.Rhistory

# Engagement profiles — change scores by activity and dimension
# Computed directly from matched_survey
# ------------------------------------------------------------
# Cohen's d helper (already defined in analysis script)
# d = mean(change) / sd(change)
compute_d <- function(data, chg_var) {
x <- data[[chg_var]]
x <- x[!is.na(x)]
round(mean(x) / sd(x), 3)
}
# Wilcoxon p-value helper for significance flagging
compute_p <- function(data, pre_var, post_var) {
x <- data[[pre_var]]
y <- data[[post_var]]
valid <- complete.cases(x, y)
d <- y[valid] - x[valid]
wilcox.test(d, mu = 0, exact = FALSE)$p.value
}
profile_data <- bind_rows(
# Podcasting
tibble(activity = "Podcasting",    dimension = "Behavioral",
d = compute_d(matched_survey, "chg_beh_podcast"),
p = compute_p(matched_survey, "beh_podcast_pre",   "beh_podcast_post")),
tibble(activity = "Podcasting",    dimension = "Emotional",
d = compute_d(matched_survey, "chg_emo_podcast"),
p = compute_p(matched_survey, "emo_podcast_pre",   "emo_podcast_post")),
tibble(activity = "Podcasting",    dimension = "Cognitive",
d = compute_d(matched_survey, "chg_cog_podcast"),
p = compute_p(matched_survey, "cog_podcast_pre",   "cog_podcast_post")),
# Case study
tibble(activity = "Case study",    dimension = "Behavioral",
d = compute_d(matched_survey, "chg_beh_casestudy"),
p = compute_p(matched_survey, "beh_casestudy_pre", "beh_casestudy_post")),
tibble(activity = "Case study",    dimension = "Emotional",
d = compute_d(matched_survey, "chg_emo_casestudy"),
p = compute_p(matched_survey, "emo_casestudy_pre", "emo_casestudy_post")),
tibble(activity = "Case study",    dimension = "Cognitive",
d = compute_d(matched_survey, "chg_cog_casestudy"),
p = compute_p(matched_survey, "cog_casestudy_pre", "cog_casestudy_post")),
# Business case — emotional pre unavailable
tibble(activity = "Business case", dimension = "Behavioral",
d = compute_d(matched_survey, "chg_beh_bizcase"),
p = compute_p(matched_survey, "beh_bizcase_pre",   "beh_bizcase_post")),
tibble(activity = "Business case", dimension = "Cognitive",
d = compute_d(matched_survey, "chg_cog_bizcase"),
p = compute_p(matched_survey, "cog_bizcase_pre",   "cog_bizcase_post")),
# Mentorship — emotional pre unavailable
tibble(activity = "Mentorship",    dimension = "Behavioral",
d = compute_d(matched_survey, "chg_beh_mentor"),
p = compute_p(matched_survey, "beh_mentor_pre",    "beh_mentor_post")),
tibble(activity = "Mentorship",    dimension = "Cognitive",
d = compute_d(matched_survey, "chg_cog_mentor"),
p = compute_p(matched_survey, "cog_mentor_pre",    "cog_mentor_post"))
) %>%
mutate(
sig = ifelse(p < .05, "*", ""),
activity  = factor(activity,
levels = c("Podcasting", "Case study",
"Business case", "Mentorship")),
dimension = factor(dimension,
levels = c("Behavioral", "Emotional", "Cognitive")),
color = case_when(
d >  0.1 ~ "Positive",
d < -0.1 ~ "Negative",
TRUE     ~ "Negligible"
)
)
# Verify values
print(profile_data %>% select(activity, dimension, d, p, sig))
profile_caption <- str_wrap(
"Figure 3. Cohen's d effect sizes for pre-to-post engagement change by activity and dimension. Positive values indicate improvement; negative values indicate decline relative to pre-semester expectations. Emotional engagement not available for business case or mentorship at pre-semester due to reliability constraints; omitted from those panels. * p < .05, Wilcoxon signed-rank test, two-sided.",
width = 120
)
ggplot(profile_data, aes(x = d, y = dimension, fill = color)) +
geom_col(width = 0.55) +
geom_vline(xintercept = 0, color = "gray30", linewidth = 0.8) +
geom_text(aes(label = paste0(ifelse(d > 0, "+", ""), round(d, 2), sig),
hjust = ifelse(d >= 0, -0.15, 1.15)),
fontface = "bold", size = 4) +
scale_fill_manual(values = c(
"Positive"   = "#1F3864",
"Negative"   = "#C8960C",
"Negligible" = "#BBBBBB"
)) +
scale_x_continuous(limits = c(-0.65, 0.55),
breaks = c(-0.4, -0.2, 0, 0.2, 0.4),
labels = c("-0.4", "-0.2", "0", "+0.2", "+0.4")) +
facet_wrap(~ activity, ncol = 4) +
labs(title = "Engagement Change by Activity and Dimension",
subtitle = "Cohen's d (post minus pre)  ·  * p < .05",
caption = profile_caption,
x = "Effect size (Cohen's d)", y = NULL) +
theme_minimal(base_size = 13) +
theme(legend.position = "none",
plot.title = element_text(face = "bold", hjust = 0.5, size = 14),
plot.subtitle = element_text(hjust = 0.5, color = "gray40", size = 11),
plot.caption = element_text(hjust = 0, size = 9,
color = "gray40", face = "italic"),
plot.caption.position = "plot",
strip.text = element_text(face = "bold", size = 12),
axis.text = element_text(face = "bold"),
panel.grid.major.y = element_blank())
ggsave("tilt/finding2_engagement_profiles.png",
width = 10, height = 4.5, dpi = 300)
# ------------------------------------------------------------
# RQX. Finding 3
# Knowledge Assessment — Pre/Post Score Distribution
# Computed directly from matched_survey
# ------------------------------------------------------------
# Compute score distributions from matched_survey
ka_data <- bind_rows(
matched_survey %>%
count(score = ka_score_pre) %>%
mutate(time = "Pre"),
matched_survey %>%
count(score = ka_score_post) %>%
mutate(time = "Post")
) %>%
# Ensure all score levels appear for both time points
# (e.g. if no one scored 0 or 1 at post, still show as 0)
complete(score, time, fill = list(n = 0)) %>%
filter(score >= 2) %>%  # drop scores below 2 if none present
mutate(
time  = factor(time, levels = c("Pre", "Post")),
score = factor(score,
levels = sort(unique(score)),
labels = paste0(sort(unique(score)), " of 5"))
)
# Verify distributions match Step 13 output
cat("Pre distribution:\n")
print(matched_survey %>% count(ka_score_pre))
cat("\nPost distribution:\n")
print(matched_survey %>% count(ka_score_post))
# Compute mean scores for annotation
mean_pre  <- round(mean(matched_survey$ka_score_pre,  na.rm = TRUE), 2)
mean_post <- round(mean(matched_survey$ka_score_post, na.rm = TRUE), 2)
ka_caption <- str_wrap(
"Figure 4. Distribution of knowledge assessment scores at pre- and post-semester (n = 38). Score = number correct out of 5 items retained for adequate response rates. Mean improved from 3.82 to 4.34 (W = 234, p = .003, r = 0.37, medium effect). The combination of active learning strategies precludes attribution to any single activity; the result is consistent with a course environment that supported content comprehension alongside engagement.",
width = 120
)
ggplot(ka_data, aes(x = score, y = n, fill = time)) +
geom_col(position = position_dodge(width = 0.7), width = 0.65) +
geom_text(aes(label = ifelse(n > 0, n, "")),
position = position_dodge(width = 0.7),
vjust = -0.4, fontface = "bold", size = 4.5) +
scale_fill_manual(values = c("Pre" = "#D5E3F0", "Post" = "#1F3864"),
name = NULL) +
scale_y_continuous(limits = c(0, 22), breaks = seq(0, 20, 5)) +
annotate("text", x = 3.5, y = 21,
label = paste0("Mean: ", mean_pre, " \u2192 ", mean_post,
"  (p = .003, r = 0.37)"),
fontface = "bold.italic", size = 4.5, color = "#1F3864") +
labs(title = "Knowledge Assessment: Pre- and Post-Semester Scores",
subtitle = "Students demonstrated significant gains in content knowledge over the semester",
caption = ka_caption,
x = "Score (items correct out of 5)",
y = "Number of students") +
theme_minimal(base_size = 13) +
theme(plot.title = element_text(face = "bold", hjust = 0.5, size = 14),
plot.subtitle = element_text(hjust = 0.5, color = "gray40", size = 11),
plot.caption = element_text(hjust = 0, size = 9,
color = "gray40", face = "italic"),
plot.caption.position = "plot",
legend.position = "top",
axis.text = element_text(face = "bold"),
panel.grid.major.x = element_blank())
ggsave("tilt/finding3_knowledge.png", width = 7, height = 5, dpi = 300)
# ------------------------------------------------------------
# RQ3. Finding 4 — Correlation heatmap
# Beliefs predict engagement intentions r = 0.50 to 0.65
# across all nine correlations.
# ------------------------------------------------------------
# Define variable groups with updated labels
bv_vars <- c(
"Growth mindset"              = "bv_growth_score_pre",
"Grit"                        = "bv_persist_score_pre",
"Entrepreneurial identity"    = "bv_relevance_score_pre"
)
beh_vars <- c(
"Podcasting"    = "beh_podcast_pre",
"Case study"    = "beh_casestudy_pre",
"Business case" = "beh_bizcase_pre",
"Mentorship"    = "beh_mentor_pre"
)
# Compute Pearson correlations between each belief/values
# composite and each anticipated behavioral engagement item
cor_long <- expand.grid(
belief   = names(bv_vars),
activity = names(beh_vars),
stringsAsFactors = FALSE
) %>%
rowwise() %>%
mutate(
r = cor(
matched_survey[[bv_vars[belief]]],
matched_survey[[beh_vars[activity]]],
use = "complete.obs",
method = "pearson"
)
) %>%
ungroup() %>%
mutate(
belief   = factor(belief,
levels = c("Growth mindset", "Grit",
"Entrepreneurial identity")),
activity = factor(activity,
levels = c("Podcasting", "Case study",
"Business case", "Mentorship"))
)
# Verify correlations match previously reported values
print(cor_long %>% arrange(belief, activity) %>%
mutate(r = round(r, 3)))
fig2_caption <- str_wrap(
"Figure 5. Pearson correlations between pre-semester beliefs and values composites and anticipated behavioral engagement across the four course activities (n = 38). Growth mindset = belief that intelligence and ability can be developed through effort (Dweck et al., 2014). Grit = tendency to follow through on goals and overcome setbacks. Entrepreneurial identity = perceived applicability of entrepreneurship to students' own lives, interests, and futures. All 12 correlations are positive and range from r = 0.50 to r = 0.65.",
width = 120
)
ggplot(cor_long, aes(x = activity, y = belief, fill = r)) +
geom_tile(color = "white", linewidth = 1) +
geom_text(aes(label = round(r, 2)), size = 5, fontface = "bold") +
scale_fill_gradient(low = "#D5E3F0", high = "#1F3864",
limits = c(0.45, 0.70),
name = "Pearson r") +
scale_y_discrete(labels = scales::label_wrap(13)) +
labs(title = "Beliefs and Values vs. Anticipated Behavioral Engagement",
caption = fig2_caption,
x = NULL, y = NULL) +
theme_minimal(base_size = 14) +
theme(axis.text = element_text(face = "bold"),
axis.text.y = element_text(face = "bold", lineheight = 0.9),
plot.title = element_text(face = "bold", hjust = 0.5),
plot.caption = element_text(hjust = 0, size = 9,
color = "gray40", face = "italic"),
plot.caption.position = "plot",
legend.position = "right")
ggsave("tilt/finding4_correlation_heatmap.png",
width = 7, height = 4, dpi = 300)
# ------------------------------------------------------------
# Beliefs vs. post-semester realized behavioral engagement
# Side-by-side with pre-semester anticipated engagement
# for direct comparison
# ------------------------------------------------------------
# Post behavioral engagement variables
beh_vars_post <- c(
"Podcasting"    = "beh_podcast_post",
"Case study"    = "beh_casestudy_post",
"Business case" = "beh_bizcase_post",
"Mentorship"    = "beh_mentor_post"
)
# Compute correlations: beliefs (pre) vs. realized engagement (post)
cor_long_post <- expand.grid(
belief   = names(bv_vars),
activity = names(beh_vars_post),
stringsAsFactors = FALSE
) %>%
rowwise() %>%
mutate(
r = cor(
matched_survey[[bv_vars[belief]]],
matched_survey[[beh_vars_post[activity]]],
use = "complete.obs",
method = "pearson"
)
) %>%
ungroup() %>%
mutate(
belief   = factor(belief,
levels = c("Growth mindset", "Grit",
"Entrepreneurial identity")),
activity = factor(activity,
levels = c("Podcasting", "Case study",
"Business case", "Mentorship")),
timing = "Realized (Post-Semester)"
)
# Add timing label to existing pre correlations
cor_long_pre <- cor_long %>%
mutate(timing = "Anticipated (Pre-Semester)")
# Combine
cor_both <- bind_rows(cor_long_pre, cor_long_post) %>%
mutate(timing = factor(timing,
levels = c("Anticipated (Pre-Semester)",
"Realized (Post-Semester)")))
# Verify post correlations
cat("Post-semester correlations:\n")
print(cor_long_post %>% arrange(belief, activity) %>%
mutate(r = round(r, 3)))
fig_cor_both_caption <- str_wrap(
"Figure X. Pearson correlations between pre-semester beliefs and values composites and behavioral engagement — anticipated at pre-semester (left) and realized at post-semester (right) — across the four course activities (n = 38). Growth mindset = belief that intelligence and ability can be developed through effort (Dweck et al., 2014). Grit = tendency to follow through on goals and overcome setbacks. Entrepreneurial identity = perceived applicability of entrepreneurship to students' own lives, interests, and futures.",
width = 120
)
# Shared color scale across both panels
r_limits <- range(cor_both$r, na.rm = TRUE)
r_limits <- c(floor(min(r_limits) * 20) / 20,
ceiling(max(r_limits) * 20) / 20)
ggplot(cor_both, aes(x = activity, y = belief, fill = r)) +
geom_tile(color = "white", linewidth = 1) +
geom_text(aes(label = round(r, 2)), size = 4.5, fontface = "bold") +
scale_fill_gradient(low = "#D5E3F0", high = "#1F3864",
limits = r_limits,
name = "Pearson r") +
scale_y_discrete(labels = scales::label_wrap(13)) +
facet_wrap(~ timing) +
labs(title = "Beliefs and Values vs. Behavioral Engagement: Anticipated and Realized",
caption = fig_cor_both_caption,
x = NULL, y = NULL) +
theme_minimal(base_size = 13) +
theme(axis.text = element_text(face = "bold"),
axis.text.y = element_text(face = "bold", lineheight = 0.9),
axis.text.x = element_text(angle = 15, hjust = 1),
plot.title = element_text(face = "bold", hjust = 0.5, size = 14),
plot.caption = element_text(hjust = 0, size = 9,
color = "gray40", face = "italic"),
plot.caption.position = "plot",
strip.text = element_text(face = "bold", size = 12),
legend.position = "right")
ggsave("tilt/finding_beliefs_pre_post_engagement.png",
width = 11, height = 4.5, dpi = 300)
fig_cor_both_caption <- str_wrap(
"Figure X. Pearson correlations between pre-semester beliefs and values composites and behavioral engagement — anticipated at pre-semester (left) and realized at post-semester (right) — across the four course activities (n = 38). Growth mindset = belief that intelligence and ability can be developed through effort (Dweck et al., 2014). Grit = tendency to follow through on goals and overcome setbacks. Entrepreneurial identity = perceived applicability of entrepreneurship to students' own lives, interests, and futures.",
width = 140
)
# Shared color scale across both panels
r_limits <- range(cor_both$r, na.rm = TRUE)
r_limits <- c(floor(min(r_limits) * 20) / 20,
ceiling(max(r_limits) * 20) / 20)
ggplot(cor_both, aes(x = activity, y = belief, fill = r)) +
geom_tile(color = "white", linewidth = 1) +
geom_text(aes(label = round(r, 2)), size = 4.5, fontface = "bold") +
scale_fill_gradient(low = "#D5E3F0", high = "#1F3864",
limits = r_limits,
name = "Pearson r") +
scale_y_discrete(labels = scales::label_wrap(13)) +
facet_wrap(~ timing) +
labs(title = "Beliefs and Values vs. Behavioral Engagement: Anticipated and Realized",
caption = fig_cor_both_caption,
x = NULL, y = NULL) +
theme_minimal(base_size = 13) +
theme(axis.text = element_text(face = "bold"),
axis.text.y = element_text(face = "bold", lineheight = 0.9),
axis.text.x = element_text(angle = 15, hjust = 1),
plot.title = element_text(face = "bold", hjust = 0.5, size = 14),
plot.caption = element_text(hjust = 0, size = 9,
color = "gray40", face = "italic"),
plot.caption.position = "plot",
strip.text = element_text(face = "bold", size = 12),
legend.position = "right")
ggsave("tilt/finding_beliefs_pre_post_engagement.png",
width = 11, height = 4.5, dpi = 300)
fig_cor_both_caption <- str_wrap(
"Figure 5. Pearson correlations between pre-semester beliefs and values composites and behavioral engagement — anticipated at pre-semester (left) and realized at post-semester (right) — across the four course activities (n = 38). Growth mindset = belief that intelligence and ability can be developed through effort (Dweck et al., 2014). Grit = tendency to follow through on goals and overcome setbacks. Entrepreneurial identity = perceived applicability of entrepreneurship to students' own lives, interests, and futures.",
width = 140
)
# Shared color scale across both panels
r_limits <- range(cor_both$r, na.rm = TRUE)
r_limits <- c(floor(min(r_limits) * 20) / 20,
ceiling(max(r_limits) * 20) / 20)
ggplot(cor_both, aes(x = activity, y = belief, fill = r)) +
geom_tile(color = "white", linewidth = 1) +
geom_text(aes(label = round(r, 2)), size = 4.5, fontface = "bold") +
scale_fill_gradient(low = "#D5E3F0", high = "#1F3864",
limits = r_limits,
name = "Pearson r") +
scale_y_discrete(labels = scales::label_wrap(13)) +
facet_wrap(~ timing) +
labs(title = "Beliefs and Values vs. Behavioral Engagement: Anticipated and Realized",
caption = fig_cor_both_caption,
x = NULL, y = NULL) +
theme_minimal(base_size = 13) +
theme(axis.text = element_text(face = "bold"),
axis.text.y = element_text(face = "bold", lineheight = 0.9),
axis.text.x = element_text(angle = 15, hjust = 1),
plot.title = element_text(face = "bold", hjust = 0.5, size = 14),
plot.caption = element_text(hjust = 0, size = 9,
color = "gray40", face = "italic"),
plot.caption.position = "plot",
strip.text = element_text(face = "bold", size = 12),
legend.position = "right")
ggsave("tilt/finding_beliefs_pre_post_engagement.png",
width = 11, height = 4.5, dpi = 300)
# Test significance of post-semester correlations
cor_long_post %>%
rowwise() %>%
mutate(
p_value = cor.test(
matched_survey[[bv_vars[belief]]],
matched_survey[[beh_vars_post[activity]]],
use = "complete.obs",
method = "pearson"
)$p.value,
p_value = round(p_value, 3),
sig = ifelse(p_value < .05, "*", "")
) %>%
select(belief, activity, r, p_value, sig) %>%
print(n = 12)
fig_cor_both_caption <- str_wrap(
"Figure 5. Pearson correlations between pre-semester beliefs and values composites and behavioral engagement — anticipated at pre-semester (left) and realized at post-semester (right) — across the four course activities (n = 38). Pre-semester correlations range from r = 0.50 to r = 0.65 (all p < .05). Post-semester correlations range from r = 0.06 to r = 0.32 (none significant; closest: growth mindset vs. case study, p = .050). Growth mindset = belief that intelligence and ability can be developed through effort (Dweck et al., 2014). Grit = tendency to follow through on goals and overcome setbacks. Entrepreneurial identity = perceived applicability of entrepreneurship to students' own lives, interests, and futures.",
width = 140
)
# Shared color scale across both panels
r_limits <- range(cor_both$r, na.rm = TRUE)
r_limits <- c(floor(min(r_limits) * 20) / 20,
ceiling(max(r_limits) * 20) / 20)
ggplot(cor_both, aes(x = activity, y = belief, fill = r)) +
geom_tile(color = "white", linewidth = 1) +
geom_text(aes(label = round(r, 2)), size = 4.5, fontface = "bold") +
scale_fill_gradient(low = "#D5E3F0", high = "#1F3864",
limits = r_limits,
name = "Pearson r") +
scale_y_discrete(labels = scales::label_wrap(13)) +
facet_wrap(~ timing) +
labs(title = "Beliefs and Values vs. Behavioral Engagement: Anticipated and Realized",
caption = fig_cor_both_caption,
x = NULL, y = NULL) +
theme_minimal(base_size = 13) +
theme(axis.text = element_text(face = "bold"),
axis.text.y = element_text(face = "bold", lineheight = 0.9),
axis.text.x = element_text(angle = 15, hjust = 1),
plot.title = element_text(face = "bold", hjust = 0.5, size = 14),
plot.caption = element_text(hjust = 0, size = 9,
color = "gray40", face = "italic"),
plot.caption.position = "plot",
strip.text = element_text(face = "bold", size = 12),
legend.position = "right")
fig_cor_both_caption <- str_wrap(
"Figure 5. Pearson correlations between pre-semester beliefs and values composites and behavioral engagement — anticipated at pre-semester (left) and realized at post-semester (right) — across the four course activities (n = 38). Pre-semester correlations range from r = 0.50 to r = 0.65 (all p < .05). Post-semester correlations range from r = 0.06 to r = 0.32 (none significant; closest: growth mindset vs. case study, p = .050). Growth mindset = belief that intelligence and ability can be developed through effort (Dweck et al., 2014). Grit = tendency to follow through on goals and overcome setbacks. Entrepreneurial identity = perceived applicability of entrepreneurship to students' own lives, interests, and futures.",
width = 150
)
# Shared color scale across both panels
r_limits <- range(cor_both$r, na.rm = TRUE)
r_limits <- c(floor(min(r_limits) * 20) / 20,
ceiling(max(r_limits) * 20) / 20)
ggplot(cor_both, aes(x = activity, y = belief, fill = r)) +
geom_tile(color = "white", linewidth = 1) +
geom_text(aes(label = round(r, 2)), size = 4.5, fontface = "bold") +
scale_fill_gradient(low = "#D5E3F0", high = "#1F3864",
limits = r_limits,
name = "Pearson r") +
scale_y_discrete(labels = scales::label_wrap(13)) +
facet_wrap(~ timing) +
labs(title = "Beliefs and Values vs. Behavioral Engagement: Anticipated and Realized",
caption = fig_cor_both_caption,
x = NULL, y = NULL) +
theme_minimal(base_size = 13) +
theme(axis.text = element_text(face = "bold"),
axis.text.y = element_text(face = "bold", lineheight = 0.9),
axis.text.x = element_text(angle = 15, hjust = 1),
plot.title = element_text(face = "bold", hjust = 0.5, size = 14),
plot.caption = element_text(hjust = 0, size = 9,
color = "gray40", face = "italic"),
plot.caption.position = "plot",
strip.text = element_text(face = "bold", size = 12),
legend.position = "right")
ggsave("tilt/finding_beliefs_pre_post_engagement.png",
width = 11, height = 4.5, dpi = 300)
# Test significance of pre-semester correlations
expand.grid(
belief   = names(bv_vars),
activity = names(beh_vars),
stringsAsFactors = FALSE
) %>%
rowwise() %>%
mutate(
r = cor(
matched_survey[[bv_vars[belief]]],
matched_survey[[beh_vars[activity]]],
use = "complete.obs",
method = "pearson"
),
p_value = cor.test(
matched_survey[[bv_vars[belief]]],
matched_survey[[beh_vars[activity]]],
use = "complete.obs",
method = "pearson"
)$p.value,
r       = round(r, 3),
p_value = round(p_value, 3),
sig     = ifelse(p_value < .05, "*", "")
) %>%
ungroup() %>%
select(belief, activity, r, p_value, sig) %>%
arrange(belief, activity) %>%
print(n = 12)
fig_cor_both_caption <- str_wrap(
"Figure 5. Pearson correlations between pre-semester beliefs and values composites and behavioral engagement — anticipated at pre-semester (left) and realized at post-semester (right) — across the four course activities (n = 38). Pre-semester correlations range from r = 0.50 to r = 0.65, all p ≤ .001. Post-semester correlations range from r = 0.06 to r = 0.32, none significant (all p > .05). Growth mindset = belief that intelligence and ability can be developed through effort (Dweck et al., 2014). Grit = tendency to follow through on goals and overcome setbacks. Entrepreneurial identity = perceived applicability of entrepreneurship to students' own lives, interests, and futures.",
width = 150
)
# Shared color scale across both panels
r_limits <- range(cor_both$r, na.rm = TRUE)
r_limits <- c(floor(min(r_limits) * 20) / 20,
ceiling(max(r_limits) * 20) / 20)
ggplot(cor_both, aes(x = activity, y = belief, fill = r)) +
geom_tile(color = "white", linewidth = 1) +
geom_text(aes(label = round(r, 2)), size = 4.5, fontface = "bold") +
scale_fill_gradient(low = "#D5E3F0", high = "#1F3864",
limits = r_limits,
name = "Pearson r") +
scale_y_discrete(labels = scales::label_wrap(13)) +
facet_wrap(~ timing) +
labs(title = "Beliefs and Values vs. Behavioral Engagement: Anticipated and Realized",
caption = fig_cor_both_caption,
x = NULL, y = NULL) +
theme_minimal(base_size = 13) +
theme(axis.text = element_text(face = "bold"),
axis.text.y = element_text(face = "bold", lineheight = 0.9),
axis.text.x = element_text(angle = 15, hjust = 1),
plot.title = element_text(face = "bold", hjust = 0.5, size = 14),
plot.caption = element_text(hjust = 0, size = 9,
color = "gray40", face = "italic"),
plot.caption.position = "plot",
strip.text = element_text(face = "bold", size = 12),
legend.position = "right")
ggsave("tilt/finding_beliefs_pre_post_engagement.png",
width = 11, height = 4.5, dpi = 300)