| 查看: 336 | 回复: 0 | ||
| 【悬赏金币】回答本帖问题，作者irunhua将赠送您 5 个金币 | ||
[求助]
论栏里的牛人，遇到四个专业统计问题，能帮忙指点解决吗
|
||
|
论栏里的牛人，投稿PLOS one遇到四个专业统计问题，能帮忙指点解决吗 1. In line 205, “All variables that were significantly associated with the outcome variable at P < 0.20 in the bivariate logistic regression were included in a multivariate logistic regression model, wherein we calculated adjusted odds ratios (AORs) and 95% CIs.” However, it seemed to be inconsistent that not all the AOR and 95%CIs for those variables with P < 0.20 in the bivariate logistic regression were showed in table 2 and 3, such as Education level, Monthly income, and Sexual orientation in table 2. This means the model built by the author has big problem, and I do not think this paper should be published before the author correct this big issue. 2. final multivariable should be refined by keeping only significant predictors to avoid inflated variance, using stepwise backward selection procedures. 3. Adding another layer of multivariable analysis related to HIV and syphilis infection in the context of CAI and testing may strengthen this paper, as the current result presentation may more fit as a short research report other than a full research article. 4. Methods used to build a final model (selection of only those variables with statistically significant univariable associations) are not widely endorsed in the health sciences community; refer to Zalpuri S et al. Association vs. causality in transfusion medicine: understanding multivariable analysis in prediction vs. etiologic research. Transfus Med Rev. 2013 Apr;27(2):74–81. |
回复此楼» 猜你喜欢
这种情况还有戏吗
已经有11人回复
-
微信指数没变化，科研之友没阅读
已经有4人回复
-
同事接到电话了，我却没有
已经有4人回复
-
准备明年的基金了
已经有5人回复
-
今年的WR进展到哪一步了？
已经有6人回复
-
时间戳他又来了
已经有11人回复
-
小木虫看见有人已经知道结果了
已经有15人回复
-
这个自发加氧反应的机理是什么？
已经有6人回复
-
J-J-W不为人知的一面
已经有18人回复
-
没消息就是被刷了呗
已经有13人回复