对话人物 李建新北京大学社会学系教授、《中国民生发展报告2012》主编。

Conversation LiJianXin Beijing university, professor at the society、《China minsheng development report 2012》editor。

　　8月5日，北京大学发布《中国民生发展报告2012》。报告称，调查显示，去年全国家庭的平均住房面积为116.4平方米，人均住房面积为36平方米。报告一出，舆论哗然。很多网民认为数据有失公允，戏称自己“被平均”“拖后腿”。

August 5,，Issued by the Beijing university《China minsheng development report 2012》。The report says，survey，Last year the national family average housing area is 116.4 square metre，Per capita housing area of 36 square meters。Report a，Public outcry。Many Internet users think that data are being unfair，Calls himself“Be average”“Dragging its heels”。

　　昨天，本报记者专访了报告主编、北京大学社会学系教授李建新。他承认，报告所得的家庭平均住房面积116.4平方米存在缺陷，掩盖了样本的差异。他表示，可能会适时再做调研，细化群体分类。

yesterday，Our reporter interviews, the report、Beijing university, professor at the LiJianXin society。He admitted that he，The report from the average family housing area of 116.4 square meters deficiencies，Cover up a sample of the differences。He said，May be timely to do research，Refining group classification。

住房平均数据掩盖个体差异 Housing average data cover up individual differences

　　京华时报：全国家庭平均住房面积为116.4平方米，这个数据是否客观？

Jinghua times：The national average family housing area is 116.4 square metre，The data are objective？

　　李建新：从8月6日到现在，我接到不少媒体的电话，很多记者对“家庭平均住房面积116.4平方米”和“人均住房面积36平方米”这两个数据提出质疑，说与自己的现实情况出入很大。其实我们报告中所得到的数据只是个概括的平均数，虽然比较简洁，但是也掩盖了我们所取样本内部的结构差异。

LiJianXin：From August 6, until now，I received a lot of media call，Many reporters to“Average family housing area of 116.4 square meters”and“Housing area of 36 square meters per capita”The two data questioned，Said the reality of the situation and himself in and out is very large。In fact we report from the data is just a general average，Although more concise，But, it covered our sampling this internal structural differences。

　　京华时报：您承认差异的存在？

Jinghua times：You admit that differences？

　　李建新：对，差异确实有。这种差异体现在两方面：一个是地域差异，中国幅员辽阔，比如城与乡、城市中心人群与非城市中心人群，农村人均住房面积会比较大，而城市人住房则比较紧张；另一个差异存在于不同的阶层，比如高收入群体和低收入群体的分别，有极端的情况，比如一个家庭拥有几套豪宅。其实在样本内部还是有很大差异的。

LiJianXin：to，Differences do have。This difference embodied in two aspects：One is the regional differences，China is a large country，Such as city and township、The city center with the city center crowd crowd，Rural per capita housing area will be relatively large，While the city housing is more nervous person；Another difference exists in different classes，Such as high income group and low-income groups respectively，With extreme situations，Such as a family has a few sets a person of extraordinary powers curtilage。In fact in the sample internally or have very big difference。

　　网民反响强烈始料未及

Internet users unexpected evokes strong repercussions

　　京华时报：报告发布后，116.4平方米引起了网民强烈的反响，有没有想到会引起这么大的反响？

Jinghua times：After the report，116.4 square meters caused the netizens strong response，Did you come up with would cause so much reaction？

　　李建新：这是始料未及的，我个人参与报告的撰写有三四年，报告提出全国家庭平均住房面积是第一次，这也是我们第一次遇到这么大的反响。

LiJianXin：This is an unexpected hit，I personally participate in writing of the report have three or four years，The report proposes the national average family housing area is the first time，This is the first time we met such a big echo。

　　京华时报：怎么看待这种反响？

Jinghua times：What about this response？

　　李建新：一方面，这个平均数指标是存在缺陷的。网友呼声强烈，这恰恰反映了存在不均衡的情况，但是被平均数掩盖了。

LiJianXin：On the one hand，The average index is the defect。The strong voice，This just reflects the existence of equilibrium，But be average hide。

　　另一方面，这个数据高于一些人的心理预期，恰巧是身为“蚁族”的一些年轻人、低于平均线的这部分网友关注了，传媒的舆论也会有导向，选择性地放大某些差异。实际上，我在微博上也回应了，但是跟“116.4”这个平均数相比，我的回应得到的关注很少。

On the other hand，This data is higher than some people expected，Happened to be as“Ant family”Of some of the young people、Below the average line on this part of the net friend......，The opinion of the media there will also be oriented，Selective amplification certain differences。In fact，I'm in a blog also responded，But with“116.4”Compared to the average，My response to get the attention of the very few。

已意识到算法存在问题 Has already realized algorithm has a problem

　　京华时报：您提到在样本内部存在差异，那么针对不同阶层、地域的人群，有没有细化的调研数据？比如乡镇人均住房面积是多少，城市是多少？

Jinghua times：You mentioned that internal differences in the sample，So according to different class、Regional crowd，Have refined research data？Such as rural per capita housing area is much，City is much？

　　李建新：我们暂时没有考虑到细化的分类。这个《中国民生发展报告》每年都会有，我们关注的是所有的民生问题，包括家庭支出、教育支出、医药卫生等等，而住房只是我们其中的一个方面，我们并没有对其分类，只是反映了大体的情况。

LiJianXin：We'll do not consider the refinement of classification。the《China minsheng development report》Every year there are，We focus on all of the livelihood of the people，Including family spending、Education spending、Medicine and health care, and so on，But housing is only one of us，And we have not classify them，Just reflect the general situation。

　　京华时报：这个数据科学吗？算法和样本的抽样方法是什么？

Jinghua times：This data science？Algorithm and the sample sampling method is??????？

　　李建新：我们是在全国范围内，按照一定的比例抽取了城乡、各阶层的样本，进行调查的。具体的算法就是用调查的总面积除以总的家庭数量，属于比较简单的求平均值算法。至少在现有的方法中，它是比较科学准确的。

LiJianXin：We are in the national scope，According to a certain proportion of the urban and rural extraction、All sectors of the samples，Survey of。Specific algorithm is with the total area of the investigation is divided by the total number of family，Belong to relatively simple averaged algorithm。At least in the existing methods，It is more scientific and accurate。

　　当然，如果横向去比，这个数据和以前的既有数据也没有太大出入，住房和城乡建设部的报告也显示，截至2009年底，中国城市人均住宅建筑面积已达30平方米。

Of course，If the transverse than to go to，The data and the data before both is there much discrepancy，Housing and rural construction report also showed，By the end of 2009，China's urban per capita residential area has reached 30 square meters。

　　不过，我们也开始意识到平均数算法存在的问题，这份报告中的研究并不深入，不可能照顾到全部人群。今后，我们可能会针对人均住房面积的样本进行细化分类，深入研究，适时推出相关的报告。

but，We began to realize that average algorithm existing problems，The report of the study doesn't further，Can't take care of all the crowd。In the future，We can lead to housing area per capita of thinning sample classification，In-depth study，Releasing the reports。