Stories. Questions. Different perspectives.
Browse WAURE
← Advice & Questions
Politics · Advice & Questions ·

Are seasonally adjusted figures always the better version to use?

Politics category photographPolitics

Our civic group is preparing a workshop on reading economic statistics. We want to ask two different questions: whether activity changed from one month to the next beyond its usual seasonal pattern, and what the busy part of the calendar actually looked like. A draft handout treats the seasonally adjusted series as the only “proper” version for both.

The ONS explanation of seasonal adjustment says it estimates and removes systematic calendar-related variation to help comparisons between consecutive periods. It also shows that irregular fluctuations remain; adjustment does not reveal a perfectly smooth underlying trend.

I think an unadjusted series can still be relevant when the seasonal pattern is part of the question, while adjusted data help with a different comparison. How can we teach that choice without suggesting people should pick whichever version produces the political story they prefer? We aren’t using either series to forecast a particular local service.

Discussion

5 community replies

Write the question before choosing the series. “What happened in that calendar period?” and “How did consecutive periods compare after usual seasonal effects?” should lead the discussion, not a preferred headline.

Keep the labels visible whenever you switch between versions. A small footer on the first page won’t help someone comparing two later screenshots.

I’d include the point about irregular movement in the handout itself. “Seasonally adjusted” can otherwise sound like “all noise removed,” which the ONS explanation does not say.

If you show both versions, won’t participants just choose the one that agrees with their politics anyway?

Ask them to state a comparison rule before seeing an outcome, then explain any change of method. That won’t eliminate selective use, but it makes the reason for choosing a series discussable rather than hiding the choice behind “proper data.”

Join the discussion

More in Politics