Despite several studies there’s no evidence that mail in voting favors one party over another.
Here’s the Brookline’s Institute, a pretty neutral bunch, on fraud: https://www.brookings.edu/blog/fixgov/2020/06/02/low-rates-o...
Some analysis from first principles on both fraud and bias: https://fivethirtyeight.com/features/there-is-no-evidence-th...
This non partisan group did detailed analysis of two elections, all of Colorado in 2014 and of Utah in 2016 looking at turnout and bias. https://voteathome.org/
If you want more academic work there’s quite a bit of it done by economics and political science profs. Just search for “mail vote bias”