For years, Texas lawyers have tried every tactic imaginable to convince the 5th U.S. Circuit Court of Appeals that binding arbitration should not resolve their clients' fates, with little success. But John Crouch is adding a new argument to the debate: statistics.

Samantha Diggs v. Citigroup was fully briefed before the 5th Circuit this month. Crouch represents Samantha Diggs, who alleges that that the American Arbitration Association (AAA) is systematically "biased" against employees in employment disputes. To prove it, Diggs' May 15 brief cites a 2009 affidavit and study performed by Cornell University labor law professor Alexander Colvin. Colvin looked at 1,213 employment awards handled by AAA between 2003 and 2007.

In that affidavit, Colvin concludes, among other things:

• The employee win rate on discrimination allegations was only 21.4 % in arbitrations, compared to 36.4% in federal court and 43.8% in state court.

• In cases where a recovery was obtained, the mean award in arbitration was $109,858 versus $384,223 for federal court litigation and $595,594 in state court litigation.

• Arbitrator bias is particularly likely when the employer is a "repeat player" before the AAA. In such situations, the employee win rate drops to 16.9 % versus a 31.6% win rate when the employer is using the AAA for the first time.

Diggs sued her former employer in U.S. District Court for the Northern District of Texas in Dallas, alleging a violation of the Family and Medical Leave Act (FMLA). Citigroup won its Daubert challenge to the affidavit and study and also won its motion to dismiss and compel arbitration.

But Crouch hopes to convince the 5th Circuit that the statistics don't lie when it comes to judging whether arbitration is a fair alternative to courtroom litigation.

"That's the whole point of the argument is: You get people to sign these contracts with arbitration clauses, because it's supposed to be fair," says Crouch, who is a partner in Dallas' Kilgore & Kilgore.

"If the employer knows . . . that they are almost always going to win in arbitration — and even when they lose, the damages will be very small — if they don't disclose that to the employee, then it's essentially fraud in the inducement to arbitration," he alleges.

"If neither side was aware of these statistics, then it's a mutual mistake of fact. And it's not enforceable. The case law is actually clear," Crouch continues.

Kristin Synder, a shareholder in Dallas' Ogletree, Deakins, Nash, Smoak & Stewart who represents Citigroup, did not return a call seeking comment. But Citigroup denies that it violated the FMLA in its July 16 reply brief at the 5th Circuit. It also notes that a Northern District judge in a different employment law case rejected the Colvin affidavit.

"Dr. Colvin's opinions are not based or tailored to the facts of this case. To the contrary, the Colvin Affidavit was signed on February 2009, over three years before this case was filed," Citigroup states in its brief.

"The trial court [in Diggs] rejected the Colvin Affidavit because it questioned the reliability and relevance of Dr. Colvin's study, because the Affidavit did not take into account a number of variables and because, in her briefing, Diggs took general statements and findings from the Affidavit that were not relevant to this case out of context," the reply brief states.

Yet Colvin says he's continued to analyze AAA data through 2012 "and the story doesn't change very much. I get the same finding for the repeat-player effect. And the outcome looks very similar."

Michael Clark, a spokesman for AAA, declines comment.

"We cannot comment on this case because it could be a possible arbitration with us," Clark says.