Baselines in Statistics

by Craig Shrives

What Are Baselines in Statistics?

In Statistics, a baseline is "a measurement, calculation or location used as a basis for comparison". In other words, it is "a starting point for a comparison".

Examples of Comparisons with Different Baselines

I like to think of the baseline as the context for a comparison. Things have to be compared in similar contexts, otherwise the comparison might be meaningless. For example, you can't claim the side wall of a building is taller than the rear wall if you measure one from the floor and the other from the foundations; you can't claim your village is safer than the town because there are fewer burglaries in the village each year; and you can't claim the residents of Nevada look after their cars better than those from Alaska because their cars last longer.

The elements being compared in each of these scenarios have different baselines. They cannot be compared accurately until the baselines are made similar or accounted for. So, to perform meaningful comparisons, you would have to measure both walls from the same point; do something like determine the number of burglaries in the village and town per household; and make some adjustments for the environmental effects on cars. Only then could useful comparisons be made. Ignoring the baseline is a common error when presenting statistics. Sometimes it's done accidentally, but often it's done intentionally. If you ignore the baseline, you can't really offer a true comparison.

Make Adjustments for Differences in Baselines

If you compare raw numbers without adjusting for baseline differences, you will not get a meaningful result. (If that suits your purposes, you could go for it – just make sure you have a plan for when your statistics are ridiculed.)

See Also