Erasmus University Rotterdam, The Netherlands.
Soc Sci Res. 2012 Jan;41(1):203-5. doi: 10.1016/j.ssresearch.2011.11.009. Epub 2011 Nov 12.
Inequality of happiness in nations can be measured using the standard deviation of responses to surveys questions. The standard-deviation is not quite independent of the mean, being zero when everybody is maximally happy or unhappy while the possible value of the standard deviation is highest when the mean is in the middle of the response scale. Delhey and Kohler see this intrinsic dependency as a problem and propose two ways to compute 'corrected' standard deviations. I advise against this medicine. One reason is that there is no real disease, since the presumed problem does not occur with commonly used numerical rating scales of 10 or more steps. The second reason is that one of Delhey and Kohler's medicines have side effects, their first correction affects the mean and their second correction is based on implausible assumptions. A third reason is that there are better ways to estimate the effect happiness-inequality net happiness-level. Partialling out mean happiness did not affect the non-correlation between inequality of income and inequality of happiness in an analysis of 116 nations.
可以使用对调查问题的反应的标准差来衡量国家之间的幸福不平等。标准差并不完全独立于平均值，当每个人都处于最大幸福或不幸福状态时，标准差为零，而当平均值处于反应量表的中间时，标准差的可能值最高。德尔海和科勒认为这种内在的依赖性是一个问题，并提出了两种计算“校正”标准差的方法。我不建议使用这种方法。原因之一是没有真正的疾病，因为通常使用的 10 步或更多步骤的数值评分量表不会出现假定的问题。第二个原因是，德尔海和科勒的一种药物有副作用，他们的第一种矫正方法会影响平均值，而第二种矫正方法则基于不合理的假设。第三个原因是，还有更好的方法来估计幸福感不平等对净幸福感水平的影响。在对 116 个国家的分析中，对平均幸福感进行部分消除并没有影响收入不平等与幸福感不平等之间的非相关性。
