正如@alistaire提到的，当你的例子是可复制的时，帮助总是更容易，但我在这里已经尝试过了。
您的数据（大约）：

df <- data.frame(
  estimate = c(3200, 3400, 2780, 3790, 3100, 3750, 2900, 4900),
  treatment = c("control", "control", "high ee2", "high ee2", "low ee2", 
                "low ee2", "virgin", "virgin"),
  day = rep(c("d14", "d21"), 4)
)

df$ci_up <- df$estimate + 500
df$ci_lo <- df$estimate - 500

以特定顺序绘制x轴的第一个挑战是对因子进行排序，该因子具有here和其他位置的答案。但要做到这一点，请使用factor()：

df$treatment = factor(df$treatment, levels = c("control", "virgin", 
                                               "low ee2", "high ee2"))

现在，绘制数据：

ggplot(df) + 
  geom_point(aes(x = treatment, y = estimate, colour = treatment)) + 
  geom_errorbar(aes(x = treatment, ymin = ci_lo, ymax = ci_up, 
                    colour = treatment), width = 0.2)

但我想你也想根据当天的情况来点菜吧？假设D14是第一个，D21是第二个，我们可以这样排列它们：

df$day = factor(df$day , levels = c("d14", "d21"))

然后绘制：

ggplot(df) + 
  geom_point(aes(x = treatment, y = estimate, colour = treatment, 
                 shape = day), 
             position = position_dodge(width = 0.5)) + 
  geom_errorbar(aes(x = treatment, ymin = ci_lo, ymax = ci_up, 
                    colour = treatment, group = day), width = 0.2,
                position = position_dodge(width = 0.5))

我们不必改变当天的形状，但它可能是有用的。为了保持误差条和点彼此对齐，而不是堆叠在日组上，我们将group = day添加到误差条调用中，并对两者使用position_dodge()。如果你想删除不同的形状，以防止点/误差条在相同的处理中相互堆叠，我们也可以将group = day添加到geom_point()：

ggplot(df) + 
  geom_point(aes(x = treatment, y = estimate, colour = treatment, 
                 group = day), 
             position = position_dodge(width = 0.5)) + 
  geom_errorbar(aes(x = treatment, ymin = ci_lo, ymax = ci_up, 
                    colour = treatment, group = day), width = 0.2,
                position = position_dodge(width = 0.5))