Research by academics, professional organizations, and businesses on ethics in the workplace often relies on surveys that ask employees to report how frequently they have observed others engaging in unethical behavior. But what do these frequencies in observer-reports say about the frequencies of committed unethical behavior? This paper is the first to address this question by empirically exploring the relationship between observer- and self-reports. Our survey research among the Swiss working population shows that for all 37 different forms of unethical behavior investigated, observer-reports show higher frequencies than self-reports. Ratios of observer- to self-reports vary substantially among these forms, ranging from 1.46 for improperly gathering competitor's confidential information to 6.4 for engaging in (sexual) harassment or creating a hostile work environment. The results indicate that researchers should not assume that the frequency in self-reports can generally be approximated by the frequency in observer-reports. Four possible explanations are presented for the differences in ratios, with recommendations for future research.