The success of open source software (OSS) projects depends on sustained contributions by developers who often display a wide variety of contribution patterns. Project leaders and stakeholders would ...strongly prefer that developers not only maintain but preferably increase their contributions over time as they gain experience. Corporations increasingly complement OSS developer motivations (such as fit in terms of shared values with the project community) by paying them to sustain contributions. However, practitioners argue whether payment helps or hurts projects because monetary compensation may dampen developer motivation in the long run, making it difficult for project leaders to understand what to expect from developers over time. Using Herzberg motivation-hygiene framework, we explore how developer perceptions of value fit with the project and being paid interact to determine the level of code contribution and its rate of change over time (i.e., growth). Using a survey of 564 developers across 431 projects on GitHub, we build a three-level growth model explaining the code contribution and its growth over a six-month period. We find that value fit with the project positively influences both the level and growth of code contribution. However, there are notable differences among paid and unpaid developers in the impact of value fit on their level and growth in code contributions over time. The implications of our work will be of interest to researchers, practitioners, and organizations investing in open source projects.
This study investigates how knowledge sharing (KS) contributes to firm performance (FP) through the enhancement of innovation and/or intellectual capital (IC) using data collected from Chinese ...high-technology firms. The paper proposes three alternative models that suggest different mediating roles of innovation and IC components in the KS→FP nomological network based on existing theory. The paper then compares these models in terms of in-sample explanatory and out-of-sample predictive powers using consistent partial least squares path modeling (PLSc). Results indicate that in the best performing model, innovation and IC simultaneously mediate the relationship between KS and FP in this specific context. The findings offer insights regarding the parallel mediation roles of innovation and IC in the KS→FP process, showcase the predictive utility of PLSc, and can help managers set priorities when leveraging KS to achieve specific performance goals.
ABSTRACT
Partial least squares path modeling (PLS‐PM) has become popular in various disciplines to model structural relationships among latent variables measured by manifest variables. To fully ...benefit from the predictive capabilities of PLS‐PM, researchers must understand the efficacy of predictive metrics used. In this research, we compare the performance of standard PLS‐PM criteria and model selection criteria derived from Information Theory, in terms of selecting the best predictive model among a cohort of competing models. We use Monte Carlo simulation to study this question under various sample sizes, effect sizes, item loadings, and model setups. Specifically, we explore whether, and when, the in‐sample measures such as the model selection criteria can substitute for out‐of‐sample criteria that require a holdout sample. Such a substitution is advantageous when creating a holdout causes considerable loss of statistical and predictive power due to an overall small sample. We find that when the researcher does not have the luxury of a holdout sample, and the goal is selecting correctly specified models with low prediction error, the in‐sample model selection criteria, in particular the Bayesian Information Criterion (BIC) and Geweke–Meese Criterion (GM), are useful substitutes for out‐of‐sample criteria. When a holdout sample is available, the best performing out‐of‐sample criteria include the root mean squared error (RMSE) and mean absolute deviation (MAD). We recommend against using standard the PLS‐PM criteria (R2, Adjusted R2, and Q2), and specifically the out‐of‐sample mean absolute percentage error (MAPE) for prediction‐oriented model selection purposes. Finally, we illustrate the model selection criteria's practical utility using a well‐known corporate reputation model.
Purpose
Researchers often stress the predictive goals of their partial least squares structural equation modeling (PLS-SEM) analyses. However, the method has long lacked a statistical test to compare ...different models in terms of their predictive accuracy and to establish whether a proposed model offers a significantly better out-of-sample predictive accuracy than a naïve benchmark. This paper aims to address this methodological research gap in predictive model assessment and selection in composite-based modeling.
Design/methodology/approach
Recent research has proposed the cross-validated predictive ability test (CVPAT) to compare theoretically established models. This paper proposes several extensions that broaden the scope of CVPAT and explains the key choices researchers must make when using them. A popular marketing model is used to illustrate the CVPAT extensions’ use and to make recommendations for the interpretation and benchmarking of the results.
Findings
This research asserts that prediction-oriented model assessments and comparisons are essential for theory development and validation. It recommends that researchers routinely consider the application of CVPAT and its extensions when analyzing their theoretical models.
Research limitations/implications
The findings offer several avenues for future research to extend and strengthen prediction-oriented model assessment and comparison in PLS-SEM.
Practical implications
Guidelines are provided for applying CVPAT extensions and reporting the results to help researchers substantiate their models’ predictive capabilities.
Originality/value
This research contributes to strengthening the predictive model validation practice in PLS-SEM, which is essential to derive managerial implications that are typically predictive in nature.
•A concise overview of open science and PLS-SEM, along with a discussion of ways in which open science practices can aid PLS-SEM’s reproducibility.•Structured literature review of open science ...practices in PLS-SEM in marketing research.•Recommendations on how to employ open science practices in PLS-SEM.•Proposing a preregistration template for PLS-SEM.
Driven by the high-profile failures to reproduce and replicate published findings, there have been increasing demands to adopt open science practices across scientific disciplines in order to enhance research transparency. Critics have highlighted the use of underpowered studies and researchers’ analytical degrees of freedom as factors contributing to these issues. Despite methodological advances and updated guidelines, similar concerns persist regarding studies utilizing partial least squares structural equation modeling (PLS-SEM). Open science practices can help alleviate these concerns by facilitating transparency in PLS-SEM-based studies. However, the current level of adherence to these practices remains unknown. In this article, we conduct a comprehensive literature review of leading marketing journals to assess the extent to which open science practices are implemented in PLS-SEM-based studies. Based on the observed lack of adoption, we propose a PLS-SEM-specific preregistration template that researchers can use to foster transparency in their analyses, thereby bolstering confidence in their findings.
Display omitted
•Customer satisfaction relationships differ in online and offline purchasing.•Perceived value is a stronger driver of satisfaction in online purchases.•Overall quality and ...expectations are stronger drivers of satisfaction offline.•Customers are more satisfaction-sensitive when purchasing online.•Differences generally persist across customer demographics and retail categories.
Retailers seek to utilize both online and offline purchase channels strategically to satisfy customers and thrive in the marketplace. Unfortunately, current multichannel research is deficient in answering what drives customers’ satisfaction, and consequently their loyalty, differently when customers purchase online versus at a physical store. This gap in knowledge can be a significant concern for retailers due to the negative impact of having dissatisfied customers on their bottom lines. Using a version of the American Customer Satisfaction Index (ACSI) model, we demonstrate several important purchase-channel differences in the antecedents of customer satisfaction and its subsequent effect on customer loyalty. Specifically, we show that when retail customers buy electronic goods online they view purchase value as a significant attribute in rating satisfaction, and are more satisfaction-sensitive when making repurchase decisions than when they purchase offline. On the other hand, the overall quality of the purchase experience and customer expectations are stronger drivers of customer satisfaction in the offline purchases. We provide evidence that these differences between the channels generally persist across customer demographics (gender, age, and education) and broader product categories, and we also discuss the specific contexts where they do not. Our work offers actionable guidance to retailers seeking to enhance customer satisfaction and loyalty across both the online and offline channels.
ABSTRACT
Management researchers often develop theories and policies that are forward‐looking. The prospective outlook of predictive modeling, where a model predicts unseen or new data, can complement ...the retrospective nature of causal‐explanatory modeling that dominates the field. Partial least squares (PLS) path modeling is an excellent tool for building theories that offer both explanation and prediction. A limitation of PLS, however, is the lack of a statistical test to assess whether a proposed or alternative theoretical model offers significantly better out‐of‐sample predictive power than a benchmark or an established model. Such an assessment of predictive power is essential for theory development and validation, and for selecting a model on which to base managerial and policy decisions. We introduce the cross‐validated predictive ability test (CVPAT) to conduct a pairwise comparison of predictive power of competing models, and substantiate its performance via multiple Monte Carlo studies. We propose a stepwise predictive model comparison procedure to guide researchers, and demonstrate CVPAT's practical utility using the well‐known American Customer Satisfaction Index (ACSI) model.
Driven by the growing importance of the digital provision of government services (e-government), recent research has sought to develop and test conceptual models of citizen satisfaction and trust ...with these services. Yet, there remains little agreement on how to optimally model these relationships with regards to the somewhat divergent goals of explanation and prediction of citizen trust. In this paper, we test two prominent modeling paradigms of the e-government satisfaction-trust relationship: the “service quality” model and the “expectancy-disconfirmation” model. We compare several variations of these models for their in-sample explanatory abilities, out-of-sample predictive abilities, and parsimony. To test the models, we examine a pooled, cross-agency sample of survey data measuring citizens' experiences with and perceptions of three important and widely accessed U.S. federal e-government services—the webpages of the Social Security Administration, the Internal Revenue Service, and the U.S. Census Bureau. Our findings suggest that while the expectancy-disconfirmation paradigm performs well in explanation, a parsimonious model with an “overall quality-satisfaction-trust” link is best suited for predicting trust. In addition, the service quality paradigm offers the best compromise between predictive accuracy and explanatory power. These findings offer new insights for academic researchers, government agencies, and practitioners, especially those deciding upon an empirical model to adopt to measure e-government satisfaction and its impact upon citizen trust.
•A comparison between Expectancy-Disconfirmation & Service Quality paradigms for citizen E-Government satisfaction & trust.•The role of explanation & prediction of citizen trust as empirical modeling goals is highlighted.•Expectancy-Disconfirmation paradigm well-suited suited for explaining citizen trust.•A parsimonious model with an “overall quality-satisfaction-trust” link is best suited for predicting trust.•The service quality paradigm offers the best compromise between predictive accuracy & explanatory power.
The ongoing scientific discourse surrounding the replication crisis in behavioral research, including management information systems (MIS) research, underscores the importance of innovative and ...rigorous approaches to theory development and validation. This article proposes the EP-mixed framework, which addresses the necessity of an ontological distinction between explanation and prediction in MIS theories, along with the epistemological challenges associated with conflating exploratory and confirmatory research during the design of robust, replicable theories. EP-mixed refers to theories that explain and predict (i.e., EP theories) developed using a mixed mode that combines the strengths of both exploratory and confirmatory research. The EP-mixed framework guides researchers in selecting appropriate analytical approaches based on their research goals and the type of theory being developed. While it can be applied in conjunction with a broad spectrum of statistical methods to enhance the robustness and replicability of MIS theories, we elaborate on the predictive analytic tools available in partial least squares structural equation modeling (PLS-SEM) as an exemplar for operationalizing the framework.
•Proposes a new framework for conducting replicable research.•Clarifies the distinction between exploratory and confirmatory research.•Clarifies the distinction between explanatory and predictive modeling.•Guides in creating and validating replicable explanatory-predictive theories.•Discusses tools in PLS-SEM to operationalize the framework.
While the U.S. federal government has adopted myriad initiatives mandating collection of citizen evaluations of its services, scant research exists into how prior biases such as those arising from ...political partisanship affect these performance metrics. In this study, we examine a multi-year sample asking U.S. citizens about their experiences with federal government services (n = 8,341). Guided by motivated reasoning theory, the results show that partisanship affects citizen satisfaction, confidence, and trust in the federal government during both Democratic (2015-2016) and Republican (2017-2018) presidential administrations. However, the results indicate an asymmetric 'president-in-power' effect, complicating efforts to interpret this data dynamically and over time as power changes hands.