我正在进行逻辑回归,我的自变量之一是分类的(年份:2010、2012、2016)。我将2010设置为参考变量。但是,当我运行回归时,输出会降低2012年的水平……以前有人见过这个,还是知道为什么它会降低其中一个水平比较?谢谢!
更新为显示自变量和完整的输出
变量设置:
year <- factor(year)
year <- relevel(year, ref="2010")
gender[gender==2] <- 0 # female
gender <- factor(gender)
> table(data$Year, data$A1_Gender)
1 2
2010 104 13
2012 99 11
2016 115 16
输出:
> log <- glm(d2~year + gender, family="binomial"(link="logit"))
> summary(log)
Call:
glm(formula = d2 ~ year + gender, family = binomial(link = "logit"))
Deviance Residuals:
Min 1Q Median 3Q Max
-0.9989 -0.9441 -0.9339 1.4302 1.4422
Coefficients:
Estimate Std. Error z value Pr(>|z|)
(Intercept) -0.43561 0.52008 -0.838 0.402
year2016 -0.02682 0.31424 -0.085 0.932
gender1 -0.14151 0.51136 -0.277 0.782
(Dispersion parameter for binomial family taken to be 1)
Null deviance: 232.52 on 177 degrees of freedom
Residual deviance: 232.44 on 175 degrees of freedom
(181 observations deleted due to missingness)
AIC: 238.44
Number of Fisher Scoring iterations: 4
来自用户20650建议的结果
> md = na.omit(ca[c("d2", "year", "gender")])
> table(md$year)
2016 2010 2012
98 0 80
user20650是正确的,我有一个缺少数据的问题。谢谢大家的帮助和建议!