R为每个参与者的每个条件选择n行

Question

我有关于多个单一研究的数据框架.对三名参与者(ID_AB2)实施行为干预,并收集基线和干预条件的数据(PhaseAB2:A =基线; B =干预).最后,"Occassions"是一个包含会话编号的变量.

我试图做的是编写一个代码来获取每个参与者的每个阶段的最后三个值,然后计算平均值.

 RateAB2 <- c(8, 4, 5, 5, 5, 4, 8, 7, 8, 5, 4, 8, 7, 4, 7, 7, 4, 3, 4, 4, 4, 5, 9, 6, 16, 13, 8, 9, 15, 17, 9, 10, 7, 3, 2, 4, 3, 3, 3, 7, 13, 18, 12, 18, 14, 17, 19, 15) %>% as.numeric()
 PhaseAB2 <- rep(c("A", "B"), each = 8, len = 48)
 OccasionsAB2 <- rep(1:16, len = 48)
 ID_AB2 <- rep(c("C1", "C2", "C3"), each = 16)
 db5 <- data.frame(ID_AB2, OccasionsAB2, PhaseAB2, RateAB2)

我正在考虑使用,dbdplyr::filter(OccasionsAB2 == ...)但是代码将严格依赖于每个特定的数据集,并且它不能为具有不同观察数量的参与者选择不同的位置.

谢谢你的帮助!

Answer 1

library(dplyr)

RateAB2 <- c(8, 4, 5, 5, 5, 4, 8, 7, 8, 5, 4, 
             8, 7, 4, 7, 7, 4, 3, 4, 4, 4, 5, 
             9, 6, 16, 13, 8, 9, 15, 17, 9, 
             10, 7, 3, 2, 4, 3, 3, 3, 7, 13, 
             18, 12, 18, 14, 17, 19, 15) %>% 
                as.numeric()
PhaseAB2 <- rep(c("A", "B"), each = 8, len = 48)
OccasionsAB2 <- rep(1:16, len = 48)
ID_AB2 <- rep(c("C1", "C2", "C3"), each = 16)
db5 <- data.frame(ID_AB2, OccasionsAB2, PhaseAB2, RateAB2)

db5 %>%
  group_by(ID_AB2, PhaseAB2) %>%       # for each ID and Phase
  top_n(3, OccasionsAB2) %>%           # keep last 3 occasions
  summarise(MEAN = mean(RateAB2)) %>%  # get the average
  ungroup()                            # forget the grouping

# # A tibble: 6 x 3
#   ID_AB2 PhaseAB2  MEAN
#   <fct>  <fct>    <dbl>
# 1 C1     A         6.33
# 2 C1     B         6   
# 3 C2     A         6.67
# 4 C2     B        12   
# 5 C3     A         4.33
# 6 C3     B        17 

您可以将summarise过去更新为

summarise(MEAN = mean(RateAB2), OccasionSequence = paste0(OccasionsAB2, collapse = ","))

还获得每种情况下使用的场合的ID(作为字符串).