目的通过分析生活满意度高、低水平的微博用户在微博语言和行为上的差异,归纳出可以区分生活满意度高低的微博特征。方法于2012年10月通过新浪微博招募被试,收集被试的人口学资料、生活满意度和微博特征数据。以生活满意度得分的平均数加减前后一个标准差为条件筛选出高低分组,得到325名高分组和303名低分组被试,对2组被试的各个特征进行差异检验。结果生活满意度总均分为(13.6 ± 9.8)分; 高分组生活满意度为(22.8 ± 2.7)分,低分组为(3.7 ± 2.0)分,差异有统计学意义(P < 0.001); 非参数检验中,高分组(M,QR)特定人称代名词(3.92,2.67) 、第一人称代词单数(2.04,1.29) 、第一人称代词复数(0.06,0.00) 、相对词(9.21, 7.55) 、空间词(3.91, 3.07)和金钱词(0.42, 0.00) 6个语言特征的使用频率均高于低分组的(3.55, 2.39) 、(1.91, 1.02) 、(0.00, 0.00) 、(8.71, 7.23) 、(3.59, 2.84) 、(0.28, 0.00),差异有统计学意义(P < 0.05),高分组排除词的使用频率为(1.76, 1.13),低于低分组的(1.98, 1.22),差异有统计学意义(P < 0.05); 行为特征方面,高分组的互粉数与关注数之比(0.42, 0.23) 、互粉数(103, 55) 、粉丝数(463, 309)和认证数(0.00, 0.20)均高于低分组的(0.35, 0.19) 、(86, 46) 、(415, 221) 、(0.00, 0.00),差异有统计学意义(P < 0.05) 。结论不同生活满意度的微博用户在部分微博语言特征和行为特征上存在差异。

Objective To analyze differences and characteristics in linguistic content and behaviors among the microblog users with different life satisfaction level. Methods The participants were recruited from active microblog users via Sina microblog. The information on demography,life satisfaction level,and online traces of the participants were collected with web crawler. We calculated the mean（ M） and standard deviation（ SD） of life satisfaction score and M± SD value was adopted as the threshold to assign the participants into a high（ 325） and low（ 303） satisfaction group. Then the disparities in the characteristics between the two groups were assessed. Results The overall average life satisfaction score was 13. 6 ± 9. 8 for all the participants; the average score was 22. 8 ± 2. 7 for the participants of high satisfaction group and 3. 7 ± 2. 0 for those of lowsatisfaction group,with a significant difference（ P〈0. 001）. The results of nonparametric test showed that compared to those of lowsatisfaction group,the participants of high satisfaction group used following words more frequently（ mean frequency; quartile range of the frequencies） ： particular personal pronoun（ 3. 92 vs. 3. 55; 2. 67 vs. 2. 39）,first-person singular pronoun（ 2. 04 vs. 1. 91; 1. 29 vs. 1. 02）,first-person plural pronoun（ 0. 06 vs. 0. 00; 0. 00 vs. 0. 00）,relative term（ 9. 21 vs. 8. 71; 7. 55 vs. 7. 23）,special word（ 3. 91 vs. 3. 59; 3. 07 vs. 2. 84）,words for money（ 0. 42 vs. 0. 28; 0. 00 vs. 0. 00）（ all P〈0. 05）; but they used exclusive words less frequently（ 1. 76 vs. 1. 98; 1. 13 vs. 1. 22）（ P〈0. 05）. For the behavior features,the participants of high life satisfaction group had a higher ratio of mutual follower versus following（ mean value; quartile range of the values）（ 0. 42 vs. 0. 35; 0. 23 vs. 0. 19）,a greater number of mutual followers（ 103 vs. 86; 55 vs. 46）,a greater number of followers（ 463 vs. 415; 309 vs. 221）,and a high ratio of verification（ 0. 00 vs. 0. 00; 0. 20 vs. 0. 00） compared with those of participants of lowlife satisfaction group（ all P〈0. 05）. Conclusion There are significant differences in some online linguistic and behavior features between the microblog users with high and lowlife satisfaction.