Ad analysis (t-test)

Rationale

Framing theory suggests that the way information is presented can influence how audiences interpret a message. Media messages contain cues that can encourage people to interpret the same information in different ways. In an advertisement, characteristics of the person presenting a product can serve as one of those cues.

This experiment examines whether the physical appearance of an actor in a fast-food advertisement influenced viewers’ judgments about the advertised product. Specifically, the study tests whether viewers estimated different numbers of calories for a sandwich depending on whether the advertisement featured a “Buff” actor or a “Dad bod” actor.

Hypothesis

Average sandwich calorie estimates among viewers of the advertisement featuring the Buff actor will differ from average sandwich calorie estimates among viewers of the advertisement featuring the Dad bod actor.

Variables & method

The dependent variable was the estimated number of calories participants believed were in the advertised sandwich. Calories is a continuous variable. The independent variable was the type of actor featured in the advertisement. Actor is a categorical variable with two groups: Buff and Dad bod.

The study included 100 participants who were randomly divided into two groups of 50. Each group watched one version of an advertisement for a Cajun Grill Chicken Sandwich at McDonald’s. The advertisements were identical except for the physique of the actor. After viewing the advertisement, participants completed a questionnaire that included a question asking them to estimate the number of calories in the sandwich.

An independent-samples t-test was used to determine whether the average calorie estimates differed significantly between viewers who saw the Buff actor and viewers who saw the Dad bod actor.

Results & discussion

The graph below shows the distribution of calorie estimates for each actor group and the average estimate for each group.

The descriptive statistics are shown below.

Actor Group	N	Mean	SD	Minimum	Maximum
Descriptive Statistics
Calorie estimates by actor group
Buff	50	584.26	128.03	316.00	827.00
Dad bod	50	656.28	141.25	358.00	946.00

A Shapiro-Wilk test was used to examine whether the calorie estimates in each group differed significantly from a normal distribution.

Actor Group	W Statistic	p-value
Shapiro-Wilk Normality Test Results
Tests of normality within each actor group
Buff	0.976	0.385
Dad bod	0.985	0.790

The independent-samples t-test was conducted using Welch’s t-test, which does not assume equal variances between the two groups.

Group 1	Group 2	Mean (Group 1)	Mean (Group 2)	t Statistic	Degrees of Freedom	p-value	95% CI (Lower)	95% CI (Upper)
Independent Samples t-Test Results
Welch's t-test (unequal variances assumed)
Buff	Dad bod	584.26	656.28	−2.671	97.069	0.009	−125.53	−18.51

The calorie estimates were higher among participants who viewed the advertisement featuring the Dad bod actor than among participants who viewed the advertisement featuring the Buff actor. The Buff group had an average estimate of approximately 584.26 calories, while the Dad bod group had an average estimate of approximately 656.28 calories.

The Welch independent-samples t-test found that this difference was statistically significant, t(97.069) = -2.671, p = .009. The 95% confidence interval for the difference ranged from approximately -125.53 to -18.51 calories.

Because the p-value was less than .05, the results support the hypothesis that average calorie estimates differed depending on which actor appeared in the advertisement.

Code

# Load packages
library(dplyr)
library(ggplot2)
library(gt)
library(gtExtras)

options(scipen = 999)

# Load data
mydata <- read.csv("SandwichAd.csv")

# Identify dependent and independent variables
mydata <- mydata %>%
  mutate(
    DV = Calories,
    IV = Actor
  )

# Calculate group means
group_means <- mydata %>%
  group_by(IV) %>%
  summarise(
    mean_DV = mean(DV, na.rm = TRUE),
    .groups = "drop"
  )

# Create graph
Graphic <- ggplot(mydata, aes(x = DV)) +
  geom_histogram(
    binwidth = diff(range(mydata$DV, na.rm = TRUE)) / 30,
    color = "black",
    fill = "#1f78b4",
    alpha = 0.7
  ) +
  geom_vline(
    data = group_means,
    aes(xintercept = mean_DV),
    color = "red",
    linetype = "dashed",
    linewidth = 1
  ) +
  facet_grid(IV ~ .) +
  labs(
    title = "DV Distributions by IV Group",
    x = "Estimated Calories",
    y = "Count"
  ) +
  theme_minimal()

Graphic

# Descriptive statistics
Descriptives <- mydata %>%
  group_by(IV) %>%
  summarise(
    N = n(),
    Mean = mean(DV, na.rm = TRUE),
    SD = sd(DV, na.rm = TRUE),
    Minimum = min(DV, na.rm = TRUE),
    Maximum = max(DV, na.rm = TRUE),
    .groups = "drop"
  )

Descriptives

## # A tibble: 2 × 6
##   IV          N  Mean    SD Minimum Maximum
##   <chr>   <int> <dbl> <dbl>   <int>   <int>
## 1 Buff       50  584.  128.     316     827
## 2 Dad bod    50  656.  141.     358     946

# Shapiro-Wilk normality test
shapiro_summary <- mydata %>%
  group_by(IV) %>%
  summarise(
    W = unname(shapiro.test(DV)$statistic),
    p_value = shapiro.test(DV)$p.value,
    .groups = "drop"
  )

shapiro_summary

## # A tibble: 2 × 3
##   IV          W p_value
##   <chr>   <dbl>   <dbl>
## 1 Buff    0.976   0.385
## 2 Dad bod 0.985   0.790

# Welch independent-samples t-test
t_res <- t.test(
  DV ~ IV,
  data = mydata,
  var.equal = FALSE
)

t_res