See this report. A snippet:

Two political science professors — Jonathan Katz of the California Institute of Technology and Anthony Gill of the University of Washington — broke down the data for the GOP.
Both subtracted improper votes from the candidates in proportion to the overall vote each received. When possible, precinct returns were used to establish the pattern; otherwise, county results set the percentages. For example, if Gregoire received 60 percent of a vote in Precinct A and Rossi 40 percent, and there were 100 improper votes from that precinct, Gregoire’s total would be reduced by 60 votes and Rossi’s by 40.
In most cases, the number of improper votes traceable to a single precinct is very small — often, as few as one. In that case, using the same percentages, Gregoire would lose 0.6 votes and Rossi 0.4.
Katz analyzed 1,053 allegedly improper votes statewide and concluded that eliminating them from the count would yield a roughly 100-vote margin for Rossi. (The total of improper votes in Katz’s report is less than the total of claimed by the GOP in separate documents.)
Gill focused on votes from felons that could be tied to individual precincts. Subjecting the 879 votes in his data base to the statistical analysis would produce a Rossi edge of 191.38 votes, Gill said.

It is unclear to me from the article whether the GOP is using the argument to argue that the court should order a new election based on this data, or should reverse the outcome of the election and declare Rossi the winner. If it is the latter, I think this would be a terrible precedent, very much like what happened in the notorious Compton, California election challenge a few years ago, where based on some flimsy social science statistical data about the “ballot order effect,” a trial judge reversed the outcome of a mayoral race. Fortunately, the trial court decision was quickly reversed on appeal. [Disclosure: I consulted for the city of Compton in that litigation.]