dissenting.
Pretext discrimination cases inherently involve a question of credibility because they require the determination of whether the stated reason for the adverse action is true or false-in other words, “the defendant’s mens rea.” “Whether the defendant was in fact motivated by discrimination is of course for the finder of fact to decide[.]” Reeves v. Sanderson Plumbing Prods., Inc., 530 U.S. 133, 154, 120 S.Ct. *312097, 2112, 147 L.Ed.2d 105 (2000) (Ginsburg, J., concurring). For this reason, a jury verdict in a pretext discrimination case should rarely be overturned. See id. at 155, 120 S.Ct. at 2112 (absent conclusive determination that discrimination could not be true motivation, “the ultimate question of liability ordinarily should not be taken from the jury” once plaintiff has introduced evidence establishing prima fa-cie case and evidence that employer’s proffered explanation for adverse action was false).
Appellate courts have nothing but written words on a page. A jury has the opportunity to hear the tone, inflection, hesitation, or assuredness of the witnesses. It can evaluate the witnesses’ expressions, gestures, mannerisms, and attitudes. It can sense nervousness, arrogance, honesty, and deception. In short, it can do what we cannot-determine when someone is lying.
In this case, Mr. Williams was required to prove two things to the jury: (1) the stated reason for his termination was a pretext (i.e., not the true reason) and (2) the true reason was discrimination. The jury, after hearing all the witnesses and reviewing all the evidence, simply “did not believe the Baker Hughes witnesses,” and found that the alleged basis for Williams’s termination was not the real reason he was discharged. Rather, the jury found that discrimination was the real reason behind Williams’s termination.
In reversing that jury verdict, the majority found that there was no evidence supporting the jury’s determination that race was a motivating factor in Baker Hughes decision to fire Williams. But there was evidence at trial, viewed in the light most favorable to the verdict, that (1) Baker Hughes acquiesced in the “common practice” of the machinists reporting problems to Moisés Banda, (2) other black machinists in the Fishing Tools department who — prior to Fendle/s and Schulz’s supervision — had good work records, were subjected to increased performance citations after Fendley and Schulz became supervisors, and five were eventually terminated,1 (3) the termination rate of black employees at Baker Hughes’ Fishing Tools section for performance issues was higher than that of white or Hispanic employees, and (4) Williams was replaced by a white employee. The “aggregate effect of all these pieces” of evidence amounts to at least sufficient evidence to rise “to a level that would enable reasonable and fair-minded people to differ in their conclusions” about whether race was a motivating factor in Williams’s termination. See King Ranch v. Chapman, 118 S.W.3d 742, 751 (Tex.2003) (defining when “more than a scintilla of evidence” exists).
The “jury heard all the evidence in context and reached a permissible conclusion about what happened” and there is more than a scintilla of evidence supporting their conclusion. The evidence is therefore legally sufficient to support the verdict. Because the majority has held otherwise, I dissent.
. There was testimony that one worker resigned, but the jury, as fact finder, was free to disbelieve this testimony,