Rationale

Media priming suggests that content you process can temporarily influence subsequent attitudes and behaviours. In this experiment fifth‑graders were randomly assigned to three different lessons just before recess. One group played a violent exponent game where correct answers let the player punch an on‑screen opponent. Another group played a non‑violent exponent game that rewarded correct answers by adding blocks to build a structure. The third group completed an art project using a colour wheel. After the lessons, observers recorded how many minutes each child spent engaged in aggressive acts during a thirty minute recess. This analysis applies priming theory to see if exposure to violent game content is associated with more aggressive behaviour compared with the non‑violent game or the art project.

Hypothesis

On average, students who played the violent exponent game will spend more time engaged in aggressive acts during recess than students who played the non‑violent exponent game or who completed the art project.

Variables & method

Results & discussion

Descriptive statistics showed that the art project group averaged about 5 minutes of aggression (SD ≈ 1.0), the non‑violent game group averaged about 5.9 minutes (SD ≈ 1.1) and the violent game group averaged about 10.7 minutes of aggression (SD ≈ 1.0). The interactive box plot below displays the distribution of aggression times for each lesson group and includes individual data points for transparency.

Below the box plot, the ANOVA table summarises the omnibus test. The F statistic was very large (F ≈ 410) and the associated p‑value was effectively zero, indicating a statistically significant difference among the group means. Tukey’s HSD comparisons showed that every pair of groups differed significantly: the violent game group spent on average about 5.8 minutes more in aggressive acts than the art project group and about 4.9 minutes more than the non‑violent game group, while the non‑violent game group spent about 0.9 minutes more than the art project group. The effect size was large (η² ≈ 0.86). These results support the hypothesis that exposure to the violent game primed greater aggression during recess.

## ── Attaching core tidyverse packages ──────────────────────── tidyverse 2.0.0 ──
## ✔ dplyr     1.1.4     ✔ readr     2.1.5
## ✔ forcats   1.0.0     ✔ stringr   1.5.1
## ✔ lubridate 1.9.4     ✔ tibble    3.3.0
## ✔ purrr     1.1.0     ✔ tidyr     1.3.1
## ── Conflicts ────────────────────────────────────────── tidyverse_conflicts() ──
## ✖ dplyr::filter() masks plotly::filter(), stats::filter()
## ✖ dplyr::lag()    masks stats::lag()
## ℹ Use the conflicted package (<http://conflicted.r-lib.org/>) to force all conflicts to become errors
## Loading required package: carData
## 
## 
## Attaching package: 'car'
## 
## 
## The following object is masked from 'package:dplyr':
## 
##     recode
## 
## 
## The following object is masked from 'package:purrr':
## 
##     some
One‑way ANOVA results
Term df sumsq meansq statistic p.value
Between groups 2 872.6253 436.312667 409.7299 0
Within groups 132 140.5640 1.064879 NA NA
Tukey HSD pairwise comparisons
Comparison diff lwr upr p_adj
Nonviolent game-Art project 0.900000 0.3843098 1.415690 0.0001831
Violent game-Art project 5.786667 5.2709765 6.302357 0.0000000
Violent game-Nonviolent game 4.886667 4.3709765 5.402357 0.0000000

The homogeneity of variances assumption held reasonably well (Levene’s test p > 0.05), and normality checks did not reveal extreme departures. Nevertheless, Welch’s ANOVA produced a similarly large F value, confirming that the result was not sensitive to the equal variance assumption. In sum, fifth‑graders who played the violent exponent game showed markedly more aggressive behaviour during recess than peers who played a non‑violent exponent game or completed an art project. The findings align with media priming theory and illustrate how brief exposure to violent game content can prime aggressive behaviour, at least in the short term.

Code

The code below reproduces the analysis described above. It loads the necessary packages, reads the data, checks assumptions, runs the ANOVA and Tukey tests, computes effect size, and creates the interactive plot.

## Install packages if they are missing
packages <- c("tidyverse", "plotly", "car", "broom", "effectsize")
to_install <- packages[!(packages %in% installed.packages())]
if(length(to_install)) install.packages(to_install)

library(tidyverse)  # for data manipulation
library(plotly)     # for interactive plots
library(car)        # for Levene's test
library(broom)      # for tidying model outputs
library(effectsize) # for effect size measures

## Read the data
mydata <- read.csv("Priming.csv")
mydata$Group <- as.factor(mydata$Group)  # ensure Group is a factor

## Descriptive statistics
mydata %>%
  group_by(Group) %>%
  summarise(
    count = n(),
    mean  = mean(Value, na.rm = TRUE),
    sd    = sd(Value, na.rm = TRUE),
    min   = min(Value, na.rm = TRUE),
    max   = max(Value, na.rm = TRUE)
  )
## # A tibble: 3 × 6
##   Group           count  mean    sd   min   max
##   <fct>           <int> <dbl> <dbl> <dbl> <dbl>
## 1 Art project        45  4.96 1.04    2.3   7.3
## 2 Nonviolent game    45  5.86 1.06    3.6   8.4
## 3 Violent game       45 10.7  0.998   8.5  12.8
## Normality tests (Shapiro–Wilk per group)
mydata %>%
  group_by(Group) %>%
  summarise(p_value = shapiro.test(Value)$p.value)
## # A tibble: 3 × 2
##   Group           p_value
##   <fct>             <dbl>
## 1 Art project       0.923
## 2 Nonviolent game   0.594
## 3 Violent game      0.699
## Homogeneity of variances (Levene's test)
leveneTest(Value ~ Group, data = mydata)
## Levene's Test for Homogeneity of Variance (center = median)
##        Df F value Pr(>F)
## group   2  0.2387  0.788
##       132
## One‑way ANOVA
anova_model <- aov(Value ~ Group, data = mydata)
summary(anova_model)
##              Df Sum Sq Mean Sq F value Pr(>F)    
## Group         2  872.6   436.3   409.7 <2e-16 ***
## Residuals   132  140.6     1.1                   
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## Welch's ANOVA as a robustness check
oneway.test(Value ~ Group, data = mydata, var.equal = FALSE)
## 
##  One-way analysis of means (not assuming equal variances)
## 
## data:  Value and Group
## F = 420.86, num df = 2.000, denom df = 87.947, p-value < 2.2e-16
## Effect size (eta squared)
eta_squared(anova_model)
## For one-way between subjects designs, partial eta squared is equivalent
##   to eta squared. Returning eta squared.
## # Effect Size for ANOVA
## 
## Parameter | Eta2 |       95% CI
## -------------------------------
## Group     | 0.86 | [0.83, 1.00]
## 
## - One-sided CIs: upper bound fixed at [1.00].
## Tukey HSD post hoc comparisons
TukeyHSD(anova_model)
##   Tukey multiple comparisons of means
##     95% family-wise confidence level
## 
## Fit: aov(formula = Value ~ Group, data = mydata)
## 
## $Group
##                                  diff       lwr      upr     p adj
## Nonviolent game-Art project  0.900000 0.3843098 1.415690 0.0001831
## Violent game-Art project     5.786667 5.2709765 6.302357 0.0000000
## Violent game-Nonviolent game 4.886667 4.3709765 5.402357 0.0000000
## Interactive box plot with plotly
fig <- plot_ly(mydata, x = ~Group, y = ~Value, type = "box", color = ~Group,
               boxpoints = 'all', jitter = 0.3, pointpos = -1.8,
               hovertemplate = paste('Group: %{x}<br>Minutes: %{y}<extra></extra>'))
fig <- fig %>% layout(title = "Aggressive time by lesson group",
                      xaxis = list(title = "Lesson group"),
                      yaxis = list(title = "Minutes of aggressive acts"),
                      showlegend = FALSE)
fig