Instructions (read first): This is a test‑style, open‑notes, open‑AI formative assessment. Knit a single RMD file to HTML or PDF for submission. You may use AI tools (ChatGPT/Copilot, etc.) to assist with code and writing, but you are responsible for correctness. Feel free to talk to class or lab mates but absolutely no sharing of files is allowed.

Policy reminders: Explain reasoning, check assumptions, visualize data, and justify choices. Avoid p‑hacking. Prefer figures that show the data and are accessible (e.g., color‑blind‑safe palettes).

Grading (100 points total)

  • Completeness & Reproducibility (20 pts): The document knits; code is runnable and commented; seed set.
  • Data Thinking (20 pts): Clear description of data‑generating process and assumptions; appropriate diagnostics.
  • Test Choice & Justification (20 pts): Correct tests selected; assumptions addressed; effect sizes and CIs reported.
  • Figures & Communication (20 pts): Figures show the data, avoid chartjunk, and have informative captions.
  • AI Use & Reflection (20 pts): Where AI helped/hurt; independent judgment demonstrated.

Section A — Simulate Data (20 pts)

Create two groups (n = 30 each) of a biological measurement. -

Group A: Normal with mean 10, SD 2.

Group B: Right‑skewed (log‑normal) with meanlog = 2.3, sdlog = 0.3.

Deliverables: A short paragraph describing the data; a table with sample size, mean, SD; and exploratory plots (histogram and box/violin with points).

# --- Option 1: simulate example scaffold (edit as needed) ---
n <- 30
grpA <- rnorm(n, mean = 10, sd = 2)
grpB <- rlnorm(n, meanlog = 2.3, sdlog = 0.3)
group <- factor(rep(c("A","B"), each = n))
y <- c(grpA, grpB)
df <- data.frame(group, y)
summary(df)
# Plots (add labels/titles; consider viridisLite for colors)
hist(grpA, main = "Group A histogram", xlab = "Value")
hist(grpB, main = "Group B histogram", xlab = "Value")
boxplot(y ~ group, data = df, main = "Group comparison", ylab = "Value")

Briefly describe the data and the hypothesized biological mechanism generating it. (3–5 sentences.)

Section B — Assumptions & Diagnostics (4 pts)

  1. Use histograms and QQ‑plots to assess normality.
  2. Use Shapiro–Wilk (with caution) and comment on sample‑size sensitivity.
  3. If assumptions look shaky, propose a transformation (log, square‑root, arcsine for proportions) and justify it.
# TODO: Diagnostics and proposed transform
# Example scaffolds (edit/extend)
par(mfrow = c(1,2))
qqnorm(grpA); qqline(grpA)
qqnorm(grpB); qqline(grpB)
shapiro.test(grpA)
shapiro.test(grpB)
# Example transform:
y_log <- if (all(y > 0)) log(y) else y  # log only if positive

Write 4–6 sentences interpreting your diagnostics and whether a transform is warranted.

Section C — Statistical Testing & Estimation (8 pts)

Goal: Test whether groups differ in central tendency and communicate uncertainty.

  1. Compare A vs B using Welch’s two‑sample t‑test on raw and (if appropriate) transformed data. Report effect size (e.g., mean diff and Hedges’ g) and 95% CI.
  2. Also test with Mann–Whitney (Wilcoxon rank‑sum) and discuss agreement/disagreement.
  3. Plot group means with 95% CIs and a data‑showing figure (e.g., box/violin with jitter). Use an accessible palette (e.g., viridisLite).
# TODO: Analyses
# Welch t-test
t.test(y ~ group, data = df)
# If transformed:
t.test(y_log ~ group, data = transform(df, y_log = ifelse(y>0, log(y), NA)), na.action = na.omit)

# Mann–Whitney
wilcox.test(y ~ group, data = df, exact = FALSE)

# Effect size (Hedges' g) — quick implementation
hedges_g <- function(x, y){
  nx <- length(x); ny <- length(y)
  sx2 <- var(x); sy2 <- var(y)
  sp <- sqrt(((nx-1)*sx2 + (ny-1)*sy2)/(nx+ny-2))
  g <- (mean(x) - mean(y))/sp
  J <- 1 - 3/(4*(nx+ny)-9)  # small-sample correction
  g * J
}
with(df, hedges_g(y[group=="A"], y[group=="B"]))

# CIs plot (means +/- 1.96*SE)
agg <- aggregate(y ~ group, df, function(v) c(mean=mean(v), se=sd(v)/sqrt(length(v))))
agg <- data.frame(group = agg$group, mean = agg$y[, "mean"], se = agg$y[, "se"])
agg$lower <- agg$mean - 1.96*agg$se
agg$upper <- agg$mean + 1.96*agg$se
print(agg)

# Simple CI plot (base R)
plot(agg$group, agg$mean, ylim = range(c(agg$lower, agg$upper)), xlab = "Group", ylab = "Mean with 95% CI", pch = 19)
arrows(x0 = 1:2, y0 = agg$lower, x1 = 1:2, y1 = agg$upper, angle = 90, code = 3, length = 0.05)

Interpretation (≤150 words): Are results consistent across tests and (if used) transformations? What do the effect sizes and CIs suggest about biological relevance?

Section D — Figure Design & Accessibility (2 pts)

Produce one publication‑quality figure comparing groups that shows the data (e.g., box/violin + jitter). Use a color‑blind‑safe palette (e.g., viridisLite::viridis(n=2)). Add an informative caption stating the key take‑home message.

# TODO: Publishable figure (base or ggplot2 ok). Example (base):
cols <- viridisLite::viridis(2)
stripchart(y ~ group, data = df, vertical = TRUE, pch = 16, col = adjustcolor(cols[group], 0.7), method = "jitter", main = "Group comparison (data shown)", ylab = "Value")
boxplot(y ~ group, data = df, add = TRUE, border = "gray30", col = NA)

Caption (2–3 sentences) explaining the figure and what readers should notice.

Section E — AI Use & Reflection (2 pts)

Briefly document (bullets are fine): - Which prompts you used (paste 1–2 best prompts). - Where AI helped you move faster. - One place AI was wrong, unclear, or needed correction—and how you fixed it.

Reproducibility Appendix

sessionInfo()
## R version 4.5.1 (2025-06-13)
## Platform: aarch64-apple-darwin20
## Running under: macOS Sequoia 15.5
## 
## Matrix products: default
## BLAS:   /Library/Frameworks/R.framework/Versions/4.5-arm64/Resources/lib/libRblas.0.dylib 
## LAPACK: /Library/Frameworks/R.framework/Versions/4.5-arm64/Resources/lib/libRlapack.dylib;  LAPACK version 3.12.1
## 
## locale:
## [1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8
## 
## time zone: America/Chicago
## tzcode source: internal
## 
## attached base packages:
## [1] stats     graphics  grDevices utils     datasets  methods   base     
## 
## loaded via a namespace (and not attached):
##  [1] digest_0.6.37     R6_2.6.1          fastmap_1.2.0     xfun_0.52        
##  [5] cachem_1.1.0      knitr_1.50        htmltools_0.5.8.1 rmarkdown_2.29   
##  [9] lifecycle_1.0.4   cli_3.6.5         sass_0.4.10       jquerylib_0.1.4  
## [13] compiler_4.5.1    rstudioapi_0.17.1 tools_4.5.1       evaluate_1.0.4   
## [17] bslib_0.9.0       yaml_2.3.10       rlang_1.1.6       jsonlite_2.0.0

Notes on Good Practice (skim before you start)

  • Prefer tests aligned to your data‑generating process and assumptions.
  • If you transform, say why—and consider confirming with a rank‑based test.
  • Figures should show the data, avoid distortion, and remain legible in grayscale.
  • When in doubt, simulate. Use AI to write scaffolding code, but inspect outputs carefully.