A recent Implicit Association Test study using political parties found unconscious partisan biases were about 1.5 times as strong as unconscious racial biases.

Judged as of 2014-09-30. True: a recent study (for example Iyengar & Westwood, 2014/2015) that ran an Implicit Association Test with political parties found implicit partisan bias about 1.5 times as strong as implicit racial bias (an effect-size ratio of roughly 1.3–1.7). Mostly true: the study found partisan implicit bias clearly stronger than racial implicit bias, but the ratio is somewhat outside 'half again' (for example 1.2 or 1.9). Mixed: partisan bias was stronger on some measures and not on others. Mostly false: the study exists but found the two biases about equal, or the ratio comes from a different kind of measure (for example explicit bias). False: no such study exists, or it found racial implicit bias stronger.

Set before the research began.

1. E1 Fear and Loathing across Party Lines: New Evidence on Group Polarization American Journal of Political Science (Shanto Iyengar and Sean J. Westwood; PDF hosted on Stanford course site) · July 2015

   “the black-white difference in implicit bias was substantial, t(740.10) = 11.04, p < .001, but the effect size for race (Cohen’s d = .61) was not nearly as strong as the corresponding effect of party (Cohen’s d = .95).”

2. E2 Fear and Loathing across Party Lines: New Evidence on Group Polarization American Journal of Political Science (Shanto Iyengar and Sean J. Westwood; PDF hosted on Stanford course site) · July 2015

   “Our evidence demonstrates that hostile feelings for the opposing party are ingrained or automatic in voters’ minds, and that affective polarization based on party is just as strong as polarization based on race.”

3. E3 “Political identity is fair game for hatred”: how Republicans and Democrats discriminate Vox · 7 December 2015

   “According to the test, Americans are more automatically partisan than they are automatically racist.”

The published Iyengar & Westwood study [E1] gives Implicit Association Test effect sizes of Cohen's d = .95 for party and d = .61 for race. That is a ratio of about 1.56, within the 1.3–1.7 range for "half again as strong". The paper also states that affective polarization based on party is "just as strong" as polarization based on race [E2], but the implicit-bias effect sizes it reports show a larger effect for party (Cohen's d = .95) than for race (d = .61) [E1]. A Vox article [E3] independently reports that the test found partisan automatic bias stronger than racial automatic bias.