Overview

This report presents the findings from the Lone Wolf experimental study examining how experimental condition affects feelings of uncertainty, and whether this effect is moderated by individual difference variables.

Research Question: Does the experimental manipulation affect feelings of uncertainty, and is this effect moderated by narcissism, uncertainty aversion, attachment style, self-control, or self-esteem?

Data Preparation

library(tidyverse)
library(psych)
library(car)
library(effsize)
library(interactions)
library(reghelper)
library(effectsize)
library(emmeans)
library(knitr)
library(kableExtra)

# Load process if available
if(file.exists("process.R")) {
  source("process.R")
}

## 
## **************** PROCESS Procedure for R Version 5.0 ****************** 
##  
##            Written by Andrew F. Hayes, Ph.D.  www.afhayes.com              
##    Documentation available in Hayes (2022). www.guilford.com/p/hayes3   
##  
## *********************************************************************** 
##  
## PROCESS is now ready for use.
## Copyright 2013-2025 by Andrew F. Hayes ALL RIGHTS RESERVED
## Workshop schedule at haskayne.ucalgary.ca/CCRAM 
## Information about PROCESS available at processmacro.org/faq.html

# Load data
the.data <- read.csv("lonewolfog.csv")

# Display sample size
cat("Total sample size:", nrow(the.data), "\n")

## Total sample size: 278

Main Effect: Condition on Feelings of Uncertainty

Descriptive Statistics

# Convert FelUn items to numeric
the.data <- the.data %>%
  mutate(across(FelUn1:FelUn19, as.numeric))

# Calculate FelUn total
the.data <- the.data %>%
  mutate(FelUn_total = rowMeans(across(FelUn1:FelUn19), na.rm = TRUE))

# Descriptive statistics by condition
desc_stats <- the.data %>%
  group_by(Condition) %>%
  summarise(
    N = n(),
    Mean = mean(FelUn_total, na.rm = TRUE),
    SD = sd(FelUn_total, na.rm = TRUE),
    SE = SD / sqrt(N),
    .groups = "drop"
  )

kable(desc_stats, digits = 3, caption = "Feelings of Uncertainty by Condition") %>%
  kable_styling(bootstrap_options = c("striped", "hover"))

Feelings of Uncertainty by Condition
Condition	N	Mean	SD	SE
	23	NaN	NA	NA
0	127	3.056	0.663	0.059
1	126	2.712	0.659	0.059
Condition	1	NaN	NA	NA
{“ImportId”:“Condition”}	1	NaN	NA	NA

Distribution

ggplot(the.data, aes(x = FelUn_total)) +
  geom_histogram(bins = 30, fill = "steelblue", color = "black", alpha = 0.7) +
  facet_wrap(~Condition, labeller = labeller(Condition = c("0" = "Control", "1" = "Treatment"))) +
  labs(title = "Distribution of Feelings of Uncertainty by Condition",
       x = "Feelings of Uncertainty (Mean Score)",
       y = "Count") +
  theme_minimal()

T-Test Results

# T-test
ttest_result <- t.test(FelUn_total ~ Condition, data = the.data)
cohen_d_result <- cohen.d(FelUn_total ~ Condition, data = the.data)

# Display results
cat("Independent Samples t-test\n")

## Independent Samples t-test

cat("=====================================\n")

## =====================================

cat(sprintf("t(%.2f) = %.3f, p = %.4f\n",
            ttest_result$parameter,
            ttest_result$statistic,
            ttest_result$p.value))

## t(249.99) = 4.129, p = 0.0000

cat(sprintf("Cohen's d = %.3f\n", cohen_d_result$estimate))

## Cohen's d = 0.520

cat(sprintf("95%% CI: [%.3f, %.3f]\n",
            ttest_result$conf.int[1],
            ttest_result$conf.int[2]))

## 95% CI: [0.180, 0.508]

Assumptions Checks

# Normality tests
shapiro_0 <- shapiro.test(the.data$FelUn_total[the.data$Condition == "0"])
shapiro_1 <- shapiro.test(the.data$FelUn_total[the.data$Condition == "1"])

cat("Shapiro-Wilk Normality Tests\n")

## Shapiro-Wilk Normality Tests

cat("=====================================\n")

## =====================================

cat(sprintf("Condition 0: W = %.4f, p = %.4f\n", shapiro_0$statistic, shapiro_0$p.value))

## Condition 0: W = 0.9770, p = 0.0300

cat(sprintf("Condition 1: W = %.4f, p = %.4f\n", shapiro_1$statistic, shapiro_1$p.value))

## Condition 1: W = 0.9805, p = 0.0666

# Q-Q plot
par(mfrow = c(1, 2))
qqnorm(the.data$FelUn_total[the.data$Condition == "0"], main = "Q-Q Plot: Condition 0")
qqline(the.data$FelUn_total[the.data$Condition == "0"])
qqnorm(the.data$FelUn_total[the.data$Condition == "1"], main = "Q-Q Plot: Condition 1")
qqline(the.data$FelUn_total[the.data$Condition == "1"])

par(mfrow = c(1, 1))

# Levene's test
levene_result <- leveneTest(FelUn_total ~ Condition, data = the.data)
cat("\nLevene's Test for Equality of Variances\n")

## 
## Levene's Test for Equality of Variances

cat("=====================================\n")

## =====================================

print(levene_result)

## Levene's Test for Homogeneity of Variance (center = median)
##        Df F value Pr(>F)
## group   1  0.0627 0.8025
##       250

Moderator 1: Narcissistic Personality Inventory (NPI)

Scale Reliability

# Define NPI items
authority_items <- paste0("NarcisAuthority", 1:8)
exhibitionism_items <- paste0("NarcisExhib", 1:7)
superiority_items <- c("NarcisSuper1", "NarcisSuper2", "NarcisSuper3",
                       "NarcisSuper4x", "NarcisSuper5")
entitlement_items <- paste0("NarcisEnti", 1:6)
exploitativeness_items <- c("NarcisExpl1", "NarcisExpl2", "NarcisExpl3",
                            "NarcisExpl4", "NarcisExpl6")
vanity_items <- paste0("NarcisVan", 1:3)
selfsuff_items <- paste0("NarcisSelfSuff", 1:6)
all_npi_items <- c(authority_items, exhibitionism_items, superiority_items,
                   entitlement_items, exploitativeness_items, vanity_items,
                   selfsuff_items)

# Convert to numeric and calculate scores
the.data <- the.data %>%
  mutate(across(all_of(all_npi_items), as.numeric)) %>%
  mutate(NPI_Total_mean = rowMeans(select(., all_of(all_npi_items)), na.rm = TRUE))

# Reliability
alpha_npi <- psych::alpha(the.data[, all_npi_items])
cat(sprintf("NPI Total Scale: Cronbach's α = %.3f\n", alpha_npi$total$raw_alpha))

## NPI Total Scale: Cronbach's α = 0.914

Distribution

ggplot(the.data, aes(x = NPI_Total_mean)) +
  geom_histogram(bins = 30, fill = "coral", color = "black", alpha = 0.7) +
  labs(title = "Distribution of Narcissistic Personality Inventory Scores",
       x = "NPI Mean Score",
       y = "Count") +
  theme_minimal()

Moderation Analysis

# Mean-center narcissism
the.data <- the.data %>%
  mutate(Condition = as.numeric(Condition),
         NPI_centered = NPI_Total_mean - mean(NPI_Total_mean, na.rm = TRUE))

# Regression model
mod_npi <- lm(FelUn_total ~ Condition * NPI_centered, data = the.data)
summary(mod_npi)

## 
## Call:
## lm(formula = FelUn_total ~ Condition * NPI_centered, data = the.data)
## 
## Residuals:
##      Min       1Q   Median       3Q      Max 
## -2.06370 -0.40706  0.09275  0.51530  1.47225 
## 
## Coefficients:
##                        Estimate Std. Error t value Pr(>|t|)    
## (Intercept)             3.05848    0.05964  51.281  < 2e-16 ***
## Condition              -0.33769    0.08418  -4.012 8.02e-05 ***
## NPI_centered           -0.01938    0.12359  -0.157    0.876    
## Condition:NPI_centered -0.12018    0.17076  -0.704    0.482    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 0.6635 on 245 degrees of freedom
##   (29 observations deleted due to missingness)
## Multiple R-squared:  0.06811,    Adjusted R-squared:  0.0567 
## F-statistic: 5.968 on 3 and 245 DF,  p-value: 0.0006084

# Effect sizes
cat("\nPartial Eta-Squared:\n")

## 
## Partial Eta-Squared:

print(eta_squared(mod_npi, partial = TRUE))

## # Effect Size for ANOVA (Type I)
## 
## Parameter              | Eta2 (partial) |       95% CI
## ------------------------------------------------------
## Condition              |           0.06 | [0.02, 1.00]
## NPI_centered           |       3.79e-03 | [0.00, 1.00]
## Condition:NPI_centered |       2.02e-03 | [0.00, 1.00]
## 
## - One-sided CIs: upper bound fixed at [1.00].

Interaction Plot

interact_plot(mod_npi,
              pred = Condition,
              modx = NPI_centered,
              interval = TRUE,
              int.width = 0.95,
              x.label = "Condition",
              y.label = "Feelings of Uncertainty",
              legend.main = "Narcissism (Centered)") +
  theme_minimal() +
  ggtitle("Condition × Narcissism Interaction")

Johnson-Neyman Regions of Significance

johnson_neyman(mod_npi, pred = Condition, modx = NPI_centered, alpha = 0.05)

## JOHNSON-NEYMAN INTERVAL
## 
## When NPI_centered is INSIDE the interval [-0.61, 1.43], the slope of
## Condition is p < .05.
## 
## Note: The range of observed values of NPI_centered is [-1.49, 1.88]

Moderator 2: Uncertainty Aversion

Scale Reliability

uncert_items <- paste0("UncerAver", 1:15)

# Convert and calculate
the.data <- the.data %>%
  mutate(across(all_of(uncert_items), as.numeric)) %>%
  mutate(UncertAver_mean = rowMeans(select(., all_of(uncert_items)), na.rm = TRUE))

# Reliability
alpha_uncert <- psych::alpha(the.data[, uncert_items])
cat(sprintf("Uncertainty Aversion: Cronbach's α = %.3f\n", alpha_uncert$total$raw_alpha))

## Uncertainty Aversion: Cronbach's α = 0.911

Distribution

ggplot(the.data, aes(x = UncertAver_mean)) +
  geom_histogram(bins = 30, fill = "purple", color = "black", alpha = 0.7) +
  labs(title = "Distribution of Uncertainty Aversion Scores",
       x = "Uncertainty Aversion Mean Score",
       y = "Count") +
  theme_minimal()

Moderation Analysis

# Center the moderator
the.data <- the.data %>%
  mutate(UncertAver_centered = UncertAver_mean - mean(UncertAver_mean, na.rm = TRUE))

# Regression model
mod_uncert <- lm(FelUn_total ~ Condition * UncertAver_centered, data = the.data)
summary(mod_uncert)

## 
## Call:
## lm(formula = FelUn_total ~ Condition * UncertAver_centered, data = the.data)
## 
## Residuals:
##      Min       1Q   Median       3Q      Max 
## -2.02916 -0.39894  0.04897  0.45206  1.52211 
## 
## Coefficients:
##                               Estimate Std. Error t value Pr(>|t|)    
## (Intercept)                    3.05017    0.05781  52.759  < 2e-16 ***
## Condition                     -0.30626    0.08180  -3.744 0.000225 ***
## UncertAver_centered            0.11390    0.07998   1.424 0.155648    
## Condition:UncertAver_centered  0.19864    0.11394   1.743 0.082513 .  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 0.6424 on 246 degrees of freedom
##   (28 observations deleted due to missingness)
## Multiple R-squared:  0.1231, Adjusted R-squared:  0.1124 
## F-statistic: 11.52 on 3 and 246 DF,  p-value: 4.328e-07

# Effect sizes
cat("\nPartial Eta-Squared:\n")

## 
## Partial Eta-Squared:

print(eta_squared(mod_uncert, partial = TRUE))

## # Effect Size for ANOVA (Type I)
## 
## Parameter                     | Eta2 (partial) |       95% CI
## -------------------------------------------------------------
## Condition                     |           0.07 | [0.03, 1.00]
## UncertAver_centered           |           0.05 | [0.02, 1.00]
## Condition:UncertAver_centered |           0.01 | [0.00, 1.00]
## 
## - One-sided CIs: upper bound fixed at [1.00].

Interaction Plot

interact_plot(mod_uncert,
              pred = Condition,
              modx = UncertAver_centered,
              interval = TRUE,
              int.width = 0.95,
              x.label = "Condition",
              y.label = "Feelings of Uncertainty",
              legend.main = "Uncertainty Aversion",
              colors = c("darkgreen", "blue", "darkred")) +
  theme_minimal() +
  ggtitle("Condition × Uncertainty Aversion Interaction")

Johnson-Neyman Analysis

johnson_neyman(mod_uncert, pred = Condition, modx = UncertAver_centered, alpha = 0.05)

## JOHNSON-NEYMAN INTERVAL
## 
## When UncertAver_centered is INSIDE the interval [-11.72, 0.53], the slope
## of Condition is p < .05.
## 
## Note: The range of observed values of UncertAver_centered is [-2.32, 1.68]

Moderator 3: Attachment Style (ECR-R)

Scale Reliability

# Define items
ecr_anx_items <- paste0("ECR.anx_", 1:18)
ecr_avoi_items <- paste0("ECR.avoi_", 1:18)
all_ecr_items <- c(ecr_anx_items, ecr_avoi_items)

# Convert to numeric
the.data <- the.data %>%
  mutate(across(all_of(all_ecr_items), as.numeric))

# Reverse code identified items
anx_reverse_items <- c("ECR.anx_9", "ECR.anx_11")
avoi_reverse_items <- c("ECR.avoi_1", "ECR.avoi_3", "ECR.avoi_5",
                        "ECR.avoi_6", "ECR.avoi_7", "ECR.avoi_14")

the.data <- the.data %>%
  mutate(
    across(all_of(anx_reverse_items), ~ 6 - .),
    across(all_of(avoi_reverse_items), ~ 6 - .)
  )

# Calculate subscales
the.data <- the.data %>%
  mutate(
    ECR_Anxiety_mean = rowMeans(select(., all_of(ecr_anx_items)), na.rm = TRUE),
    ECR_Avoidance_mean = rowMeans(select(., all_of(ecr_avoi_items)), na.rm = TRUE)
  )

# Reliability
alpha_anx <- psych::alpha(the.data[, ecr_anx_items])
alpha_avoi <- psych::alpha(the.data[, ecr_avoi_items])

cat(sprintf("Attachment Anxiety: Cronbach's α = %.3f\n", alpha_anx$total$raw_alpha))

## Attachment Anxiety: Cronbach's α = 0.923

cat(sprintf("Attachment Avoidance: Cronbach's α = %.3f\n", alpha_avoi$total$raw_alpha))

## Attachment Avoidance: Cronbach's α = 0.934

Distribution

par(mfrow = c(1, 2))
hist(the.data$ECR_Anxiety_mean, main = "Attachment Anxiety",
     xlab = "Mean Score", col = "coral", breaks = 20)
hist(the.data$ECR_Avoidance_mean, main = "Attachment Avoidance",
     xlab = "Mean Score", col = "steelblue", breaks = 20)

par(mfrow = c(1, 1))

Attachment Anxiety Moderation

# Center
the.data <- the.data %>%
  mutate(ECR_Anxiety_centered = ECR_Anxiety_mean - mean(ECR_Anxiety_mean, na.rm = TRUE))

# Model
mod_anxiety <- lm(FelUn_total ~ Condition * ECR_Anxiety_centered, data = the.data)
summary(mod_anxiety)

## 
## Call:
## lm(formula = FelUn_total ~ Condition * ECR_Anxiety_centered, 
##     data = the.data)
## 
## Residuals:
##      Min       1Q   Median       3Q      Max 
## -2.02881 -0.43135  0.05666  0.46610  1.40393 
## 
## Coefficients:
##                                Estimate Std. Error t value Pr(>|t|)    
## (Intercept)                     3.05305    0.05978  51.072  < 2e-16 ***
## Condition                      -0.33217    0.08451  -3.931  0.00011 ***
## ECR_Anxiety_centered            0.05912    0.07668   0.771  0.44145    
## Condition:ECR_Anxiety_centered  0.02293    0.11438   0.201  0.84125    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 0.6622 on 247 degrees of freedom
##   (27 observations deleted due to missingness)
## Multiple R-squared:  0.07073,    Adjusted R-squared:  0.05944 
## F-statistic: 6.267 on 3 and 247 DF,  p-value: 0.0004085

Interaction Plot: Anxiety

interact_plot(mod_anxiety,
              pred = Condition,
              modx = ECR_Anxiety_centered,
              interval = TRUE,
              int.width = 0.95,
              x.label = "Condition",
              y.label = "Feelings of Uncertainty",
              legend.main = "Attachment Anxiety") +
  theme_minimal() +
  ggtitle("Condition × Attachment Anxiety Interaction")

Attachment Avoidance Moderation

# Center
the.data <- the.data %>%
  mutate(ECR_Avoidance_centered = ECR_Avoidance_mean - mean(ECR_Avoidance_mean, na.rm = TRUE))

# Model
mod_avoidance <- lm(FelUn_total ~ Condition * ECR_Avoidance_centered, data = the.data)
summary(mod_avoidance)

## 
## Call:
## lm(formula = FelUn_total ~ Condition * ECR_Avoidance_centered, 
##     data = the.data)
## 
## Residuals:
##      Min       1Q   Median       3Q      Max 
## -2.10838 -0.41215  0.04433  0.49307  1.41134 
## 
## Coefficients:
##                                  Estimate Std. Error t value Pr(>|t|)    
## (Intercept)                       3.06685    0.05983  51.263   <2e-16 ***
## Condition                        -0.36679    0.08429  -4.351    2e-05 ***
## ECR_Avoidance_centered           -0.09638    0.08225  -1.172    0.242    
## Condition:ECR_Avoidance_centered  0.10664    0.11894   0.897    0.371    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 0.6572 on 242 degrees of freedom
##   (32 observations deleted due to missingness)
## Multiple R-squared:  0.08063,    Adjusted R-squared:  0.06923 
## F-statistic: 7.074 on 3 and 242 DF,  p-value: 0.0001413

Interaction Plot: Avoidance

interact_plot(mod_avoidance,
              pred = Condition,
              modx = ECR_Avoidance_centered,
              interval = TRUE,
              int.width = 0.95,
              x.label = "Condition",
              y.label = "Feelings of Uncertainty",
              legend.main = "Attachment Avoidance") +
  theme_minimal() +
  ggtitle("Condition × Attachment Avoidance Interaction")

Categorical Attachment Styles

# Create attachment style categories
the.data <- the.data %>%
  mutate(
    Attachment_Style = case_when(
      ECR_Anxiety_mean < median(ECR_Anxiety_mean, na.rm = TRUE) &
        ECR_Avoidance_mean < median(ECR_Avoidance_mean, na.rm = TRUE) ~ "Secure",
      ECR_Anxiety_mean >= median(ECR_Anxiety_mean, na.rm = TRUE) &
        ECR_Avoidance_mean < median(ECR_Avoidance_mean, na.rm = TRUE) ~ "Preoccupied",
      ECR_Anxiety_mean < median(ECR_Anxiety_mean, na.rm = TRUE) &
        ECR_Avoidance_mean >= median(ECR_Avoidance_mean, na.rm = TRUE) ~ "Dismissing",
      ECR_Anxiety_mean >= median(ECR_Anxiety_mean, na.rm = TRUE) &
        ECR_Avoidance_mean >= median(ECR_Avoidance_mean, na.rm = TRUE) ~ "Fearful-Avoidant"
    )
  )

# Distribution
attach_table <- table(the.data$Attachment_Style)
kable(as.data.frame(attach_table),
      col.names = c("Attachment Style", "N"),
      caption = "Distribution of Attachment Styles") %>%
  kable_styling(bootstrap_options = c("striped", "hover"))

Distribution of Attachment Styles
Attachment Style	N
Dismissing	84
Fearful-Avoidant	44
Preoccupied	79
Secure	40

ANOVA: Attachment Style

aov_attachment <- aov(FelUn_total ~ Condition * Attachment_Style, data = the.data)
summary(aov_attachment)

##                             Df Sum Sq Mean Sq F value   Pr(>F)    
## Condition                    1   8.57   8.568  19.926 1.24e-05 ***
## Attachment_Style             3   1.16   0.387   0.900    0.442    
## Condition:Attachment_Style   3   1.64   0.546   1.269    0.286    
## Residuals                  238 102.34   0.430                     
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 32 observations deleted due to missingness

# Effect sizes
cat("\nPartial Eta-Squared:\n")

## 
## Partial Eta-Squared:

print(eta_squared(aov_attachment, partial = TRUE))

## # Effect Size for ANOVA (Type I)
## 
## Parameter                  | Eta2 (partial) |       95% CI
## ----------------------------------------------------------
## Condition                  |           0.08 | [0.03, 1.00]
## Attachment_Style           |           0.01 | [0.00, 1.00]
## Condition:Attachment_Style |           0.02 | [0.00, 1.00]
## 
## - One-sided CIs: upper bound fixed at [1.00].

Visualization

ggplot(the.data %>% filter(!is.na(Attachment_Style)),
       aes(x = factor(Condition), y = FelUn_total, fill = Attachment_Style)) +
  stat_summary(fun = mean, geom = "bar", position = position_dodge(0.9)) +
  stat_summary(fun.data = mean_se, geom = "errorbar",
               position = position_dodge(0.9), width = 0.2) +
  scale_fill_manual(values = c("Secure" = "darkgreen",
                                "Preoccupied" = "red",
                                "Dismissing" = "blue",
                                "Fearful-Avoidant" = "purple")) +
  labs(x = "Condition",
       y = "Feelings of Uncertainty",
       fill = "Attachment Style",
       title = "Condition Effect by Attachment Style") +
  theme_minimal()

Moderator 4: Self-Control

Scale Reliability

# Define items
selfcon_items <- c("SelfCon1x", "SelfCon2x", "SelfCon3x", "SelfCon4x", "SelfCon5",
                   "SelfCon6x", "SelfCon7", "SelfCon8", "SelfCon9x", "SelfCon10",
                   "SelfCon11x", "SelfCon12x", "SelfCon13x")

# Convert to numeric
the.data <- the.data %>%
  mutate(across(all_of(selfcon_items), as.numeric))

# Reverse code items with 'x' suffix
selfcon_reverse <- c("SelfCon1x", "SelfCon2x", "SelfCon3x", "SelfCon4x",
                     "SelfCon6x", "SelfCon9x", "SelfCon11x", "SelfCon12x", "SelfCon13x")

the.data <- the.data %>%
  mutate(across(all_of(selfcon_reverse), ~ 6 - .))

# Calculate scores
the.data <- the.data %>%
  mutate(SelfCon_mean = rowMeans(select(., all_of(selfcon_items)), na.rm = TRUE))

# Reliability
alpha_selfcon <- psych::alpha(the.data[, selfcon_items])
cat(sprintf("Self-Control: Cronbach's α = %.3f\n", alpha_selfcon$total$raw_alpha))

## Self-Control: Cronbach's α = 0.826

Distribution

ggplot(the.data, aes(x = SelfCon_mean)) +
  geom_histogram(bins = 30, fill = "steelblue", color = "black", alpha = 0.7) +
  labs(title = "Distribution of Self-Control Scores",
       x = "Self-Control Mean Score",
       y = "Count") +
  theme_minimal()

Moderation Analysis

# Center
the.data <- the.data %>%
  mutate(SelfCon_centered = SelfCon_mean - mean(SelfCon_mean, na.rm = TRUE))

# Model
mod_selfcon <- lm(FelUn_total ~ Condition * SelfCon_centered, data = the.data)
summary(mod_selfcon)

## 
## Call:
## lm(formula = FelUn_total ~ Condition * SelfCon_centered, data = the.data)
## 
## Residuals:
##      Min       1Q   Median       3Q      Max 
## -2.11584 -0.42853  0.07154  0.45520  1.46431 
## 
## Coefficients:
##                            Estimate Std. Error t value Pr(>|t|)    
## (Intercept)                 3.05858    0.05846  52.323  < 2e-16 ***
## Condition                  -0.34332    0.08251  -4.161 4.38e-05 ***
## SelfCon_centered           -0.11862    0.09132  -1.299    0.195    
## Condition:SelfCon_centered -0.12677    0.13268  -0.955    0.340    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 0.6535 on 247 degrees of freedom
##   (27 observations deleted due to missingness)
## Multiple R-squared:  0.09497,    Adjusted R-squared:  0.08398 
## F-statistic:  8.64 on 3 and 247 DF,  p-value: 1.786e-05

# Effect sizes
cat("\nPartial Eta-Squared:\n")

## 
## Partial Eta-Squared:

print(eta_squared(mod_selfcon, partial = TRUE))

## # Effect Size for ANOVA (Type I)
## 
## Parameter                  | Eta2 (partial) |       95% CI
## ----------------------------------------------------------
## Condition                  |           0.07 | [0.03, 1.00]
## SelfCon_centered           |           0.03 | [0.00, 1.00]
## Condition:SelfCon_centered |       3.68e-03 | [0.00, 1.00]
## 
## - One-sided CIs: upper bound fixed at [1.00].

Interaction Plot

interact_plot(mod_selfcon,
              pred = Condition,
              modx = SelfCon_centered,
              interval = TRUE,
              int.width = 0.95,
              x.label = "Condition",
              y.label = "Feelings of Uncertainty",
              legend.main = "Self-Control") +
  theme_minimal() +
  ggtitle("Condition × Self-Control Interaction")

Johnson-Neyman Analysis

johnson_neyman(mod_selfcon, pred = Condition, modx = SelfCon_centered, alpha = 0.05)

## JOHNSON-NEYMAN INTERVAL
## 
## When SelfCon_centered is INSIDE the interval [-0.73, 2.41], the slope of
## Condition is p < .05.
## 
## Note: The range of observed values of SelfCon_centered is [-1.56, 1.82]

Moderator 5: Self-Esteem (Rosenberg)

Scale Reliability

# Define items
selfest_items <- c("SelfEst1", "SelfEst2", "SelfEst3x", "SelfEst4", "SelfEst5x",
                   "SelfEst6", "SelfEst7", "SelfEst8x", "SelfEst9x", "SelfEst10x")

# Convert to numeric
the.data <- the.data %>%
  mutate(across(all_of(selfest_items), as.numeric))

# Reverse code items with 'x' suffix
selfest_reverse <- c("SelfEst3x", "SelfEst5x", "SelfEst8x", "SelfEst9x", "SelfEst10x")

the.data <- the.data %>%
  mutate(across(all_of(selfest_reverse), ~ 6 - .))

# Calculate scores
the.data <- the.data %>%
  mutate(SelfEst_mean = rowMeans(select(., all_of(selfest_items)), na.rm = TRUE))

# Reliability
alpha_selfest <- psych::alpha(the.data[, selfest_items])
cat(sprintf("Self-Esteem (Rosenberg): Cronbach's α = %.3f\n", alpha_selfest$total$raw_alpha))

## Self-Esteem (Rosenberg): Cronbach's α = 0.908

Distribution

ggplot(the.data, aes(x = SelfEst_mean)) +
  geom_histogram(bins = 30, fill = "darkseagreen", color = "black", alpha = 0.7) +
  labs(title = "Distribution of Self-Esteem Scores",
       x = "Self-Esteem Mean Score",
       y = "Count") +
  theme_minimal()

Moderation Analysis

# Center
the.data <- the.data %>%
  mutate(SelfEst_centered = SelfEst_mean - mean(SelfEst_mean, na.rm = TRUE))

# Model
mod_selfest <- lm(FelUn_total ~ Condition * SelfEst_centered, data = the.data)
summary(mod_selfest)

## 
## Call:
## lm(formula = FelUn_total ~ Condition * SelfEst_centered, data = the.data)
## 
## Residuals:
##      Min       1Q   Median       3Q      Max 
## -2.08021 -0.39647  0.06019  0.45562  1.48461 
## 
## Coefficients:
##                            Estimate Std. Error t value Pr(>|t|)    
## (Intercept)                 3.04720    0.05825  52.315  < 2e-16 ***
## Condition                  -0.30416    0.08246  -3.689 0.000278 ***
## SelfEst_centered           -0.12637    0.07240  -1.745 0.082160 .  
## Condition:SelfEst_centered -0.11504    0.10412  -1.105 0.270308    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 0.6466 on 246 degrees of freedom
##   (28 observations deleted due to missingness)
## Multiple R-squared:  0.1116, Adjusted R-squared:  0.1008 
## F-statistic:  10.3 on 3 and 246 DF,  p-value: 2.057e-06

# Effect sizes
cat("\nPartial Eta-Squared:\n")

## 
## Partial Eta-Squared:

print(eta_squared(mod_selfest, partial = TRUE))

## # Effect Size for ANOVA (Type I)
## 
## Parameter                  | Eta2 (partial) |       95% CI
## ----------------------------------------------------------
## Condition                  |           0.07 | [0.02, 1.00]
## SelfEst_centered           |           0.05 | [0.01, 1.00]
## Condition:SelfEst_centered |       4.94e-03 | [0.00, 1.00]
## 
## - One-sided CIs: upper bound fixed at [1.00].

Interaction Plot

interact_plot(mod_selfest,
              pred = Condition,
              modx = SelfEst_centered,
              interval = TRUE,
              int.width = 0.95,
              x.label = "Condition",
              y.label = "Feelings of Uncertainty",
              legend.main = "Self-Esteem") +
  theme_minimal() +
  ggtitle("Condition × Self-Esteem Interaction")

Johnson-Neyman Analysis

johnson_neyman(mod_selfest, pred = Condition, modx = SelfEst_centered, alpha = 0.05)

## JOHNSON-NEYMAN INTERVAL
## 
## When SelfEst_centered is INSIDE the interval [-0.72, 3.17], the slope of
## Condition is p < .05.
## 
## Note: The range of observed values of SelfEst_centered is [-2.16, 1.74]

Summary of All Moderations

# Create summary table
moderation_summary <- data.frame(
  Moderator = c("Narcissism (NPI)", "Uncertainty Aversion", "Attachment Anxiety",
                "Attachment Avoidance", "Self-Control", "Self-Esteem"),
  Interaction_Beta = c(
    coef(mod_npi)["Condition:NPI_centered"],
    coef(mod_uncert)["Condition:UncertAver_centered"],
    coef(mod_anxiety)["Condition:ECR_Anxiety_centered"],
    coef(mod_avoidance)["Condition:ECR_Avoidance_centered"],
    coef(mod_selfcon)["Condition:SelfCon_centered"],
    coef(mod_selfest)["Condition:SelfEst_centered"]
  ),
  P_value = c(
    summary(mod_npi)$coefficients["Condition:NPI_centered", "Pr(>|t|)"],
    summary(mod_uncert)$coefficients["Condition:UncertAver_centered", "Pr(>|t|)"],
    summary(mod_anxiety)$coefficients["Condition:ECR_Anxiety_centered", "Pr(>|t|)"],
    summary(mod_avoidance)$coefficients["Condition:ECR_Avoidance_centered", "Pr(>|t|)"],
    summary(mod_selfcon)$coefficients["Condition:SelfCon_centered", "Pr(>|t|)"],
    summary(mod_selfest)$coefficients["Condition:SelfEst_centered", "Pr(>|t|)"]
  ),
  Significant = c(
    summary(mod_npi)$coefficients["Condition:NPI_centered", "Pr(>|t|)"] < 0.05,
    summary(mod_uncert)$coefficients["Condition:UncertAver_centered", "Pr(>|t|)"] < 0.05,
    summary(mod_anxiety)$coefficients["Condition:ECR_Anxiety_centered", "Pr(>|t|)"] < 0.05,
    summary(mod_avoidance)$coefficients["Condition:ECR_Avoidance_centered", "Pr(>|t|)"] < 0.05,
    summary(mod_selfcon)$coefficients["Condition:SelfCon_centered", "Pr(>|t|)"] < 0.05,
    summary(mod_selfest)$coefficients["Condition:SelfEst_centered", "Pr(>|t|)"] < 0.05
  )
)

kable(moderation_summary,
      digits = 4,
      col.names = c("Moderator", "Interaction β", "p-value", "Significant (p < .05)"),
      caption = "Summary of Moderation Analyses") %>%
  kable_styling(bootstrap_options = c("striped", "hover")) %>%
  row_spec(which(moderation_summary$Significant), bold = TRUE, color = "white", background = "#3498db")

Summary of Moderation Analyses
	Moderator	Interaction β	p-value	Significant (p < .05)
Condition:NPI_centered	Narcissism (NPI)	-0.1202	0.4822	FALSE
Condition:UncertAver_centered	Uncertainty Aversion	0.1986	0.0825	FALSE
Condition:ECR_Anxiety_centered	Attachment Anxiety	0.0229	0.8412	FALSE
Condition:ECR_Avoidance_centered	Attachment Avoidance	0.1066	0.3708	FALSE
Condition:SelfCon_centered	Self-Control	-0.1268	0.3403	FALSE
Condition:SelfEst_centered	Self-Esteem	-0.1150	0.2703	FALSE

Correlation Matrix

# Create correlation matrix of all key variables
cor_vars <- the.data %>%
  select(FelUn_total, NPI_Total_mean, UncertAver_mean,
         ECR_Anxiety_mean, ECR_Avoidance_mean, SelfCon_mean, SelfEst_mean) %>%
  rename(
    `Feelings of Uncertainty` = FelUn_total,
    `Narcissism` = NPI_Total_mean,
    `Uncertainty Aversion` = UncertAver_mean,
    `Attachment Anxiety` = ECR_Anxiety_mean,
    `Attachment Avoidance` = ECR_Avoidance_mean,
    `Self-Control` = SelfCon_mean,
    `Self-Esteem` = SelfEst_mean
  )

cor_matrix <- cor(cor_vars, use = "pairwise.complete.obs")

kable(cor_matrix, digits = 3, caption = "Correlation Matrix of All Variables") %>%
  kable_styling(bootstrap_options = c("striped", "hover"))

Correlation Matrix of All Variables
	Feelings of Uncertainty	Narcissism	Uncertainty Aversion	Attachment Anxiety	Attachment Avoidance	Self-Control	Self-Esteem
Feelings of Uncertainty	1.000	-0.071	0.249	0.111	-0.076	-0.167	-0.240
Narcissism	-0.071	1.000	-0.259	-0.221	0.214	0.124	0.405
Uncertainty Aversion	0.249	-0.259	1.000	0.486	-0.168	-0.387	-0.504
Attachment Anxiety	0.111	-0.221	0.486	1.000	-0.440	-0.356	-0.561
Attachment Avoidance	-0.076	0.214	-0.168	-0.440	1.000	0.225	0.388
Self-Control	-0.167	0.124	-0.387	-0.356	0.225	1.000	0.464
Self-Esteem	-0.240	0.405	-0.504	-0.561	0.388	0.464	1.000

Conclusions

This report presents a comprehensive analysis of the Lone Wolf study data. The main effect of condition on feelings of uncertainty was tested, along with five potential moderators:

Narcissistic Personality (NPI)
Uncertainty Aversion
Attachment Style (Anxiety and Avoidance)
Self-Control
Self-Esteem (Rosenberg)

Key findings are summarized in the moderation summary table above, showing which moderators significantly influenced the relationship between experimental condition and feelings of uncertainty.

Report generated on 2025-11-27

Lone Wolf Study: Analysis Report

Psychology 390

2025-11-27

Overview

Data Preparation

Main Effect: Condition on Feelings of Uncertainty

Descriptive Statistics

Distribution

T-Test Results

Assumptions Checks

Moderator 1: Narcissistic Personality Inventory (NPI)

Scale Reliability

Distribution

Moderation Analysis

Interaction Plot

Johnson-Neyman Regions of Significance

Moderator 2: Uncertainty Aversion

Scale Reliability

Distribution

Moderation Analysis

Interaction Plot

Johnson-Neyman Analysis

Moderator 3: Attachment Style (ECR-R)

Scale Reliability

Distribution

Attachment Anxiety Moderation

Interaction Plot: Anxiety

Attachment Avoidance Moderation

Interaction Plot: Avoidance

Categorical Attachment Styles

ANOVA: Attachment Style

Visualization

Moderator 4: Self-Control

Scale Reliability

Distribution

Moderation Analysis

Interaction Plot

Johnson-Neyman Analysis

Moderator 5: Self-Esteem (Rosenberg)

Scale Reliability

Distribution

Moderation Analysis

Interaction Plot

Johnson-Neyman Analysis

Summary of All Moderations

Correlation Matrix

Conclusions