Table 1 Reasons for combining experimental or observational research data across cohorts

Theoretical

 Pooling results in an area of research

 Increasing generalizability

 Interest in known or potential sources of heterogeneity

 Replication of results

 Questions of interest centre on association/modelling

 Interest in a subset of data

 Identifying directions for future research

Statistical

 Effects of interest are small but important

 More sophisticated models necessary

 Increasing observations of infrequent events

 Minimising effects of attrition over time

 Standardizing modelling used in predicting outcomes

 Aggregation of data from repeated experiments

Practical

 Maximising existing resources

 Time efficiency/producing information of current public interest

 Cost efficiency

 Preliminary exploration to support funding for more comprehensive research design or inform later phases of research

 Features appealing to funding bodies