I am a master's student in organisational psychology. My dissertation is a systematic review of the measures of workload in air traffic control. This will be a narrative analysis of the findings and implications. Workload in air traffic control is tested under many heterogeneous conditions, which authors have commented prevents from doing a meta-analysis. My analysis will focus on which types of measures (self-assessment, eye movement, EEG, heart rate) have successfully discriminated levels of workload, commenting on the design, methodology and participants in the study, which are often different. Is it possible to borrow elements of meta-synthesis to achieve a structured and comprehensive level of analysis? Thank you!
Workload becomes the DV and will do a meta anslysi of all the bivariate relations for ivs effecting if
Hi Kalina,
why exactly should it be not possible to do a meta-analysis? Heterogeneity is the most prevalent reasons to do a meta-analysis....You could try to code characteristics of the studies that might explain the heterogeneity.
Best,
Holger
Kalina Tcholakova
I think reasons not to do a meta-analysis revolve around intolerable differences between studies. Ones that may be mentioned in the literature of your field and proved to make a difference in the outcome. For example, in clinical research different doses of a drug can be a problem, specially if the difference was established in published articles.
To manage these discrepancies, you can do a "subgroup analysis" which makes a separate MA for each set of circumstances and compare each group with certain differences. Or you can do a sensitivity analysis, which involves removing studies that are substantially different (by some defined factor) and see whether removing them makes changes on the pooled estimate. However, it doesn't completely solve the problem, sometimes it's just not a good idea to compare studies using a meta-analysis. If so, try to report results of studies narratively, many systematic reviews are published without MAs. Good luck!
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Techniques
- Meta-Analysis