ABSTRACT

Convenience sampling dominates social science research. But there is a paucity of studies comparing the impact of sample source type based on composite-based theoretical model relationships. This study empirically tests four different sample sources (e.g. student, crowdsourced, professional panel, and respondent driven social network) to assess differences in construct measurement and structural models to determine how sample source can impact empirical results. Results show that in comparison to meta-analytic findings, there was no meaningful difference in either measurement or structural parameters between sample source groups. Further, results highlight the ability to pool the multiple sample sources into a single, more generalizable data set. We use the results to make a recommendation for data pooling from multiple sample sources after careful consideration of the population of interest. Insights from this study contribute to the debate on the impact sample sources have on empirical analyses of theoretical composites and their relationships.

摘要

方便抽样在社会科学研究中占主导地位。但是很少有研究比较基于复合理论模型关系的样本源类型的影响。本研究实证测试了四种不同的样本来源（例如学生、众包、专业小组和受访者驱动的社交网络），以评估构建测量和结构模型的差异，以确定样本来源如何影响实证结果。结果表明，与荟萃分析结果相比，样本来源组之间的测量或结构参数没有有意义的差异。此外，结果突出了将多个样本源汇集到一个更通用的数据集中的能力。在仔细考虑了感兴趣的人群之后，我们使用这些结果对来自多个样本源的数据池提出了建议。这项研究的见解有助于辩论样本来源对理论复合材料及其关系的实证分析的影响。