If you aren't correcting for errors in data collection then your conclusions will be wrong.
When you're dealing with measurements made over decades, the only way to deal with errors in data collection of those past measurements is to apply corrections. You cant co back and take new measurements. Also, it's not as though they're just saying, "this doesn't agree with my opinion, I'll just change all the measurements by this much". They're looking at unexpected discrepancies and trying to determine why they exist.
For example, recently researchers noticed ocean temperature measurements collected by buoys read consistently lower than measurements collected by ships. The buoys are reading accurately, and the ships appear to be reading high. (not surprising, given that ships have engines that release heat) Before the pause, most of our ocean temperature measurements were made via ships. At the beginning of the "pause" we were moving away from ships and toward primarily buoy based measurements. This created an artificial drop in the average temperature that was due to changes in measurement, not changes in the actual temperature. If this is corrected for (adjust the measurements made by ships down a specific amount determined by the typical difference between ships and buoys) then the global warming pause completely disappears. The researchers didn't set out to make it disappear, it just came out of the data.