gistfile1.txt

library("tidyverse")
library("MASS")

        #GPr   Gpo   #belPr #gmsPr gmspo  belpo
mat <- c(1,   .7,   .15,    .21,    .2,  .14,    # GPr
         .7,   1,   .17,    .38,    .22,   .15,    # Gpo
         .15, .17,   1,     .5,     .41,  .85,    # belPr
         .21, .19,  .38,    1,      .91,   .32,    # gmsPr
         .18, .22,  .41,    .91,    1,    .35,    # gmspo
         .2, .22,   .85,    .32,    .35,  1     # belpr
         )    

S <- matrix(mat, nrow = sqrt(length(mat)), ncol = sqrt(length(mat)))

mus <- c(3, # GPA_pre
         3, # GPA_post,
         4.5, # belonging pre
         4.7, # growth mindset pre
         4.5, # belonging post
         4.7) # growth mindset post


# assemble the data
fake_data <- mvrnorm(n = n_obs, mu = mus, Sigma = S) %>%
  data.frame() %>%
  setNames(c("GPA_pre", "GPA_post", "belonging_pre", "growth_mindset_pre",
             "belonging_post", "growth_mindset_post")) %>%
  mutate(subject_id = 1:nrow(.),
         treatment = rep(c("curriculum A", "curriculum B"), nrow(.)/2))


# now build in some interesting effects

# let's say the treatment has a small effect on growth mindset, belonging, and GPA

GPA_post_B <- fake_data$GPA_post[fake_data$treatment == "curriculum B"]
GPA_effect <- rnorm(nrow(fake_data)/2, mean = .08, sd = .25)
fake_data$GPA_post[fake_data$treatment == "curriculum B"] <- GPA_post_B + GPA_effect


gms_post_B <- fake_data$growth_mindset_post[fake_data$treatment == "curriculum B"]
gms_effect <- rnorm(nrow(fake_data)/2, mean = .16, sd = .25)
fake_data$GPA_pre[fake_data$treatment == "curriculum B"] <- gms_post_B + gms_effect

bel_post_B <- fake_data$belonging_post[fake_data$treatment == "curriculum B"]
bel_effect <- rnorm(nrow(fake_data)/2, mean = .16, sd = .25)
fake_data$GPA_pre[fake_data$treatment == "curriculum B"] <- bel_post_B + bel_effect


# now check

lm(GPA_post ~ treatment, data = fake_data) %>%
  summary()

# Coefficients:
#   Estimate Std. Error t value Pr(>|t|)    
# (Intercept)            2.98240    0.04409  67.645   <2e-16 ***
#   treatmentcurriculum B  0.16082    0.06235   2.579     0.01 *  
#   ---
#   Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
# 
# Residual standard error: 0.9859 on 998 degrees of freedom
# Multiple R-squared:  0.006622,	Adjusted R-squared:  0.005627 
# F-statistic: 6.653 on 1 and 998 DF,  p-value: 0.01004