the introduction of the crowdfunding mechanisms in enterprises (e.g., proposal characteristics or decision-making styles) leaving socio-economic effects on the organization and workforce unexplored. This work investigates the relationship between enterprise crowdfunding and the engagement of participating employees. By conducting an online survey with 321 employees of a multi-national manufacturing and electronics corporation, we find increased levels of employee engagement contingent upon participation in enterprise crowdfunding. These findings contribute to the understanding of effects related to the introduction of crowd-innovation platforms and enterprise social systems. From a practical perspective, they may foster the spread of enterprise crowdfunding as a tool being recognized to promote both, crowd-based innovation and employee engagement. Collapse

Topics: crowdfunding crowdsourcing innovation management missing data enterprise social media

Methods: survey experimental group longitudinal research Mann Whitney U test factor analysis

Explaining Data-Driven Decisions made by AI Systems: The Counterfactual Approach

2022 | Management Information Systems Quarterly | Citations: 0

Authors: Fernández-Loría, Carlos; Provost, Foster; Han, Xintian

Abstract: We examine counterfactual explanations for explaining the decisions made by mode ... Expand

Abstract: We examine counterfactual explanations for explaining the decisions made by model-based AI systems. The counterfactual approach we consider defines an explanation as a set of the system’s data inputs that causally drives the decision (i.e., changing the inputs in the set changes the decision) and is irreducible (i.e., changing any subset of the inputs does not change the decision). We (1) demonstrate how this framework may be used to provide explanations for decisions made by general data-driven AI systems that can incorporate features with arbitrary data types and multiple predictive models, and (2) propose a heuristic procedure to find the most useful explanations depending on the context. We then contrast counterfactual explanations with methods that explain model predictions by weighting features according to their importance (e.g., Shapley additive explanations [SHAP], local interpretable model-agnostic explanations [LIME]) and present two fundamental reasons why we should carefully consider whether importance-weight explanations are well suited to explain system decisions. Specifically, we show that (1) features with a large importance weight for a model prediction may not affect the corresponding decision, and (2) importance weights are insufficient to communicate whether and how features influence decisions. We demonstrate this with several concise examples and three detailed case studies that compare the counterfactual approach with SHAP to illustrate conditions under which counterfactual explanations explain data-driven decisions better than importance weights. Collapse

Topics: shapley additive explanation artificial intelligence Facebook decision making laptop computer

Methods: case study computational algorithm data imputation substitution by the mean data imputation regression analysis method

Something’s Missing? A Procedure for Extending Item Content Data Sets in the Context of Recommender Systems

2022 | Information Systems Frontiers | Citations: 0

Authors: Heinrich, Bernd; Hopf, Marcus; Lohninger, Daniel; Schiller, Alexander; Szubartowicz, Michael

Abstract: The rapid development of e-commerce has led to a swiftly increasing number of co ... Expand

Abstract: The rapid development of e-commerce has led to a swiftly increasing number of competing providers in electronic markets, which maintain their own, individual data describing the offered items. Recommender systems are popular and powerful tools relying on this data to guide users to their individually best item choice. Literature suggests that data quality of item content data has substantial influence on recommendation quality. Thereby, the dimension completeness is expected to be particularly important. Herein resides a considerable chance to improve recommendation quality by increasing completeness via extending an item content data set with an additional data set of the same domain. This paper therefore proposes a procedure for such a systematic data extension and analyzes effects of the procedure regarding items, content and users based on real-world data sets from four leading web portals. The evaluation results suggest that the proposed procedure is indeed effective in enabling improved recommendation quality. Collapse

Topics: missing data recommender system data quality electronic commerce decision support system

Methods: computational algorithm data imputation literature study binary classification literature sample

Business Process Management Culture in Public Administration and Its Determinants

2022 | Business & Information Systems Engineering | Citations: 0

Authors: Kregel, Ingo; Distel, Bettina; Coners, André

Abstract: Public administration institutions increasingly use business process management ... Expand

Abstract: Public administration institutions increasingly use business process management (BPM) to innovate internal operations, increase process performance and improve their services. Research on private sector companies has shown that organizational culture may impact an organization's BPM and this culture is often referred to as BPM culture. However, similar research on public administration is yet missing. Thus, this article assesses BPM culture in Germany’s municipal administration. 733 online survey responses were gathered and analyzed using MANOVA and follow-up discriminant analyses to identify possible determinants of public administration’s BPM culture. The results indicate that the employees’ professional experience and their responsibility influence the assessment of BPM culture, as does the size of a municipality. Based on these findings, the article proposes testable relationships and an agenda for further research on BPM culture in public administration. Collapse

Topics: innovation management business process management missing data organizational impact IT personell turnover

Methods: survey multivariate analysis of variance discriminant analysis survey design qualitative interview

Using Machine Learning Techniques to Explore Extra-Role Security Behavior

2021 | International Conference on Information Systems | Citations: 0

Authors: Frank, Muriel; Ranft, Lukas Manuel

Abstract: Many contemporary organizations pit on information security policy compliance t ... Expand

Abstract: Many contemporary organizations pit on information security policy compliance to combat information security threats originating from their own workforce. However, recent findings suggest that employees' adherence to security rules and regulations alone is insufficient to protect organizational assets. Instead, extra-role security behavioractions that go beyond what is specified in policies and are beneficial to the firms -is needed. So far, research with regard to extra-role security behavior is meager, in particular concerning contextual determinants influencing whether employees exhibit prosocial behaviors or not. Hence, this paper uses predictive modeling, or more precisely supervised machine learning, to classify employees according to their likelihood of exhibiting extra-role security behaviors. Results indicate that informational, social, and task context factors significantly impact the performance of extra-role security behavior. Collapse

Topics: phishing electronic mail data security computer crime IT security training

Methods: machine learning computational algorithm decision tree classification supervised learning data imputation

Theories: social exchange theory

Predicting the Outcome of a Football Game: A Comparative Analysis of Single and Ensemble Analytics Methods

2019 | HICSS | Citations: 0

Authors: Eryarsoy, Enes; Delen, Dursun

Abstract: As analytical tools and techniques advance, increasingly large numbers of resear ... Expand

Abstract: As analytical tools and techniques advance, increasingly large numbers of researchers apply these techniques on a variety of different sports. With nearly 4 billion followers, it is estimated that association football, or soccer, is the most popular sports for fans across the world by a large margin. The objective of this study is to develop a model to predict the outcomes of soccer (or association football) games (win-loss-draw), and determine factors that influence game outcomes. We used 10 years of comprehensive game-level data spanning the years 2007-2017 in the Turkish Super League, and tested a variety of classifiers to identify the most promising methods for outcome predictions. Collapse

Topics: missing data logistics management

Methods: decision tree classification random forest classification computational algorithm machine learning data imputation

Counterfactual Explanations for Data-Driven Decisions

2019 | International Conference on Information Systems | Citations: 0

Authors: Fernandez, Carlos; Provost, Foster; Han, Xintian

Abstract: Users’ lack of understanding of systems that use predictive models to make autom ... Expand

Abstract: Users’ lack of understanding of systems that use predictive models to make automated decisions is one of the main barriers for their adoption. We adopt the increasingly accepted view of a counterfactual explanation for a system decision: a set of the system inputs that is causal (meaning that removing them changes the decision) and irreducible (meaning that removing any subset of the inputs in the explanation does not change the decision). We generalize previous work on counterfactual explanations in three ways: we explain system decisions rather than model predictions; we do not enforce any specific method for removing inputs, and our explanations can incorporate inputs with arbitrary data structures. We also show how model-agnostic algorithms can be tweaked to find the most useful explanations depending on the context. Finally, we showcase our approach using a real data set to illustrate its advantages over other explanation methods when the goal is to understand system decisions better. Collapse

Topics: shapley additive explanation decision making missing data logistics management

Methods: computational algorithm regression analysis method machine learning substitution by the mean data imputation data imputation

Demand Heterogeneity in Platform Markets: Implications for Complementors

2018 | Organization Science | Citations: 32

Authors: Rietveld, Joost; Eggers, J. P.

Abstract: While two-sided platforms (e.g., video game consoles) depend on complements (e.g ... Expand

Abstract: While two-sided platforms (e.g., video game consoles) depend on complements (e.g., games) for their success, the success of complements is also influenced by platform-level dynamics. Research suggests that greater platform adoption benefits complements by providing more potential users, but this assumes that platform adopters are homogeneous. We build on extensive research exploring the heterogeneity between early and late platform adopters to identify counterintuitive dynamics for complements. Complements launched early in a platform’s life cycle face an audience entirely of early platform adopters, whereas later-launching complements face a mixed audience of both early and late adopters, and we argue that differences in preferences and behavior between early and late adopters affect whether complements will succeed and which types will be most successful. We explore these dynamics in the context of the console video game industry using a unique data set of 2,918 video games released in the United Kingdom from 2000 to 2007. We show that despite the increase in the potential user pool as the platform evolves, video games launched later in the platform life cycle realize lower sales than those launched earlier. While increased competition explains part of this effect, we show substantial evidence consistent with our theory of preference differences between early and late adopters. This includes the finding that the negative effect is stronger for novel games and that the gap between popular and less popular complements widens as later adopters move into the platform, consistent with late adopters being risk averse and seeking to avoid purchasing mistakes. The e-companion is available at https://doi.org/10.1287/orsc.2017.1183. Collapse

Topics: video game network effect innovation management willingness to pay digital platform

Methods: descriptive statistic robustness check theory development qualitative interview statistical hypothesis test

Can Information Availability Increase Patient Compliance? Mitigating Uncertainty Perceptions in the Provider-Patient Relationship

2017 | HICSS | Citations: 0

Authors: Bozan, Karoly

Abstract: Patient compliance with provider directions is central to patients’ well being, ... Expand

Abstract: Patient compliance with provider directions is central to patients’ well being, and non-compliance has been identified as a leading cause of increasing healthcare costs. While numerous factors may affect patient compliance, we investigate the mediating effect of patient health information availability on the relationship between perceived uncertainty and patients’ motivation to comply with providers’ orders. To understand how to mitigate perceived uncertainty, we extend the underlying principles of principal-agent theory—hidden information and hidden action—and propose three uncertainty-mitigating factors: perceived information asymmetry, fear of opportunism, and physician quality. The proposed structural model is empirically tested using data from 184 patients. Our model is supported, and the results provide an understanding of the process by which patients engage in their care through the support of information technology. We discuss the implications for understanding and facilitating the providerpatient relationship and its effect on patients' motivation to comply through the principal-agent perspective. Collapse

Topics: healthcare data health information system missing data electronic health electronic health record

Methods: survey partial least squares regression data transformation literature study exploratory factor analysis

Theories: agency theory lemon market theory