For scientific accuracy, test two ads in the same Ad Group

Each Ad Group contains only one keyword combination of 2 or 3 search words.
Don’t let Google run ads when people put other words into Google search that you don’t choose. Only use Phrase Match Keywords and Exact Match Keywords. Never enter Broad Match Keywords.
Write both ads with the search words in Headline 1. Write a different sales pitch for each ad in Headline 1, Headline 2 and Description 1.
Run both ads at least 200 times for statistical accuracy -- in Settings / additional settings at bottom, change Ad Rotation to "run ads indefinitely" so Google doesn't interrupt the test before both ads run 200 times.
Run the ad that gets a higher CTR Click Thru Rate, which is the percentage of times people clicked on the ad.
You know, with scientific certainty, which ad people liked best, because everyone saw their search words in Headline 1 for both ads and the only difference was the sales pitch they saw.