英语 demographic bias 是什么意思

　　我的个人见解是：不同人群偏好。

　　下面我来说一下，为什么给出“不同人群偏好”这样的回答，以及“不同人群偏好”的含义。

　　首先，demographic表示与人口有关的，bias表示“倾斜的”、“不均等的”，如果贬用就指“偏颇”、“不公正”，如果褒用则指“偏爱”、“爱好”。这里对bias的解释是为了方便读者对其含义的理解，不再做进一步的考据。

　　重点是，demographic跟bias放在一起时，用来表示什么呢？

　　我们从“文以致用”入手，以“The demographic bias of email as a survey method in a pediatric emergency population”和“An exploration of demographic bias in a questionnaire survey of hearing-impaired children: Implications for comparisons of children with and without cochlear implants”两篇论文为例，对demographic bias的含义进行揣度：

　　前者在摘要中指出：Email has been considered as a communication medium between patients and clinicians in pediatric emergency departments, but the demographic bias involved in using email has not been fully explored. We developed a paper based survey to explore access, willingness to participate and the demographic bias of email within our parent population.在结论中继续说：We have demonstrated a degree of demographic bias in email access rates, negatively affecting those individuals with lower income, less employment, and lower education. Email based surveys directed at parents in pediatric emergency departments should include questions on income, employment and education in order to permit those who analyze the data to correct for these variables. More research is needed to confirm these findings.提炼一下此处demographic bias大意就是：人们收入、就业、受教育水平不同，对问卷调查邮件的响应也是不同的。

　　后者在摘要中所述如下：

　　OBJECTIVES:Responders to questionnaire surveys, who are self-selecting, are generally accepted to be unrepresentative of the total available population, at least in demographic terms. Since demographic and other variables are known to be predictive of outcome, it is important to understand the extent of that unrepresentativeness when using survey data to report comparisons of outcome. This paper aims (i) to evaluate the extent to which a sample of hearing-impaired children surveyed by postal questionnaire was representative of the population of hearing-impaired children in the United Kingdom (UK), and (ii) to identify demographic differences between children with and without cochlear implants.

　　METHODS:Data from a previously reported total ascertainment of hearing-impaired children in the UK which identified 17,160 with permanent bilateral hearing impairment >40 dB HL were compared with data collected by postal questionnaire for a sample of 3224 children, including 527 with cochlear implants.

　　RESULTS:The sampled children were similar to the ascertained population in gender, age at onset of hearing impairment, and number of additional disabilities, but came from more recent birth cohorts and from more affluent families. Compared with profoundly impaired non-implanted children, implanted children had greater degrees of hearing loss, fewer additional disabilities, a later age of onset, were younger, came from more affluent families, were more likely to use spoken language at home, and to be taught using spoken language only.

　　CONCLUSIONS:Comparisons of outcomes and generalization of results require adjustment for relevant variables to avoid confounding estimates of the effectiveness of interventions including cochlear implantation.

　　除了标题，后者在摘要中并没有再次使用demographic bias这个词组，但是在目的一段中指出：问卷调查的响应者一般不能代表全部群体（即问卷调查的目标群体），目的一就是要用样本推及总体。该文将本次调查数据与之前的数据进行对比，在结果一段中指出不同，并在结论一段中强调要调整相关变量以减少干扰。

　　如果要说在后者中demographic bias意味着什么，大概是由样本推及总体时人口因素所造成的误差吧，说到这里，似乎很难给出一个确切的定义，但是仍然不影响我们去理解这个词组，在编辑上述答案时，不由自主地想到了一个美国总统大选时的统计学故事：

　　1936年，民主党人兰登与时任总统罗斯福竞选下届总统。《文学文摘》杂志承担了选情预测的任务。《文学文摘》曾在1920年、1924年、1928年、1932年连续4届美国总统大选中成功地预测总统归属。1936年，《文学文摘》再次进行民意调查，不同于前几次的调查，这次调查把范围拓展得更广。当时大家都相信，数据越大，预测越准。《文学文摘》寄出1000万份调查问卷，覆盖当时四分之一的选民，最终在两个多月内收到了惊人的240万份回执，在统计完成以后，《文学文摘》宣布，艾尔弗雷德兰登将会以55比41的优势，击败富兰克林·罗斯福赢得大选，另外4%的选民则会零散地投给第三候选人。然而，真实的选举结果与《文学文摘》的预测大相径庭：罗斯福以61比37的压倒性优势获胜。让《文学文摘》脸上挂不住的是，新民意调查的开创者盖洛普仅仅通过一场规模小得多的问卷——一个3000人的问卷调查，得出了准确得多的预测结果：罗斯福将稳操胜券。3000人“小”抽样，居然逆转240万的“大”调查，在于科学地抽样，保证抽样的随机性，根据选民的不同特征，根据职业、年龄、肤色等在3000人的比重，再确定电话访问、邮件访问和街头调查等各种方式所在比例。《文学文摘》的失败在于，取样存在严重偏差，它的调查对象主要锁定为它自己的订户，多集中在中上阶层，样本从一开始就是有偏差的（sample bias），而且兰登的支持者更乐于寄回问卷结果，这使得调查的错误更进了一步。这两种偏差的结合，注定了《读者文摘》调查的失败。

　　通过上面两篇论文与一个统计学小故事，如果要给demographic bias一个定义，那么它是样本偏差的一种，由不同人群所对应的不同偏好对调查问卷的影响使得所获得的样本不能代表总体。

　　提炼成一个词，就是“不同人群偏好”啦。

　　本人是人口学专业，对统计学相关知识使用有所不对的地方，还请统计学相关专业人士指正。

温馨提示：答案为网友推荐，仅供参考

当前网址：https://verywind.cn/ee/v3ef2jr3r.html