{"title":"Should A Rule of Thumb be used to Calculate PLS-SEM Sample Size","authors":"Chanta Jhantasana","doi":"10.48048/asi.2023.254658","DOIUrl":null,"url":null,"abstract":"Partial least square structural equation modeling (PLS-SEM) is more commonly used in marketing research because small sample sizes can be used. The main advantage is that the rule of thumb is often used to determine sample size, but the results may be underpowered. Therefore, appropriate sample size is still required to reach the acquired power of 0.80, which should be adequate to avoid false positive and false adverse effects arising from sample sizes that are too large or too small.\nThis research investigates the impact of different sample sizes on the power of analysis, the effect size, and the significance level of the model fitness and parameter estimation process. Many methods are used to generate study sample sizes, such as minimum R2, ten-time rule, inverse square root method, Marsh et al. method, Soper method, and Yamane method. The rule of thumb methods of minimum R2 and ten-time rule generate sample sizes that are too small and inappropriate for PLS-SEM.\nHowever, the findings have shown that PLS-SEM can be effective with small sample sizes, but the sample size should be more significant than that generated by the rule-of –thumb methods. The appropriate sample size for this study was 50, with a power of 0.81 and an effect size (f2) ranging between 0.437 and 0.506.","PeriodicalId":43547,"journal":{"name":"SOJOURN-Journal of Social Issues in Southeast Asia","volume":null,"pages":null},"PeriodicalIF":0.3000,"publicationDate":"2023-04-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"SOJOURN-Journal of Social Issues in Southeast Asia","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.48048/asi.2023.254658","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q3","JCRName":"AREA STUDIES","Score":null,"Total":0}
引用次数: 0
Abstract
Partial least square structural equation modeling (PLS-SEM) is more commonly used in marketing research because small sample sizes can be used. The main advantage is that the rule of thumb is often used to determine sample size, but the results may be underpowered. Therefore, appropriate sample size is still required to reach the acquired power of 0.80, which should be adequate to avoid false positive and false adverse effects arising from sample sizes that are too large or too small.
This research investigates the impact of different sample sizes on the power of analysis, the effect size, and the significance level of the model fitness and parameter estimation process. Many methods are used to generate study sample sizes, such as minimum R2, ten-time rule, inverse square root method, Marsh et al. method, Soper method, and Yamane method. The rule of thumb methods of minimum R2 and ten-time rule generate sample sizes that are too small and inappropriate for PLS-SEM.
However, the findings have shown that PLS-SEM can be effective with small sample sizes, but the sample size should be more significant than that generated by the rule-of –thumb methods. The appropriate sample size for this study was 50, with a power of 0.81 and an effect size (f2) ranging between 0.437 and 0.506.