R Coefplot根据治疗组大小

Question

假设您有250个观察值，并且可以根据某些条件定义治疗/控制。

constants = ['(a, b)', '(x, y, z)', 'i']
literals = [eval(k) for k in constants]    # get rid of the strings
part1 = [literals[0]]                      # individually make each of the three parts of your list
part2 = [k for k in range(literals[1][0], literals[1][1] + 1, literals[1][2])]    # or if you don't need to include y then you could just do range(literals[1])
part3 = [literals[2]]
transformed = part1 + part2 + part3

现在，您重新定义了250个观测值，就得到了

Treatment 1 - 100 observations
Control 1 - 150 observations

您估计：

Treatment 2 -- 160 observations
Control 3 -- 90 observations

其中data1和data2对应于定义治疗/控制的两种方式。

绘制回归系数：

m1 <- lm(y~x+treat, data=data1)
m2 <- lm(y~x+treat, data=data2)

我希望系数图显示（例如，通过增加点数），第二个模型中的治疗组比第一个模型大。任何线索将不胜感激。