Skip to main content

数据科学中的 R 语言

第 52 章 双样本t检验

本章需要的宏包，希望大家提前安装

install.packages(c("bayesplot", "palmerpenguins", "rstatix", "broom", "ggstatsplot", "infer", "ggthemes"))

52.1 实验设计

研究某种药物的疗效，一般采用大样本随机双盲对照试验，比较在特定条件下被试的反应，获取相关数据后，会进行组内比较或者组间比较：

  • 组内比较， 同一组人，每个人要完成多次测量（重复测量），比如服药第一天的情况，服药第二天的情况，服药第三天的情况…，每组的人数是恒定的。

  • 组间比较，A组的被试吃1mg，B组被试吃2mg, C组吃3mg…，每组的人数不要求是恒定的。

这个过程可能会使用two sample t-tests。

52.2 提问

我们以企鹅体征数据作为案例，假定企鹅就是我们的被试

library(tidyverse)
theme_set(bayesplot::theme_default())

penguins <- palmerpenguins::penguins %>%
  drop_na()

提出问题：

  • 企鹅有男女两种性别(female, male)，不同性别的bill_length_mm的均值是否相同？

  • 企鹅种类有三种(Adelie, Chinstrap, Gentoo)，比较在每个种类下男企鹅和女企鹅bill_length_mm的均值？

  • 两两比较不同种类的bill_length_mm的均值？

52.2.1 不同性别的嘴峰长度的均值是否相同

强烈推荐大家先可视化探索

penguins %>%
  ggplot(aes(x = sex, y = bill_length_mm)) +
  geom_boxplot() +
  geom_jitter() +
  theme(legend.position = "none")

接着简单计算，不同性别bill_length_mm均值以及差值

penguins %>%
  group_by(sex) %>%
  summarize(avg_rating = mean(bill_length_mm, na.rm = TRUE)) %>%
  mutate(diff_means = avg_rating - lag(avg_rating))
## # A tibble: 2 × 3
##   sex    avg_rating diff_means
##   <fct>       <dbl>      <dbl>
## 1 female       42.1      NA   
## 2 male         45.9       3.76

52.2.1.1 using t.test() 

t.test(
  bill_length_mm ~ sex, 
  data = penguins, 
  var.equal = TRUE    # `var.equal = ` 假定两个样本方差是否相等
)
## 
##  Two Sample t-test
## 
## data:  bill_length_mm by sex
## t = -6.667, df = 331, p-value = 1.094e-10
## alternative hypothesis: true difference in means between group female and group male is not equal to 0
## 95 percent confidence interval:
##  -4.866557 -2.649027
## sample estimates:
## mean in group female   mean in group male 
##             42.09697             45.85476
t.test(
  bill_length_mm ~ sex, 
  data = penguins, 
  var.equal = TRUE
) %>%
  broom::tidy()
## # A tibble: 1 × 10
##   estimate estimate1 estimate2 statistic  p.value parameter conf.low conf.high
##      <dbl>     <dbl>     <dbl>     <dbl>    <dbl>     <dbl>    <dbl>     <dbl>
## 1    -3.76      42.1      45.9     -6.67 1.09e-10       331    -4.87     -2.65
## # ℹ 2 more variables: method <chr>, alternative <chr>

52.2.1.2 using rstatix::t_test()

rstatix宏包提供了类似dplyr风格的语法

## 
## Attaching package: 'rstatix'
## The following object is masked from 'package:stats':
## 
##     filter
penguins %>%
  rstatix::t_test(bill_length_mm ~ sex, var.equal = TRUE)
## # A tibble: 1 × 8
##   .y.            group1 group2    n1    n2 statistic    df        p
## * <chr>          <chr>  <chr>  <int> <int>     <dbl> <dbl>    <dbl>
## 1 bill_length_mm female male     165   168     -6.67   331 1.09e-10

52.2.1.3 using ggstatsplot::ggbetweenstats()

探索性数据分析，将包含数据可视化和统计建模两个阶段，可视化为建模提供依据，模型反过来又可以提出不同的可视化方法。ggstatsplot将这两个阶段统一在图形中，即绘制带有统计检验信息的图形，提高数据探索的速度和效率。

## Warning: package 'ggstatsplot' was built under R version 4.2.3
## You can cite this package as:
##      Patil, I. (2021). Visualizations with statistical details: The 'ggstatsplot' approach.
##      Journal of Open Source Software, 6(61), 3167, doi:10.21105/joss.03167
penguins %>% 
  ggbetweenstats( 
    x = sex,
    y = bill_length_mm,
    pairwise.comparisons = TRUE,
    pairwise.display = "all",
    var.equal = TRUE
  )

52.2.1.4 using infer: 基于模拟的检验

Hypothesis Testing Framework

图 52.1: Hypothesis Testing Framework

  • 实际观察的差别
## Warning: package 'infer' was built under R version 4.2.2
## 
## Attaching package: 'infer'
## The following objects are masked from 'package:rstatix':
## 
##     chisq_test, prop_test, t_test
obs_diff <- penguins %>%
  specify(formula = bill_length_mm ~ sex) %>%
  calculate(
    stat = "diff in means",
    order = c("male", "female")
  )
obs_diff
## Response: bill_length_mm (numeric)
## Explanatory: sex (factor)
## # A tibble: 1 × 1
##    stat
##   <dbl>
## 1  3.76
  • 模拟
null_dist <- penguins %>%
  specify(formula = bill_length_mm ~ sex) %>%
  hypothesize(null = "independence") %>%
  generate(reps = 5000, type = "permute") %>% 
  calculate(
    stat = "diff in means",
    order = c("male", "female")
  )
head(null_dist)
## Response: bill_length_mm (numeric)
## Explanatory: sex (factor)
## Null Hypothesis: independence
## # A tibble: 6 × 2
##   replicate   stat
##       <int>  <dbl>
## 1         1  0.946
## 2         2 -0.321
## 3         3 -1.09 
## 4         4  0.126
## 5         5  0.341
## 6         6 -1.32

  1. specify() 指定解释变量和被解释变量 (y ~ x)

  2. hypothesize() 指定零假设 (比如, independence= y 和 x 彼此独立)

  3. generate() 从基于零假设的平行世界中抽样:

    • reps，指定抽样次数
    • type，指定重抽样的类型。

  4. calculate() 计算每次抽样的统计值 (stat = "diff in means")

  • 可视化
null_dist %>%
  visualize() +
  shade_p_value(obs_stat = obs_diff, direction = "both")
  • 计算p值
pvalue <- null_dist %>%
  get_pvalue(obs_stat = obs_diff, direction = "two_sided")
## Warning: Please be cautious in reporting a p-value of 0. This result is an
## approximation based on the number of `reps` chosen in the `generate()` step.
## See `?get_p_value()` for more information.
pvalue
## # A tibble: 1 × 1
##   p_value
##     <dbl>
## 1       0

52.2.1.5 using lm() 

model <- lm(bill_length_mm ~ sex, data = penguins)
broom::tidy(model)
## # A tibble: 2 × 5
##   term        estimate std.error statistic   p.value
##   <chr>          <dbl>     <dbl>     <dbl>     <dbl>
## 1 (Intercept)    42.1      0.400    105.   2.18e-256
## 2 sexmale         3.76     0.564      6.67 1.09e- 10
confint(model)
##                 2.5 %    97.5 %
## (Intercept) 41.309431 42.884509
## sexmale      2.649027  4.866557

可以看到，95%的置信区间与用t.test()的结果完全一样。

52.2.2 每个种类下男企鹅和女企鹅bill_length_mm的均值

企鹅种类有三种，比较在每个种类下男企鹅和女企鹅bill_length_mm的均值？意思是多次t-test

penguins %>%
  ggplot(aes(x = species, y = bill_length_mm, color = sex)) +
  geom_boxplot(position = position_dodge(0.8)) +
  geom_jitter(
    position = position_jitterdodge()
  ) +
  scale_x_discrete(
    expand = expansion(mult = c(0.3, 0.3))
  ) +
  theme(legend.position = "none")

52.2.2.1 using group_modify() + t.test() 

penguins %>%
  group_by(species) %>%
  group_modify(
    ~ t.test(bill_length_mm ~ sex, data = .x, var.equal = TRUE) %>%
      broom::tidy()
  )
## # A tibble: 3 × 11
## # Groups:   species [3]
##   species   estimate estimate1 estimate2 statistic  p.value parameter conf.low
##   <fct>        <dbl>     <dbl>     <dbl>     <dbl>    <dbl>     <dbl>    <dbl>
## 1 Adelie       -3.13      37.3      40.4     -8.78 4.44e-15       144    -3.84
## 2 Chinstrap    -4.52      46.6      51.1     -7.57 1.53e-10        66    -5.71
## 3 Gentoo       -3.91      45.6      49.5     -8.82 1.30e-14       117    -4.79
## # ℹ 3 more variables: conf.high <dbl>, method <chr>, alternative <chr>

52.2.2.2 using rstatix::t_test() 

library(rstatix)

penguins %>%
  group_by(species) %>%
  rstatix::t_test(bill_length_mm ~ sex, var.equal = TRUE)
## # A tibble: 3 × 9
##   species   .y.            group1 group2    n1    n2 statistic    df        p
## * <fct>     <chr>          <chr>  <chr>  <int> <int>     <dbl> <dbl>    <dbl>
## 1 Adelie    bill_length_mm female male      73    73     -8.78   144 4.44e-15
## 2 Chinstrap bill_length_mm female male      34    34     -7.57    66 1.53e-10
## 3 Gentoo    bill_length_mm female male      58    61     -8.82   117 1.3 e-14

52.2.2.3 using ggstatsplot::grouped_ggbetweenstats() 

library(ggstatsplot)

penguins %>% 
  grouped_ggbetweenstats(
    x = sex,
    y = bill_length_mm,
    pairwise.comparisons = TRUE,
    pairwise.display = "all",
    var.equal = TRUE,
    grouping.var = species  # group
  )

52.2.3 两两比较不同种类的bill_length_mm的均值

企鹅种类有三种，两两比较不同种类的bill_length_mm的均值。

  • Adelie - Chinstrap
  • Adelie - Gentoo
  • Chinstrap - Gentoo
t.test(bill_length_mm ~ species, data = penguins) 
## Error in t.test.formula(bill_length_mm ~ species, data = penguins): grouping factor must have exactly 2 levels

species 有三组，也就说有三个层级，程序不接受。方法是：成对pairwise t-tests

52.2.3.1 using pairwise.t.test() 

pairwise.t.test(x, y) # x is a vector of the data, y is the group factor
pairwise.t.test(
  penguins$bill_length_mm, penguins$species,
  alternative = "two.sided",
  paired = FALSE,     
  p.adj = "holm"
) %>%
  broom::tidy()
## # A tibble: 3 × 3
##   group1    group2     p.value
##   <chr>     <chr>        <dbl>
## 1 Chinstrap Adelie    5.05e-70
## 2 Gentoo    Adelie    3.16e-73
## 3 Gentoo    Chinstrap 5.37e- 3

注意：pairwise t-tests并不是简单地把每一个可能的配对都做一次t-test

penguins %>%
  filter(species %in% c("Gentoo", "Chinstrap")) %>%
  t.test(bill_length_mm ~ species, data = .) %>%
  broom::tidy()
## # A tibble: 1 × 10
##   estimate estimate1 estimate2 statistic p.value parameter conf.low conf.high
##      <dbl>     <dbl>     <dbl>     <dbl>   <dbl>     <dbl>    <dbl>     <dbl>
## 1     1.27      48.8      47.6      2.56  0.0117      131.    0.286      2.25
## # ℹ 2 more variables: method <chr>, alternative <chr>

52.2.3.2 using rstatix::pairwise_t_test() 

library(rstatix)

penguins %>%
  pairwise_t_test(
    bill_length_mm ~ species,
    p.adjust.method = "holm",
    alternative = "two.sided",
    paired = FALSE 
  )
## # A tibble: 3 × 9
##   .y.          group1 group2    n1    n2        p p.signif    p.adj p.adj.signif
## * <chr>        <chr>  <chr>  <int> <int>    <dbl> <chr>       <dbl> <chr>       
## 1 bill_length… Adelie Chins…   146    68 2.52e-70 ****     5.05e-70 ****        
## 2 bill_length… Adelie Gentoo   146   119 1.05e-73 ****     3.16e-73 ****        
## 3 bill_length… Chins… Gentoo    68   119 5.37e- 3 **       5.37e- 3 **

52.2.3.3 using ggstatsplot::ggbetweenstats() 

penguins %>%
  ggstatsplot::ggbetweenstats( 
    x = species, 
    y = bill_length_mm,
    pairwise.comparisons = TRUE,
    pairwise.display = "all",
    p.adjust.method = "holm",
    messages = FALSE,
    var.equal = TRUE,
    alternative = "two.sided",
    ggtheme = ggthemes::theme_economist(),
    package = "wesanderson",
    palette = "Darjeeling1"
  )