Start Date
11-12-2016 12:00 AM
Description
This research-in-progress paper describes cumulative meta-analysis, or meta-trend analysis, a form of meta-analysis that considers temporal trends in effect sizes. While this method is common in medical sciences, it is just starting to gain traction in behavioral research, and temporal trends have typically not been addressed in IS research. A review of 64 meta-analysis papers from 15 IS journals confirms that IS research is generally blind to time. No IS paper has employed meta-trend analysis to test for temporal trends, and less than a quarter of the papers reviewed have any treatment or mention of the possible impact of time. Support from ecological systems theory, in particular the idea of proximal processes, is used to explain why IS researchers may expect temporal trends in effect sizes. To illustrate this, meta-trend analysis is conducted on several frequently examined relationships between IS constructs. Preliminary evidence of temporal trends is observed.
Recommended Citation
Agogo, David and Hess, Traci, "Blind to Time? Temporal Trends in Effect Sizes in IS Research" (2016). ICIS 2016 Proceedings. 10.
https://aisel.aisnet.org/icis2016/Methodological/Presentations/10
Blind to Time? Temporal Trends in Effect Sizes in IS Research
This research-in-progress paper describes cumulative meta-analysis, or meta-trend analysis, a form of meta-analysis that considers temporal trends in effect sizes. While this method is common in medical sciences, it is just starting to gain traction in behavioral research, and temporal trends have typically not been addressed in IS research. A review of 64 meta-analysis papers from 15 IS journals confirms that IS research is generally blind to time. No IS paper has employed meta-trend analysis to test for temporal trends, and less than a quarter of the papers reviewed have any treatment or mention of the possible impact of time. Support from ecological systems theory, in particular the idea of proximal processes, is used to explain why IS researchers may expect temporal trends in effect sizes. To illustrate this, meta-trend analysis is conducted on several frequently examined relationships between IS constructs. Preliminary evidence of temporal trends is observed.