library(tidyverse)
library(gt)
library(jmv)
library(janitor)

all <- read_csv("all_ratings.csv")

How many “good” participants in each group so far?

all %>% 
  tabyl(group)
##  group  n   percent
##      1 55 0.4583333
##      2 65 0.5416667

summary stats

manipcheck <- all %>%    
  group_by(group) %>%
  summarise(gf_M = mean(manipulation_check_gf_1, na.rm = TRUE),
            gfSD = sd(manipulation_check_gf_1, na.rm = TRUE), 
            gfN = n(), 
           gfStdErr =  gfSD/sqrt(gfN),
            lf_M = mean(manipulation_check_lf_1, na.rm = TRUE),
           lfSD = sd(manipulation_check_lf_1, na.rm = TRUE), 
           lfN = n(), 
            lfStdErr =  lfSD/sqrt(lfN))

manipcheck %>% select(group, gf_M, lf_M, gfN) %>%
  gt()
group gf_M lf_M gfN
1 4.256731 3.319796 55
2 4.102500 3.595893 65

How do other studies check manipulation has worked?

Option 1: within subjects

  • test whether gain participants rate info stressed benefits > stressed negative consequences AND loss participants rate info stressed negative consequences > stressed benefits (within subjects comparison) - Table above looks like both gain and loss participants rate benefits > negatives

Option 2: between subjects

  • test whether gain participants rate that the information stressed benefits more than loss participants? AND loss participants rating that information stressed negative consequences more than gain participants (between subjects comparison), Table above looks like means are in the right direction, but are they sig different?

between subjects

gain frame plot

manipcheck <- manipcheck %>%
  mutate(group = factor(group, 
                        levels = c(1, 2),
                        labels = c("gain", "loss")))

manipcheck %>%
    ggplot(aes(x = group, y = gf_M)) +
    geom_col() +
    scale_y_continuous(limits = c(0,5)) +
    geom_errorbar(aes(ymin=gf_M-gfStdErr, ymax=gf_M+gfStdErr),
                  size=.3, 
                  width=.2) +
  theme_classic() +
  labs(title = "To what extent did the information stress the positive benefits \n of adopting open science practices?", caption = "in right direction but not sig p = .29")

gain frame t-test

ttestIS(formula = manipulation_check_gf_1 ~ group, data = all)
## 
##  INDEPENDENT SAMPLES T-TEST
## 
##  Independent Samples T-Test                                                       
##  ──────────────────────────────────────────────────────────────────────────────── 
##                                              Statistic    df          p           
##  ──────────────────────────────────────────────────────────────────────────────── 
##    manipulation_check_gf_1    Student's t     1.116715    110.0000    0.2665495   
##  ────────────────────────────────────────────────────────────────────────────────

loss frame plot

manipcheck %>%
  ggplot(aes(x = group, y = lf_M)) +
  geom_col() +
  scale_y_continuous(limits = c(0,5)) +
  geom_errorbar(aes(ymin=lf_M - lfStdErr, ymax=lf_M + lfStdErr),
                size=.3, 
                width=.2) +
  theme_classic() +
  labs(title = "To what extent did the information stress the negative consequences \n of not adopting open science practices?", caption = "in right direction, but not sig p = .27")

loss frame t-test

ttestIS(formula = manipulation_check_lf_1 ~ group, data = all)
## 
##  INDEPENDENT SAMPLES T-TEST
## 
##  Independent Samples T-Test                                                       
##  ──────────────────────────────────────────────────────────────────────────────── 
##                                              Statistic    df          p           
##  ──────────────────────────────────────────────────────────────────────────────── 
##    manipulation_check_lf_1    Student's t    -1.154928    103.0000    0.2507921   
##  ────────────────────────────────────────────────────────────────────────────────