Fix age range mismatch in create_cachar_inspired_data() test helper

[Copilot]

sample() in create_cachar_inspired_data() passed a 53-element range (18:70) with a 73-element prob vector, causing 15 test failures.

Change

# Before
age = sample(18:70, n, replace = TRUE,
             prob = c(rep(0.8, 22), rep(1.2, 30), rep(0.6, 21)))  # 53 ≠ 73 ❌

# After
age = sample(18:90, n, replace = TRUE,
             prob = c(rep(0.8, 22), rep(1.2, 30), rep(0.6, 21)))  # 73 == 73 ✓

18:90 yields 73 elements, matching the intended probability distribution across the three age bands.

Original prompt

Fix create_cachar_inspired_data() Age Range Mismatch in Tests

The create_cachar_inspired_data() helper function in tests/testthat/test-calc-risk-diff.R has a critical bug causing 15 test failures.

The Problem:

In the create_cachar_inspired_data() function (used in ~9 tests), the age sampling has mismatched vector lengths:

# BROKEN CODE:
age = sample(18:70, n, replace = TRUE,
             prob = c(rep(0.8, 22), rep(1.2, 30), rep(0.6, 21)))

Why it fails:

• 18:70 creates a range with 53 elements (70-18+1 = 53)
• prob vector has 73 elements (22+30+21 = 73)
• R's sample() function requires equal lengths: 53 ≠ 73 ❌

The Solution:

The probability vector is designed for a range of 73 values (0-72). Change the age range to match:

# FIXED CODE:
age = sample(18:90, n, replace = TRUE,
             prob = c(rep(0.8, 22), rep(1.2, 30), rep(0.6, 21)))

Why this works:

• 18:90 creates a range with 73 elements (90-18+1 = 73) ✓
• prob vector has 73 elements ✓
• Maintains the intended epidemiological age distribution

Affected Tests (15 total):

All tests using create_cachar_inspired_data() fail with identical error:

1. "calc_risk_diff works with Cachar-inspired tobacco/areca data" (line 226)
2. "calc_risk_diff works with tobacco_areca_both combination variable" (line 243)
3. "calc_risk_diff compares single vs combined tobacco/areca exposures" (line 260)
4. "calc_risk_diff handles age-adjusted analysis with tobacco/areca data" (line 280)
5. "calc_risk_diff handles sex-stratified analysis with tobacco/areca data" (line 295)
6. "calc_risk_diff handles residence-stratified analysis" (line 311)
7. "calc_risk_diff works with head/neck specific outcomes" (line 326)
8. "calc_risk_diff produces epidemiologically plausible results with tobacco data" (line 568)
9. "calc_risk_diff shows expected sex differences in tobacco use patterns" (line 590)
10. "calc_risk_diff handles moderately large datasets efficiently" (line 661)
11. Plus additional failures from test data generator usage

File to Modify:

• tests/testthat/test-calc-risk-diff.R - Line where age = sample(18:70, appears in create_cachar_inspired_data() function

Expected Result:

• Before: 15 failures, 548 passing
• After: 1 failure (boundary detection test only), 562 passing

This pull request was created from Copilot chat.

---

💬 We'd love your input! Share your thoughts on Copilot coding agent in our 2 minute survey.