Resources

An educational researcher compares a peer-tutoring method to standard lecture instruction for an introductory statistics unit. A sample of n1=64n_1 = 64 students taught via peer tutoring achieves a mean exam score of xˉ1=78.4\bar{x}_1 = 78.4 with s1=8.2s_1 = 8.2, while a sample of n2=64n_2 = 64 students taught via traditional lecture achieves xˉ2=74.0\bar{x}_2 = 74.0 with s2=7.8s_2 = 7.8. (a) Determine the pooled standard deviation spooleds_{\text{pooled}} and the standardized effect size (Cohen's dd). (b) Rather than reporting a simple pp-value, calculate the 95%95\% confidence interval for the true difference in population means μ1−μ2\mu_1 - \mu_2 using degrees of freedom df=126df = 126 (critical value t∗≈1.979t^* \approx 1.979). (c) Interpret whether the difference is both statistically significant at α=0.05\alpha = 0.05 and practically significant if course coordinators consider any effect size with d≥0.50d \ge 0.50 practically meaningful.

Answer and steps

Final answer

(a) spooled=(64−1)(8.2)2+(64−1)(7.8)264+64−2=67.24+60.842=64.04≈8.002s_{\text{pooled}} = \sqrt{\frac{(64 - 1)(8.2)^2 + (64 - 1)(7.8)^2}{64 + 64 - 2}} = \sqrt{\frac{67.24 + 60.84}{2}} = \sqrt{64.04} \approx 8.002; Cohen's d=78.4−74.08.002=4.48.002≈0.550d = \frac{78.4 - 74.0}{8.002} = \frac{4.4}{8.002} \approx 0.550. (b) Standard error SE=8.002164+164=8.002×0.17678≈1.415SE = 8.002 \sqrt{\frac{1}{64} + \frac{1}{64}} = 8.002 \times 0.17678 \approx 1.415; Margin of error ME=1.979×1.415≈2.80ME = 1.979 \times 1.415 \approx 2.80; 95%95\% CI is 4.4±2.804.4 \pm 2.80, or [1.60,7.20][1.60, 7.20] points. (c) The intervention is statistically significant at α=0.05\alpha = 0.05 because 00 is not contained in the 95%95\% CI ([1.60,7.20][1.60, 7.20]), and it is practically significant because d=0.550≥0.50d = 0.550 \ge 0.50, indicating a moderate, pedagogically substantial improvement.

Steps

Part (a): Calculate the Pooled Standard Deviation and Cohen's dd

When comparing two independent groups with assumed equal variances, the pooled sample standard deviation spooleds_{\text{pooled}} is a weighted average of the two sample standard deviations:

spooled=(n1−1)s12+(n2−1)s22n1+n2−2s_{\text{pooled}} = \sqrt{\frac{(n_1 - 1)s_1^2 + (n_2 - 1)s_2^2}{n_1 + n_2 - 2}}

Since both groups have equal sample sizes (n1=n2=64n_1 = n_2 = 64), the formula simplifies to the root-mean-square of the two sample standard deviations:

spooled=63(8.2)2+63(7.8)264+64−2=67.24+60.842=64.04≈8.002s_{\text{pooled}} = \sqrt{\frac{63(8.2)^2 + 63(7.8)^2}{64 + 64 - 2}} = \sqrt{\frac{67.24 + 60.84}{2}} = \sqrt{64.04} \approx 8.002

Next, Cohen's dd measures the standardized effect size, representing the mean difference in units of pooled standard deviation:

d=xˉ1−xˉ2spooled=78.4−74.08.002=4.48.002≈0.550d = \frac{\bar{x}_1 - \bar{x}_2}{s_{\text{pooled}}} = \frac{78.4 - 74.0}{8.002} = \frac{4.4}{8.002} \approx 0.550


Part (b): Calculate the 95%95\% Confidence Interval for μ1−μ2\mu_1 - \mu_2

  1. Calculate the Standard Error (SESE):

    SE=spooled1n1+1n2=8.002164+164=8.002264=8.002×132≈8.002×0.17678≈1.415SE = s_{\text{pooled}} \sqrt{\frac{1}{n_1} + \frac{1}{n_2}} = 8.002 \sqrt{\frac{1}{64} + \frac{1}{64}} = 8.002 \sqrt{\frac{2}{64}} = 8.002 \times \frac{1}{\sqrt{32}} \approx 8.002 \times 0.17678 \approx 1.415

  2. Calculate the Margin of Error (MEME): Using the given critical value t∗≈1.979t^* \approx 1.979 for df=64+64−2=126df = 64 + 64 - 2 = 126:

    ME=t∗⋅SE=1.979×1.415≈2.80ME = t^* \cdot SE = 1.979 \times 1.415 \approx 2.80

  3. Construct the 95%95\% Confidence Interval: The point estimate of the difference is:

    xˉ1−xˉ2=78.4−74.0=4.4\bar{x}_1 - \bar{x}_2 = 78.4 - 74.0 = 4.4

    Therefore, the interval is:

    (xˉ1−xˉ2)±ME=4.4±2.80  ⟹  [1.60, 7.20] points(\bar{x}_1 - \bar{x}_2) \pm ME = 4.4 \pm 2.80 \implies [1.60,\, 7.20]\text{ points}


Part (c): Evaluate Statistical and Practical Significance

  • Statistical Significance: A hypothesis test at significance level α=0.05\alpha = 0.05 checks whether the null value of zero difference (μ1−μ2=0\mu_1 - \mu_2 = 0) is plausible. Because 00 is entirely outside the 95%95\% confidence interval [1.60, 7.20][1.60,\, 7.20], the observed difference is statistically significant at α=0.05\alpha = 0.05.
  • Practical Significance: Practical significance depends on the magnitude of the effect in context rather than just sample size. Since the standardized effect size is d=0.550d = 0.550, which exceeds the coordinator's threshold of d≥0.50d \ge 0.50 (conventionally considered a medium-to-large effect in education), the tutoring intervention also demonstrates practical significance.

Pitfall to Avoid: A statistically significant result (p<0.05p < 0.05 or 0∉CI0 \notin \text{CI}) merely indicates that the observed difference is unlikely due to random sampling variability; it does not guarantee that the difference is large enough to be meaningful in the real world. Both statistical significance and effect size must be evaluated together.

Typical errors

OverviewACEpath Teacher's Diagnosis

01 The student averages the sample standard deviations directly instead of pooling variances, which distorts the pooled standard deviation and effect size calculation.

Sample Student Work

Reviewed by ACEpath Teacher
Step-by-Step Review
Walk Through It with Your Teacher
Partially correct
  1. Step 1

    spooled=s1+s22=8.2+7.82=8.0s_{\text{pooled}} = \frac{s_1 + s_2}{2} = \frac{8.2 + 7.8}{2} = 8.0

    Correct

    spooled=(n1−1)s12+(n2−1)s22n1+n2−2=63(8.2)2+63(7.8)2126≈8.002s_{\text{pooled}} = \sqrt{\frac{(n_1 - 1)s_1^2 + (n_2 - 1)s_2^2}{n_1 + n_2 - 2}} = \sqrt{\frac{63(8.2)^2 + 63(7.8)^2}{126}} \approx 8.002

    You averaged the standard deviations directly (s1s_1 and s2s_2). Standard deviations cannot be averaged linearly; you must pool the sample variances (s12s_1^2 and s22s_2^2) by taking their weighted mean and then taking the square root.

  2. Step 2

    d=78.4−74.08.0=4.48.0=0.55d = \frac{78.4 - 74.0}{8.0} = \frac{4.4}{8.0} = 0.55

    Correct

    d=xˉ1−xˉ2spooled=4.48.002≈0.550d = \frac{\bar{x}_1 - \bar{x}_2}{s_{\text{pooled}}} = \frac{4.4}{8.002} \approx 0.550

    Your formula for Cohen's dd is correct, but your calculation carried forward the unpooled value spooled=8.0s_{\text{pooled}} = 8.0 from Step 1.

  3. Step 3

    SE=8.0×164+164=8.0×264=8.0×0.1768=1.414  ⟹  ME=1.979×1.414=2.80  ⟹  CI=4.4±2.80=(1.60,7.20)SE = 8.0 \times \sqrt{\frac{1}{64} + \frac{1}{64}} = 8.0 \times \sqrt{\frac{2}{64}} = 8.0 \times 0.1768 = 1.414 \implies ME = 1.979 \times 1.414 = 2.80 \implies CI = 4.4 \pm 2.80 = (1.60, 7.20)

    Correct

    SE=8.002164+164≈1.415  ⟹  ME=1.979×1.415≈2.80  ⟹  4.4±2.80=[1.60,7.20]SE = 8.002 \sqrt{\frac{1}{64} + \frac{1}{64}} \approx 1.415 \implies ME = 1.979 \times 1.415 \approx 2.80 \implies 4.4 \pm 2.80 = [1.60, 7.20]

    Your steps for standard error, margin of error, and confidence interval are mathematically sound, though they relied on the unpooled standard deviation from Step 1.

  4. Step 4

    The difference is statistically significant because 00 is not in the confidence interval. It is practically significant because d=0.55≥0.50d = 0.55 \ge 0.50.

WHAT YOU DID WELL
Compliments from Your Teacher

You correctly applied the formula for Cohen's dd using the mean difference and a standard deviation estimate.

You calculated the standard error, margin of error, and 95%95\% confidence interval correctly from your intermediate values.

You properly distinguished statistical significance (zero outside the confidence interval) from practical significance (d≥0.50d \ge 0.50).

THE ERROR
Where Your ACEpath Teacher Spotted the Mistake
Averaged standard deviations instead of pooling variances

no

WHY IT HAPPENED
ACEpath Teacher's Diagnosis

You treated sample standard deviations as additive quantities rather than recognizing that variance is the additive measure of dispersion, requiring a root-mean-square combination.

REASONING PATH
Your ACEpath teacher’s suggested approach
Pooling Variances vs. Averaging Standard Deviations
  • spooled2=(n1−1)s12+(n2−1)s22n1+n2−2s_{\text{pooled}}^2 = \frac{(n_1 - 1)s_1^2 + (n_2 - 1)s_2^2}{n_1 + n_2 - 2}, so spooled=spooled2s_{\text{pooled}} = \sqrt{s_{\text{pooled}}^2}

  • Never compute s1+s22\frac{s_1 + s_2}{2}; always square the sample standard deviations, weight by their degrees of freedom, and take the square root of the pooled variance.

CORRECTED PATH
ACEpath Teacher's Demonstration
  1. Compute the pooled variance: spooled2=63(8.22)+63(7.82)126=67.24+60.842=64.04s_{\text{pooled}}^2 = \frac{63(8.2^2) + 63(7.8^2)}{126} = \frac{67.24 + 60.84}{2} = 64.04

  2. Take the square root: spooled=64.04≈8.002s_{\text{pooled}} = \sqrt{64.04} \approx 8.002

  3. Recalculate effect size: d=4.48.002≈0.550d = \frac{4.4}{8.002} \approx 0.550

  4. Recalculate standard error: SE=8.002×264≈1.415SE = 8.002 \times \sqrt{\frac{2}{64}} \approx 1.415

Final answer

(a) spooled≈8.002s_{\text{pooled}} \approx 8.002, d≈0.550d \approx 0.550; (b) 95% CI=[1.60,7.20]95\%\text{ CI} = [1.60, 7.20]; (c) Statistically significant because 0∉[1.60,7.20]0 \notin [1.60, 7.20]; practically significant because d=0.550≥0.50d = 0.550 \ge 0.50.

NEXT STEPS
Final Takeaways from Your Teacher
  • Memorize that variances add and average, whereas standard deviations do not.

  • When sample sizes are equal (n1=n2n_1 = n_2), remember that spooled=s12+s222s_{\text{pooled}} = \sqrt{\frac{s_1^2 + s_2^2}{2}} rather than s1+s22\frac{s_1 + s_2}{2}.

OverviewACEpath Teacher's Diagnosis

02 The student mistakenly omits the square root over the sample size term when calculating standard error, leading to an erroneously wide confidence interval.

Sample Student Work

Reviewed by ACEpath Teacher
Step-by-Step Review
Walk Through It with Your Teacher
Partially correct
  1. Step 1

    spooled2=63×8.22+63×7.82126=67.24+60.842=64.04  ⟹  spooled=64.04=8.00s_{\text{pooled}}^2 = \frac{63 \times 8.2^2 + 63 \times 7.8^2}{126} = \frac{67.24 + 60.84}{2} = 64.04 \implies s_{\text{pooled}} = \sqrt{64.04} = 8.00 \n d=78.4−74.08.00=4.48.00=0.55d = \frac{78.4 - 74.0}{8.00} = \frac{4.4}{8.00} = 0.55

  2. Step 2

    SE=spooled×(164+164)=8.00×(264)=0.25SE = s_{\text{pooled}} \times \left(\frac{1}{64} + \frac{1}{64}\right) = 8.00 \times \left(\frac{2}{64}\right) = 0.25

    Correct

    SE=spooled1n1+1n2=8.002164+164≈8.002×0.1768≈1.415SE = s_{\text{pooled}} \sqrt{\frac{1}{n_1} + \frac{1}{n_2}} = 8.002 \sqrt{\frac{1}{64} + \frac{1}{64}} \approx 8.002 \times 0.1768 \approx 1.415

    You omitted the square root over the sample size factor (1n1+1n2)\left(\frac{1}{n_1} + \frac{1}{n_2}\right). The standard error is the square root of the variance of the difference between means.

  3. Step 3

    ME=1.979×0.25=0.49  ⟹  CI=4.4±0.49=(3.91,4.89)ME = 1.979 \times 0.25 = 0.49 \implies CI = 4.4 \pm 0.49 = (3.91, 4.89)

    Correct

    ME=1.979×1.415≈2.80  ⟹  CI=4.4±2.80=[1.60,7.20]ME = 1.979 \times 1.415 \approx 2.80 \implies CI = 4.4 \pm 2.80 = [1.60, 7.20]

    You correctly multiplied your standard error by t∗=1.979t^* = 1.979 and constructed the interval around xˉ1−xˉ2=4.4\bar{x}_1 - \bar{x}_2 = 4.4, but the numerical limits were affected by the standard error calculation error in Step 2.

  4. Step 4

    Statistically significant because 00 is not in the interval (3.91,4.89)(3.91, 4.89). \n Practically significant because d=0.55≥0.50d = 0.55 \ge 0.50.

WHAT YOU DID WELL
Compliments from Your Teacher

You correctly calculated the pooled standard deviation and Cohen's dd.

You correctly identified the critical value multiplication step for the margin of error.

You correctly interpreted statistical significance using the confidence interval.

You properly evaluated practical significance against the designated effect size threshold.

THE ERROR
Where Your ACEpath Teacher Spotted the Mistake
Omitted square root in standard error formula

You forgot to place the radical over the term (1n1+1n2)\left(\frac{1}{n_1} + \frac{1}{n_2}\right) when calculating the two-sample standard error.

WHY IT HAPPENED
ACEpath Teacher's Diagnosis

Confusing the pooled variance formula with the standard error formula: the pooled variance σ2\sigma^2 is scaled by (1n1+1n2)\left(\frac{1}{n_1} + \frac{1}{n_2}\right), so the standard error (which is a standard deviation) must place that factor under a square root: SE=spooled1n1+1n2SE = s_{\text{pooled}}\sqrt{\frac{1}{n_1} + \frac{1}{n_2}}.

REASONING PATH
Your ACEpath teacher’s suggested approach
Standard Error Formula Structure
  • SE(xˉ1−xˉ2)=s12n1+s22n2=spooled1n1+1n2SE(\bar{x}_1 - \bar{x}_2) = \sqrt{\frac{s_1^2}{n_1} + \frac{s_2^2}{n_2}} = s_{\text{pooled}}\sqrt{\frac{1}{n_1} + \frac{1}{n_2}}

  • Always check that the sample sizes nn inside the standard error are under a square root, ensuring that units match standard deviations rather than variances.

CORRECTED PATH
ACEpath Teacher's Demonstration
  1. Recalculate the standard error with the square root included: SE=8.002×264=8.002×28≈1.415SE = 8.002 \times \sqrt{\frac{2}{64}} = 8.002 \times \frac{\sqrt{2}}{8} \approx 1.415.

  2. Recalculate the margin of error: ME=t∗×SE=1.979×1.415≈2.80ME = t^* \times SE = 1.979 \times 1.415 \approx 2.80.

  3. Construct the corrected 95%95\% confidence interval: (xˉ1−xˉ2)±ME=4.40±2.80=[1.60,7.20](\bar{x}_1 - \bar{x}_2) \pm ME = 4.40 \pm 2.80 = [1.60, 7.20].

Final answer

(a) spooled≈8.00s_{\text{pooled}} \approx 8.00, d≈0.55d \approx 0.55; (b) 95%95\% CI is [1.60,7.20][1.60, 7.20]; (c) Statistically significant because 0∉[1.60,7.20]0 \notin [1.60, 7.20], and practically significant because d=0.55≥0.50d = 0.55 \ge 0.50.

NEXT STEPS
Final Takeaways from Your Teacher
  • Review the dimensional consistency of formulas: standard deviation and standard error must have the same units as the original data, meaning nn must be square-rooted.

  • Practice calculating pooled standard error across equal and unequal sample sizes.

OverviewACEpath Teacher's Diagnosis

03 The student misinterprets practical significance as requiring d to strictly exceed rather than be greater than or equal to the threshold, or confuses the threshold rule.

Sample Student Work

Reviewed by ACEpath Teacher
Step-by-Step Review
Walk Through It with Your Teacher
Partially correct
  1. Step 1

    spooled2=63×8.22+63×7.82126=64.04  ⟹  spooled=64.04=8.00  ⟹  d=78.4−74.08.00=0.55s_{\text{pooled}}^2 = \frac{63 \times 8.2^2 + 63 \times 7.8^2}{126} = 64.04 \implies s_{\text{pooled}} = \sqrt{64.04} = 8.00 \implies d = \frac{78.4 - 74.0}{8.00} = 0.55

  2. Step 2

    SE=8.00×164+164=1.414  ⟹  ME=1.979×1.414=2.80  ⟹  CI=4.4±2.80=(1.60,7.20)SE = 8.00 \times \sqrt{\frac{1}{64} + \frac{1}{64}} = 1.414 \implies ME = 1.979 \times 1.414 = 2.80 \implies CI = 4.4 \pm 2.80 = (1.60, 7.20)

  3. Step 3

    Statistically significant because the 95%95\% confidence interval (1.60,7.20)(1.60, 7.20) excludes 00.

  4. Step 4

    Not practically significant because statistical significance does not imply practical significance, and the lower bound of difference 1.601.60 is too small to be meaningful.

    Correct

    Practically significant because d=0.550≥0.50d = 0.550 \ge 0.50, meeting the coordinator's predefined practical significance threshold.

    While it is true that statistical significance does not guarantee practical significance, the prompt defined practical significance using the objective threshold d≥0.50d \ge 0.50. Since your calculated value of d=0.55d = 0.55 satisfies 0.55≥0.500.55 \ge 0.50, the tutoring intervention is practically significant.

WHAT YOU DID WELL
Compliments from Your Teacher

You correctly calculated the pooled standard deviation and standardized effect size (dd).

You correctly computed the standard error, margin of error, and confidence interval.

You accurately evaluated statistical significance using the confidence interval.

THE ERROR
Where Your ACEpath Teacher Spotted the Mistake
Ignored Effect Size Benchmark for Practical Significance

conceptual

WHY IT HAPPENED
ACEpath Teacher's Diagnosis

You correctly recognized the conceptual distinction between statistical and practical significance, but you relied on an informal, subjective assessment of the confidence interval's lower bound instead of checking the standardized effect size against the given operational threshold (d≥0.50d \ge 0.50).

REASONING PATH
Your ACEpath teacher’s suggested approach
Statistical Significance Criterion
  • Statistical significance at α=0.05\alpha = 0.05 occurs when the pp-value <0.05< 0.05, or equivalently when 00 lies outside the 95%95\% confidence interval for μ1−μ2\mu_1 - \mu_2.

  • Check whether 00 is contained within the confidence interval.

Practical Significance Criterion
  • Practical significance is determined by comparing an effect size measure (such as Cohen's dd) against a domain-specific threshold (here, d≥0.50d \ge 0.50).

  • Compare the calculated effect size dd directly to the established cutoff value rather than making an arbitrary judgment on raw point values.

CORRECTED PATH
ACEpath Teacher's Demonstration
  1. Identify the given criterion for practical significance: d≥0.50d \ge 0.50.

  2. Recall the calculated effect size from part (a): d=0.55d = 0.55.

  3. Compare the value to the threshold: 0.55≥0.500.55 \ge 0.50 is true.

  4. Conclude that the intervention satisfies the criterion for practical significance.

Final answer

The difference is both statistically significant (because 0∉[1.60,7.20]0 \notin [1.60, 7.20]) and practically significant (because d=0.55≥0.50d = 0.55 \ge 0.50).

NEXT STEPS
Final Takeaways from Your Teacher
  • Review the formal definition of Cohen's dd thresholds (small ≈0.2\approx 0.2, medium ≈0.5\approx 0.5, large ≈0.8\approx 0.8).

  • When a problem specifies an explicit rule or threshold for practical significance, always use that quantitative benchmark instead of a subjective impression.

Lock it in

Similar question

01

Two classes take a chemistry test. Group A has n1=31n_1 = 31, s1=6.0s_1 = 6.0; Group B has n2=31n_2 = 31, s2=8.0s_2 = 8.0. Calculate the pooled standard deviation spooleds_{\text{pooled}}.

Answer

Since n1=n2=31n_1 = n_2 = 31, spooled=s12+s222=36+642=50≈7.071s_{\text{pooled}} = \sqrt{\frac{s_1^2 + s_2^2}{2}} = \sqrt{\frac{36 + 64}{2}} = \sqrt{50} \approx 7.071.

02

In an experiment comparing two training programs, xˉ1−xˉ2=3.6\bar{x}_1 - \bar{x}_2 = 3.6 points with spooled=12.0s_{\text{pooled}} = 12.0. The 95%95\% confidence interval for μ1−μ2\mu_1 - \mu_2 is [0.40,6.80][0.40, 6.80]. Is the difference statistically significant at α=0.05\alpha = 0.05? Is it practically significant if the threshold for a meaningful effect is d≥0.40d \ge 0.40?

Answer
  1. Statistically significant: Yes, because 00 lies entirely outside [0.40,6.80][0.40, 6.80]. 2. Practical significance: d=3.612.0=0.30d = \frac{3.6}{12.0} = 0.30. Since 0.30<0.400.30 < 0.40, it is not practically significant.
03

Group 1 has n1=21n_1 = 21, xˉ1=52.0\bar{x}_1 = 52.0, s1=4.0s_1 = 4.0. Group 2 has n2=41n_2 = 41, xˉ2=47.0\bar{x}_2 = 47.0, s2=5.0s_2 = 5.0. Compute spooleds_{\text{pooled}} and Cohen's dd.

Answer
  1. df1=20df_1 = 20, df2=40df_2 = 40, dftotal=60df_{\text{total}} = 60. 2. spooled=20(42)+40(52)60=320+100060=132060=22≈4.690s_{\text{pooled}} = \sqrt{\frac{20(4^2) + 40(5^2)}{60}} = \sqrt{\frac{320 + 1000}{60}} = \sqrt{\frac{1320}{60}} = \sqrt{22} \approx 4.690. 3. d=52.0−47.04.690=5.04.690≈1.066d = \frac{52.0 - 47.0}{4.690} = \frac{5.0}{4.690} \approx 1.066.
Loading workspace…