Appendix A — Exercise 1 solutions

Question 1

For each of the following research questions, identify the Population, Intervention and Comparison (if applicable) and the Outcome.

  1. Does the introduction of a 4-day working week increase productivity in government departments?

P: Staff working in government departments (these departments would be named in real life!)

I: 4-day working week

C: Regular working structure

O: Productivity, e.g. deadlines met (this would need to be more well-defined, specific and measurable!)

  1. What is the average time between an offence being committed and case completion for defendants dealt with at magistrates’ courts in the North West of England?

P: Defendants dealt with at magistrates’ courts in North West England

I: N/A

C: N/A

O: Time between offence being committed and case completion

Note

The outcome is the time between these points, not the average that is referenced in the question. Picture the dataset: the difference in time will be recorded for each defendant as a variable.

  1. How has the prison population in England and Wales changed compared to pre-pandemic levels?

P: Prisons in England and Wales

I: Pre-pandemic

C: Post-pandemic

O: Population in prisons

Note

This question can be tricky as the outcome is actually a population. Again, it helps to picture the dataset we would use to answer this research question:

  • Each row would be a prison in England or Wales
  • There would be a time period variable, where observations are either classified as pre- or post-pandemic
  • The outcome variable would be the prison population measured at each time point for each prison in England and Wales.

Question 2

If the practitioners were more likely to record information on simpler cases, as they had more time to do so, this would mean these are more likely to be observed than the more technical cases. The data will therefore not be random and will not be representative of these technical cases. This means that the results of the analysis would be biased.

If the practitioner had not recorded details because they were off work with the flu and had not had time to catch up on their work, this is not related to the data and the missing values are random. This would make our sample smaller than anticipated but still random and representative of the target population. Therefore, results from this analysis would be unbiased.