|
Meta-Analysis in Biology| old_uid | 19692 |
|---|
| title | Meta-Analysis in Biology |
|---|
| start_date | 2022/01/18 |
|---|
| schedule | 14h-16h |
|---|
| online | no |
|---|
| details | Compte tenu des conditions sanitaires, la salle de conférence de l'IHPST sera accessible uniquement sur présentation du pass sanitaire. Nous serons tenus de respecter le port du masque et les gestes barrières durant la séance. |
|---|
| summary | Meta-analysis is a statistical tool for analysing and synthesising the results of large numbers of individual studies. The primary aim of a meta-analysis is to identify causal relationships from different types of evidence. In some disciplines (such as biology), meta-analyses are also used as a means for generating generalizations. However, the status and role of meta-analysis remains a controversial and divisive issue, as its adversaries have condemned it as irrevocably biased. A second criticism concerns the heterogeneity of the primary studies which a meta-analysis attempts to synthesise, as extensive differences in terms of the type of experiment, species being investigated, sample size, type of intervention etc., can render comparisons between studies meaningless. While I agree with the critics that the quality of many meta-analyses is quite low and that heterogeneity creates practical difficulties, I believe that they are a useful tool for biologists. I will show that: (i) the quality of meta-analyses (in biology) has been steadily increasing and (ii) in some contexts, heterogeneity can be a feature rather than a bug, as a certain level of heterogeneity in primary studies can provide a better understanding of the nature of the phenomenon under investigation and can be used to test the overall robustness/scope of the phenomenon. |
|---|
| responsibles | Montjean, Delettre, Haas |
|---|
| |
|