最近看到吧友时常提中国强 j 率高达20%这个数据，所以我觉得有必要分析一下。
中国强 j 率高达20%以上这个结论是United Nations Office on Drugs and Crime调查得出的，调查报告是
Prevalence of and factors associated with non-partner rapeperpetration: findings from the UN Multi-country Cross-sectional Study on Men and Violence in Asia and the Pacific
我看了调查报告，认为中国强 j 率高达20%以上这个数据不可信。
首先，这份调查的样本太小了。在每个国家只选取 大约 800到 1800个人 。放在中国这么个人口大国，采样率差不多是一百万分之一。
其次，数据分析时采用multinomial regression models，我个人不知道这是什么，但是凡是涉及模型的分析必然还牵涉到模型的准确性和适用性，那么又多了两个疑点。
最后，这个调查是采用询问男人“是否实施过强j”的办法来统计强 j 率的。那么又涉及到男人是否说实话的问题，这么看起来数据的水分就更大了。
________________________________________________________
这是关于 UNODC的调查采样方法以及建模方法的描述 :
Methods
In this cross-sectional study, undertaken in January 2011–December 2012, for each site we chose a multistage representative sample of households and interviewed one man aged 18–49 years from each. Men self-completedquestions about rape perpetration. We present multinomial regression models of factors associated with single and multiple perpetrator rape and multivariable logistic regression models of factors associated with perpetration of malerape with population-attributable fractions.
Finding
We interviewed 10 178 men in our study (815–1812 per site). The prevalence of non-partner single perpetratorrape varied between 2·5% (28/1131; rural Bangladesh) and 26·6% (225/846; Bougainville, Papua New Guinea),multiple perpetrator rape between 1·4% (18/1246; urban Bangladesh) and 14·1% (119/846; Bougainville, Papua NewGuinea), and male rape between 1·5% (13/880; Jayapura, Indonesia) and 7·7% (65/850; Bougainville, Papua NewGuinea). 57·5% (587/1022) of men who raped a non-partner committed their fi rst rape as teenagers. Frequent reasonsfor rape were sexual entitlement (666/909; 73·3%, 95% CI 70·3–76·0), seeking of entertainment (541/921; 58·7%,55·0–62·4), and as a punishment (343/905; 37·9%, 34·5–41·4). Alcohol was a factor in 249 of 921 cases (27·0%,95% CI 24·2–30·1). Associated factors included poverty, personal history of victimisation (especially in childhood),low empathy, alcohol misuse, masculinities emphasising heterosexual performance, dominance over women, andparticipation in gangs and related activities. Only 443 of 1933 men (22·9%, 95% CI 20·7–25·3) who had committedrape had ever been sent to prison for any period.
中国强 j 率高达20%以上这个结论是United Nations Office on Drugs and Crime调查得出的，调查报告是
Prevalence of and factors associated with non-partner rapeperpetration: findings from the UN Multi-country Cross-sectional Study on Men and Violence in Asia and the Pacific
我看了调查报告，认为中国强 j 率高达20%以上这个数据不可信。
首先，这份调查的样本太小了。在每个国家只选取 大约 800到 1800个人 。放在中国这么个人口大国，采样率差不多是一百万分之一。
其次，数据分析时采用multinomial regression models，我个人不知道这是什么，但是凡是涉及模型的分析必然还牵涉到模型的准确性和适用性，那么又多了两个疑点。
最后，这个调查是采用询问男人“是否实施过强j”的办法来统计强 j 率的。那么又涉及到男人是否说实话的问题，这么看起来数据的水分就更大了。
________________________________________________________
这是关于 UNODC的调查采样方法以及建模方法的描述 :
Methods
In this cross-sectional study, undertaken in January 2011–December 2012, for each site we chose a multistage representative sample of households and interviewed one man aged 18–49 years from each. Men self-completedquestions about rape perpetration. We present multinomial regression models of factors associated with single and multiple perpetrator rape and multivariable logistic regression models of factors associated with perpetration of malerape with population-attributable fractions.
Finding
We interviewed 10 178 men in our study (815–1812 per site). The prevalence of non-partner single perpetratorrape varied between 2·5% (28/1131; rural Bangladesh) and 26·6% (225/846; Bougainville, Papua New Guinea),multiple perpetrator rape between 1·4% (18/1246; urban Bangladesh) and 14·1% (119/846; Bougainville, Papua NewGuinea), and male rape between 1·5% (13/880; Jayapura, Indonesia) and 7·7% (65/850; Bougainville, Papua NewGuinea). 57·5% (587/1022) of men who raped a non-partner committed their fi rst rape as teenagers. Frequent reasonsfor rape were sexual entitlement (666/909; 73·3%, 95% CI 70·3–76·0), seeking of entertainment (541/921; 58·7%,55·0–62·4), and as a punishment (343/905; 37·9%, 34·5–41·4). Alcohol was a factor in 249 of 921 cases (27·0%,95% CI 24·2–30·1). Associated factors included poverty, personal history of victimisation (especially in childhood),low empathy, alcohol misuse, masculinities emphasising heterosexual performance, dominance over women, andparticipation in gangs and related activities. Only 443 of 1933 men (22·9%, 95% CI 20·7–25·3) who had committedrape had ever been sent to prison for any period.
