The study’s method is straightforward. The data come from nearly 36,000 “daily diaries”, self-reporting on how Americans spent their working hours, collected from 2003 to 2012. Relying on the assumption that workers are equally honest in admitting sloth, the authors calculate the fraction of time spent not working while on the job—spent relaxing or eating, say—and find that it varies by race to a small but statistically significant degree. The gap remains, albeit in weaker form, even with the addition of extensive controls for geography, industry and union status, among others. Non-white male workers spend an additional 1.1% of the day not working while on the job, or an extra five minutes per day. Assuming their controls are adequate, that would still leave 90% of the wage difference between white workers and ethnic minorities, which was recently estimated to be 14%, unexplained. This could conceivably be the product of discrimination, or of something else.
研究采用的方法简单直接。数据来自2003-2012年的近36000份“日常日记”,日记的作者们描述了美国人如何支配日常工作时间。在假定员工们都同样诚实地承认偷懒的前提下,研究人员计算了员工在工作时间内未工作(休息或吃零食)的时间占比,其结果是,不同人种的偷懒时间存在差异,其数值不大却具有统计学意义。即使加入地理位置、行业和工会状况等各种控制条件的制约,差距会变小但却依然存在。非白种人的男性员工每天在工作时间内的偷懒时间要长出1.1%,相当于每天多出5分钟。假设他们的控制条件相同,那么白种人员工和少数族裔之间工资差距(最近的估计结果是14%)剩余的九成仍然原因未知。这一结论明显会被当作种族歧视的范例,或者是出于其他目的。