Welch t test: degrees of freedom and decision
A statistics problem from Statisticsdrill. Solve it with one answer box, with a fresh sample on every attempt and the worked steps and spreadsheet function in the answer, and it is checked the moment you submit.
The task
A plant compares the daily output of two machines. Machine A: n = 20, x̄ = 108, s = 9.7. Machine B: n = 17, x̄ = 112.1, s = 4.4 (units per day). The spreads look unequal, so use the Welch t test for H0: μ1 = μ2 against H1: μ1 ≠ μ2. Find the Welch-Satterthwaite degrees of freedom (do not round them down), then decide at α = 0.05. Round to 2 decimal places.
What this problem teaches
The Welch df is always between the smaller n - 1 and n1 + n2 - 2. Software keeps the fraction; some textbooks round it down, which is slightly conservative.
A hint if you get stuck
With a = s1²/n1 and b = s2²/n2: df = (a + b)² / [a²/(n1 - 1) + b²/(n2 - 1)]. The t statistic is (x̄1 - x̄2) / √(a + b).
Try it now, free, no account needed
The first 15 tasks on Statisticsdrill are free. Open it on the practice page, solve it, and get instant feedback with a hint. A free account adds the worked solution; Pro unlocks every task, timed mode and the certificate.
Open this problemMore problems in Inference
- Critical t value for an interval
- Margin of error for a mean (t)
- Confidence interval for a proportion
- Sample size for a mean
- Sample size for a proportion
- One-sample t statistic
- Two-sided t test: p-value and decision
- One-sided t test: p-value and decision
- Two-proportion z test
- Paired t test
- Chi-square test of independence (2 x 3)
- Chi-square 2 x 2: p-value and decision
- Two-sample t test (pooled)
- Confidence interval for a difference of means
- Chi-square goodness of fit
- One-way ANOVA F statistic
- Power of a one-sided z test
- Bonferroni: how many tests survive
- Benjamini-Hochberg: how many discoveries
- Margin of error for a mean (z, σ known)
- One-sample z test (σ known)
- One-proportion z test
- A/B test: sample size per group
- Effect size: Cohen's d
- Chi-square: expected count of a cell