If you care about accuracy, cross "Katherine Simpson" off your tree and keep looking. There is absolutely nothing to justify her entry (guess) as his wife -- there's no record of her arrival or her birth in the AGBI, every Boston Transcript entry, the NEHGS, dozens of the most trusted genealogy books available, or the hundreds of family profiles -- there's just no valid reason to connect her as his spouse. In addition, because the children are born so late, his wife would likely be significantly younger than he is, which this Katherine is not. She certainly was not giving birth in the wilds of Dover in her late 40's in the 1600s. The correct Catherine would probably have been born around 1632 (20 years before the first birth). Since the Boston birth records begin in 1630, the correct Catherine may have been born here, or... or she may have been born in England with a lost, or un-located passenger record (which we might find if we knew her name), there's no telling -- but there's virtually no chance her last name is Simpson.
Leaving incorrect information from a bad GedCom for others to find only encourages repetition -- and repeated often enough it starts to appear true, which this now does -- but that doesn't make it so.
https://www.ancestry.com/interactive/3824/gpc_newenglandmarriages-0...
