This study reports on the prevalence of racist attitudes in Germany, comparing a conventional survey question with a list experiment (LE). In a nationally representative survey, respondents were asked whether they oppose the nomination of athletes of color for German national teams. Given that opposition to athletes of color indicates a racist attitude, the questions arise whether social desirability could bias responses in a conventional survey question and whether this is less the case in an LE format. The two methods yield very different prevalence rates: 12.5% of Germans ( = one in eight respondents) expressed a racist attitude in the covert LE format, but only 1.4% ( = one in 71 respondents) in the conventional format. This is a strong indication that social desirability leads to massive underreporting of racist attitudes in surveys. Issues of race and ethnicity in national representation in elite sports are apparently more controversial and politicized than they appear at first glance.