Problem 1
A manufacturing quality engineering team conducts a comparative tool-life study between two cutting tool insert grades in a high-speed CNC turning process: Grade 1 (Titanium-Nitride coated carbide) versus Grade 2 (Uncoated carbide). Independent random samples of n1 = 16 and n2 = 16 cutting edges were tested under identical cutting speed, feed rate, and workpiece hardness until reaching 0.30 mm flank wear failure criteria. Sample summary statistics: - Grade 1: x-bar_1 = 142.5 minutes, s1 = 12.0 minutes - Grade 2: x-bar_2 = 128.0 minutes, s2 = 14.5 minutes Assuming tool lives follow normal distributions with equal population variances (σ1² = σ2²), perform a two-sided hypothesis test at significance level α = 0.05 to evaluate H0: μ1 - μ2 = 0 versus Ha: μ1 - μ2 ≠ 0. Critical value: t_0.025, 30 = 2.042. Calculate the pooled two-sample t-statistic and the 95% confidence interval for the true mean difference in tool life (μ1 - μ2).
- ATest statistic t is 1.82 (do not reject null hypothesis, 95% CI [-1.55, 30.55] min)
- BTest statistic t is 2.45 (reject null hypothesis, 95% CI [2.15, 26.85] min)
- CTest statistic t is 3.08 (reject null hypothesis, 95% CI [4.89, 24.11] min)correct
- DTest statistic t is 3.75 (reject null hypothesis, 95% CI [7.20, 21.80] min)
Show the solution
**Step 1: Compute the Pooled Sample Variance ($s_p^2$)** Degrees of freedom: $$df = n_1 + n_2 - 2 = 16 + 16 - 2 = 30$$ Pooled variance formula: $$s_p^2 = \frac{(n_1 - 1) s_1^2 + (n_2 - 1) s_2^2}{n_1 + n_2 - 2} = \frac{(15)(12.0)^2 + (15)(14.5)^2}{30}$$ $$s_p^2 = \frac{15(144.0) + 15(210.25)}{30} = \frac{2{,}160.0 + 3{,}153.75}{30} = \frac{5{,}313.75}{30} = 177.125\text{ min}^2$$ Pooled standard deviation: $$s_p = \sqrt{177.125} = 13.3088\text{ minutes}$$ **Step 2: Calculate the Standard Error of the Mean Difference ($SE$)** $$SE = s_p \sqrt{\frac{1}{n_1} + \frac{1}{n_2}} = 13.3088 \sqrt{\frac{1}{16} + \frac{1}{16}} = 13.3088 \sqrt{\frac{2}{16}} = 13.3088 \times 0.35355 = 4.7054\text{ minutes}$$ **Step 3: Compute the Two-Sample $t$-Test Statistic** $$t = \frac{(\bar{x}_1 - \bar{x}_2) - \Delta_0}{SE} = \frac{(142.5 - 128.0) - 0}{4.7054} = \frac{14.50}{4.7054} = 3.0815 \approx 3.08$$ **Step 4: Hypothesis Testing Decision and 95% Confidence Interval** With critical value $t_{0.025, 30} = 2.042$: Since $|t| = 3.08 > 2.042$, we **reject the null hypothesis $H_0$** ($p\text{-value} = 0.0044 < 0.05$). Margin of error: $$E = t_{\alpha/2, df} \times SE = 2.042 \times 4.7054 = 9.608\text{ minutes}$$ 95% Confidence Interval for $\mu_1 - \mu_2$: $$(\bar{x}_1 - \bar{x}_2) \pm E = 14.50 \pm 9.61 = [4.89, \, 24.11]\text{ minutes}$$ Because zero is strictly outside the 95% confidence interval, Grade 1 (TiN-coated) inserts provide a statistically significant improvement of approximately 14.5 minutes in tool life.
Problem 2
An industrial engineer compares the cycle completion times (in seconds) of two automated assembly cell configurations. Independent random samples are collected: • Method 1: n_1 = 12, sample mean $\bar{x}_1 = 42.5\text{ s}$, sample standard deviation $s_1 = 3.2\text{ s}$ • Method 2: n_2 = 10, sample mean $\bar{x}_2 = 38.0\text{ s}$, sample standard deviation $s_2 = 2.8\text{ s}$ Assuming normally distributed cycle times with equal population variances ($\sigma_1^2 = \sigma_2^2$), calculate the pooled two-sample t-test statistic and identify its degrees of freedom (df).
- A2.84 test statistic t with df = 22 (Equal sample size assumption)
- B4.15 test statistic t with df = 20 (Neglecting sample variance weighting)
- C3.47 test statistic t with df = 20 (Statistically significant difference at alpha = 0.01)correct
- D1.98 test statistic t with df = 21 (Using Welch-Satterthwaite approximation)
Show the solution
**Step 1: Compute the Pooled Sample Variance (s_p^2)** From the FE Reference Handbook (Probability & Statistics / Hypothesis Tests): For two independent samples assuming equal population variances $\sigma_1^2 = \sigma_2^2$: $$s_p^2 = \frac{(n_1 - 1) s_1^2 + (n_2 - 1) s_2^2}{n_1 + n_2 - 2}$$ Given: - $n_1 = 12, s_1 = 3.2 \implies s_1^2 = 10.24$ - $n_2 = 10, s_2 = 2.8 \implies s_2^2 = 7.84$ - Degrees of freedom: $df = n_1 + n_2 - 2 = 12 + 10 - 2 = 20$ $$s_p^2 = \frac{(11)(10.24) + (9)(7.84)}{20} = \frac{112.64 + 70.56}{20} = \frac{183.20}{20} = 9.160\text{ s}^2$$ $$s_p = \sqrt{9.160} = 3.02655\text{ s}$$ **Step 2: Compute Standard Error of Difference** $$\text{SE}(\bar{x}_1 - \bar{x}_2) = s_p \sqrt{\frac{1}{n_1} + \frac{1}{n_2}} = 3.02655 \times \sqrt{\frac{1}{12} + \frac{1}{10}}$$ $$\frac{1}{12} + \frac{1}{10} = 0.083333 + 0.100000 = 0.183333$$ $$\text{SE} = 3.02655 \times \sqrt{0.183333} = 3.02655 \times 0.428174 = 1.29589\text{ s}$$ **Step 3: Compute t-Test Statistic** $$t = \frac{(\bar{x}_1 - \bar{x}_2) - 0}{\text{SE}} = \frac{42.5 - 38.0}{1.29589} = \frac{4.5}{1.29589} = 3.4725 \approx 3.47$$ With $df = 20$ and $t_{critical, 0.005} = 2.845$, the result is highly significant ($p < 0.01$). Conclusion: The test statistic is **3.47 with df = 20** (Option C).
Problem 3
A quality engineering manager designs a two-sided hypothesis test for the mean tensile strength of composite brackets ($H_0: \mu = \mu_0$ versus $H_1: \mu \ne \mu_0$). The population is normally distributed with known standard deviation $\sigma = 12.0\text{ MPa}$. The test significance level is set at $\alpha = 0.05$ ($z_{\alpha/2} = 1.960$). The manager requires a statistical power of $1 - \beta = 0.90$ ($\beta = 0.10$, $z_\beta = 1.282$) to detect a critical mean deviation of $\delta = |\mu_1 - \mu_0| = 4.50\text{ MPa}$. Using the standard sample size equation: $$n = \frac{(z_{\alpha/2} + z_\beta)^2 \sigma^2}{\delta^2}$$ Calculate the minimum required sample size (round up to the nearest integer, e.g., 50):
Show the solution
Step 1: Sum the critical standard normal critical values for significance and power: $$z_{\alpha/2} + z_\beta = 1.960 + 1.282 = 3.242$$ $$(z_{\alpha/2} + z_\beta)^2 = (3.242)^2 = 10.51056$$ Step 2: Evaluate the variance and squared effect size: $$\sigma^2 = (12.0\text{ MPa})^2 = 144.0\text{ MPa}^2$$ $$\delta^2 = (4.50\text{ MPa})^2 = 20.25\text{ MPa}^2$$ Step 3: Calculate sample size $n$ and round up to the next whole integer: $$n = \frac{10.51056 \times 144.0}{20.25} = \frac{1513.52}{20.25} = 74.741$$ Rounding up to ensure the specified power yields $n = 75$ samples. Distractor Analysis: - Omitting the power term $z_\beta$ yields $n = (1.96^2 \times 144) / 20.25 = 27.3 \rightarrow 28$ (providing only 50% power). - Rounding down to 74 slightly fails the 90% power requirement.