Do you use census data? We'd like your feedback.

Last updated

27 Nov 2024

Working with census statistics

Making comparisons between groups

Census data is incredibly powerful due to its breadth and depth. We are now making this power available to users via our new Flexible Table Builder.

Census data:

  • represents the whole of population of Scotland
  • includes a wide range of characteristics data
  • includes detailed breakdowns of these characteristics
  • is available down to local area level

Census data is useful for looking at how one group of people in the population compares to another. But it is not always easy to make fair comparisons. So we need to careful when interpreting the results.

UK Armed Forces veterans example

Consider a comparison between the general health of Armed Forces veterans and the rest of the population.

We need to be careful when making this comparison. The veterans population is different from the non-veteran population in many important ways. One key difference is that the veteran population is generally a lot older on average.

The data shows that the percentage of people reporting a disability or long-term health problem is higher among veterans than non-veterans. But we know that  disability is closely related to age. So the fact that the veteran population is typically older will explain some (potentially a lot) of the difference in disability and long-term health problems.

The Scottish Government has published a paper that goes into more detail on the issue of comparing veterans to non-veterans. We know the age profile of the veterans population differs from the non-veteran population. The paper demonstrates how the two populations differ even when comparing equivalent age groups. This highlights the diversity within the veterans population itself.

Similarly, the ONS advises users of veterans data to: “Take care when comparing characteristics of UK armed forces veterans and non-veterans, as veterans are mainly male and older. Not adjusting for this fact can lead to misinterpretation, as variables such as health strongly relate to age and sex”.

The ONS has also produced a series of reports which include adjusted comparator data for the non-veteran population.

Other groups

We need to be careful when making comparisons with other groups within the population, not just veterans. Our topic reports looked at the age distributions for people with different characteristics.

There are several examples where the age distribution within specific groups is different from the rest of the population. For example, people who reported a mental health condition tend to be younger on average. And people who own their home outright tend to be older on average.

Practical tips

When making comparisons between population groups you should ask yourself:

  • what makes the group(s) I’m interested in unique/different?
  • are there any differences in the age, sex, or geographic distributions?
  • what about other characteristics?
  • are any of these differences likely to have an impact on the thing I’m measuring/comparing?

 

If the answer to these questions is yes (or maybe) then you should consider:

  • highlighting known differences in characteristics that might impact the comparison when presenting your analysis
  • breaking down the data to try to account for differences and compare like with like e.g. comparing older veterans with older non-veterans
  • applying a weighting to comparator groups, or standardising the data in some way e.g. age standardisation
Contents