{"title":"Perspective: Imputing missing values in IB research using endpoint-heterogeneous beta regression models: An application to Hofstede's LTO measures","authors":"Eugene D. Hahn , Jonathan P. Doh","doi":"10.1016/j.jwb.2025.101644","DOIUrl":null,"url":null,"abstract":"<div><div>Addressing missing values in empirical research is a longstanding issue and challenge for scholars in all disciplines. In International Business, the challenge has typically been met with an array of strategies, ranging from listwise deletion to more sophisticated approaches, including regression-based imputation. Current models for addressing missing values where data is bounded have several limitations. We present a new model for imputing values for bounded data via endpoint-heterogeneous beta regression. We use this approach to generate imputed values for additional countries in Hofstede's long-term orientation (LTO) scale. LTO was the first widely used scale to be based on a non-Western value system, however, Hofstede was not able to collect LTO data for some countries that were part of his original survey and therefore the LTO measure has numerous missing values. Our paper introduces a novel approach to addressing missing values in LTO and provides a template for scholars facing similar challenges in a wide range of IB and other domains.</div></div>","PeriodicalId":51357,"journal":{"name":"Journal of World Business","volume":"60 4","pages":"Article 101644"},"PeriodicalIF":8.9000,"publicationDate":"2025-04-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Journal of World Business","FirstCategoryId":"91","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S1090951625000331","RegionNum":1,"RegionCategory":"管理学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"BUSINESS","Score":null,"Total":0}
引用次数: 0
Abstract
Addressing missing values in empirical research is a longstanding issue and challenge for scholars in all disciplines. In International Business, the challenge has typically been met with an array of strategies, ranging from listwise deletion to more sophisticated approaches, including regression-based imputation. Current models for addressing missing values where data is bounded have several limitations. We present a new model for imputing values for bounded data via endpoint-heterogeneous beta regression. We use this approach to generate imputed values for additional countries in Hofstede's long-term orientation (LTO) scale. LTO was the first widely used scale to be based on a non-Western value system, however, Hofstede was not able to collect LTO data for some countries that were part of his original survey and therefore the LTO measure has numerous missing values. Our paper introduces a novel approach to addressing missing values in LTO and provides a template for scholars facing similar challenges in a wide range of IB and other domains.
期刊介绍:
The Journal of World Business holds a distinguished position as a leading publication within the realm of International Business. Rooted in a legacy dating back to 1965, when it was established as the Columbia Journal of World Business, JWB is committed to disseminating cutting-edge research that reflects significant advancements in the field. The journal actively seeks submissions that propel new theoretical frameworks and innovative perspectives on International Business phenomena. Aligned with its domain statement, submissions are expected to possess a clear multinational, cross-border, or international comparative focus, while remaining pertinent to the study of management and organizations. JWB particularly encourages submissions that challenge established theories or assumptions, presenting pioneering or counterintuitive findings. With an inclusive approach, the journal welcomes contributions from diverse conceptual and theoretical traditions, encompassing allied social sciences and behavioral sciences. Submissions should either develop new theories or rigorously test existing ones, employing a variety of qualitative, quantitative, or other methodological approaches. While JWB primarily caters to scholars and researchers, it values contributions that explore implications for Multinational Enterprises and their management, as well as ramifications for public policy and the broader societal role of business.