报告详情
Design Sensitivity and Estimability of Cross-Level Burnout Analytics for Software Teams: A Reproducible Simulation Study
编号：57
访问权限：仅限参会人
更新：2026-11-10 16:33:33 浏览：9次
Online
报告开始：2026年11月21日 16:30（Asia/Kolkata）
报告时间：15min
所在会场：[S5] Data Science and Intelligent Analytics [S5-3] Data Science and Intelligent Analytics
暂无文件
摘要
Decision-support systems that relate team climate to employee strain increasingly specify cross-level models in which a team-level managerial attribute conditions an employee-level pathway. It is not known whether such models can be estimated at the team levels that software organisations have, and there have been reported failures to estimate the model, which are blamed on the model design. This paper outlines a reference architecture with privacy constraints for such analytics and sets the operating characteristics of its estimand via simulation. 1,000 replications were generated per cell, using a deterministic generator, and each construct was measured using ordinal items, thus introducing measurement error as the instrument design suggests, in seven team counts and three effect conditions. Three results follow. With few clusters, the cluster-robust interaction test is anti-conservative, in that it rejects at 0.107 with 30 teams and 0.067 with 60 teams, when the nominal value is 0.05. While the employee-level mediation paths are > 0.99 power at 60 teams, power for a small interaction is 0.325 at 60 teams and 0.810 at 250 teams. The reported singular-matrix failures for mixed-effects fits of this model are not necessarily due to the non-identifiability of the model, but rather to the specific solver. A confirmatory cohort of 250 development teams (100 disjoint held-out teams) reproduces the predicted behavior end to end. The archive is accompanied by all quantities being regenerated.
关键词
Decision support systems,Design of experiments,Hierarchical models,Psychological safety,Reproducibility,Simulation,Software engineering management,Statistical power
报告人
Muthukumaran M
Assistant Professor G.T.N. Arts College稿件作者
全部评论
重要日期
-
会议日期
11月20日
2026
至11月22日
2026
-
10月31日 2026
初稿截稿日期
-
11月07日 2026
提前注册日期
-
11月22日 2026
注册截止日期
主办单位
United Societies of Science
承办单位
Pondicherry University
协办单位
IEEE Section
IEEE Madras Section
IEEE Madras Section
联系方式
- ICon INDIA 2026
- ic******@usssociety.org
发表评论