Of course when measuring X now, you have to ensure that methodology for measuring X has not changed.

For example, are medical examiners back then, or even now, more apt to rule a suicide an accidental death? Wide social trends can affect how these individual decisions are marked at population scales.