Resources

A nutritionist measures the effect of a low-sodium diet on systolic blood pressure (XX) across n=16n = 16 hypertensive patients, recording the difference as Reduction=Xpre−Xpost\text{Reduction} = X_{\text{pre}} - X_{\text{post}}. The sample mean reduction is xˉ=+2.80 mmHg\bar{x} = +2.80\text{ mmHg} with sample standard deviation s=6.40 mmHgs = 6.40\text{ mmHg}. (a) State the null and alternative hypotheses for testing if the diet reduces blood pressure, set up the test statistic tt, and determine whether the result is statistically significant at α=0.05\alpha = 0.05 (tcrit=1.753t_{\text{crit}} = 1.753 for df=15df = 15). (b) A novice researcher incorrectly defines the difference as Δ=Xpost−Xpre\Delta = X_{\text{post}} - X_{\text{pre}}, obtains xˉ=−2.80 mmHg\bar{x} = -2.80\text{ mmHg}, and mistakenly concludes: 'Because t=−1.75t = -1.75 is negative and the pp-value for Ha:μ>0H_a: \mu > 0 is 0.950.95, the diet significantly raised blood pressure and was harmful.' Diagnose this sign convention trap and state the correct one-tailed pp-value for the test of blood pressure reduction. (c) If cardiologists define practical significance as a mean systolic reduction of at least 5.0 mmHg5.0\text{ mmHg}, evaluate the practical significance of the observed +2.80 mmHg+2.80\text{ mmHg} sample result.

Answer and steps

Final answer

(a) H0:μreduction=0H_0: \mu_{\text{reduction}} = 0 versus Ha:μreduction>0H_a: \mu_{\text{reduction}} > 0. t=2.80−06.4016=2.801.60=1.750t = \frac{2.80 - 0}{\frac{6.40}{\sqrt{16}}} = \frac{2.80}{1.60} = 1.750. Because t=1.750<1.753t = 1.750 < 1.753, we fail to reject H0H_0 at α=0.05\alpha = 0.05 (p≈0.0503>0.05p \approx 0.0503 > 0.05); the reduction is not statistically significant. (b) The researcher fell into a sign convention trap: defining post minus pre reverses the sign of improvement to negative, requiring Ha:μΔ<0H_a: \mu_{\Delta} < 0. The correct one-tailed pp-value for a reduction is P(T15≤−1.750)=P(T15≥+1.750)≈0.0503P(T_{15} \le -1.750) = P(T_{15} \ge +1.750) \approx 0.0503, not 0.950.95. (c) With xˉ=2.80 mmHg<5.0 mmHg\bar{x} = 2.80\text{ mmHg} < 5.0\text{ mmHg}, the observed sample mean does not reach the clinical benchmark of 5.0 mmHg5.0\text{ mmHg}; hence, it lacks practical significance while also failing to achieve statistical significance at α=0.05\alpha = 0.05.

Steps

Part (a): Hypotheses, Test Statistic, and Statistical Significance

  1. State the Hypotheses: Since the difference is defined as a reduction, Reduction=Xpre−Xpost\text{Reduction} = X_{\text{pre}} - X_{\text{post}}, a positive value indicates that blood pressure decreased.

    • Null hypothesis (H0H_0): The diet has no effect on blood pressure: H0:μreduction=0H_0: \mu_{\text{reduction}} = 0
    • Alternative hypothesis (HaH_a): The diet reduces blood pressure (one-tailed test): Ha:μreduction>0H_a: \mu_{\text{reduction}} > 0
  2. Compute the Standard Error (SESE): SE=sn=6.4016=6.404=1.60 mmHgSE = \frac{s}{\sqrt{n}} = \frac{6.40}{\sqrt{16}} = \frac{6.40}{4} = 1.60\text{ mmHg}
  3. Calculate the tt-statistic: t=xˉ−μ0SE=2.80−01.60=1.750t = \frac{\bar{x} - \mu_0}{SE} = \frac{2.80 - 0}{1.60} = 1.750
  4. Make the Decision:

    • Degrees of freedom: df=n−1=16−1=15df = n - 1 = 16 - 1 = 15.
    • The critical value for a one-tailed test at α=0.05\alpha = 0.05 is tcrit=1.753t_{\text{crit}} = 1.753.
    • Comparing the values: t=1.750<1.753t = 1.750 < 1.753
    • Because the test statistic is slightly less than the critical value (corresponding to p≈0.0503>0.05p \approx 0.0503 > 0.05), we fail to reject H0H_0 at the α=0.05\alpha = 0.05 significance level. The sample reduction is not statistically significant.

Part (b): Diagnosing the Sign Convention Trap

  1. The Nature of the Trap:

    • Defining the change as Δ=Xpost−Xpre\Delta = X_{\text{post}} - X_{\text{pre}} means a decrease in blood pressure produces a negative number (e.g., 140→137.2  ⟹  Δ=−2.80140 \rightarrow 137.2 \implies \Delta = -2.80).
    • Therefore, evidence of a reduction corresponds to the lower tail: Ha:μΔ<0H_a: \mu_{\Delta} < 0.
    • The researcher kept the alternative hypothesis as μ>0\mu > 0 while using the negative sample mean. Testing the upper tail when the test statistic is t=−1.75t = -1.75 computes the area to the right: P(T15≥−1.750)=1−0.0503≈0.95P(T_{15} \ge -1.750) = 1 - 0.0503 \approx 0.95
    • A large pp-value in the wrong tail simply means the data do not support an increase in blood pressure; it does not indicate that the diet was harmful.
  2. Correct One-Tailed pp-value:

    • For a reduction under the Δ=Xpost−Xpre\Delta = X_{\text{post}} - X_{\text{pre}} convention, the correct upper-tail or lower-tail equivalence is: P(T15≤−1.750)=P(T15≥+1.750)≈0.0503P(T_{15} \le -1.750) = P(T_{15} \ge +1.750) \approx 0.0503

Part (c): Evaluating Practical Significance

  • Statistical Significance vs. Practical Significance:
    • Statistical significance addresses whether an observed effect is likely due to chance given the sample size (p<αp < \alpha).
    • Practical significance (clinical relevance) addresses whether the magnitude of the effect is large enough to be meaningful in the real world.
  • Evaluation:
    • Cardiologists established that a clinically meaningful reduction must be at least 5.0 mmHg5.0\text{ mmHg}.
    • The observed mean reduction is xˉ=2.80 mmHg\bar{x} = 2.80\text{ mmHg}, which is substantially smaller than the 5.0 mmHg5.0\text{ mmHg} threshold (2.80<5.02.80 < 5.0).
    • Therefore, the observed reduction lacks practical significance, in addition to failing to meet the threshold for statistical significance at α=0.05\alpha = 0.05.

Typical errors

OverviewACEpath Teacher's Diagnosis

01 The student forgets to divide the sample standard deviation by the square root of n when computing the test statistic.

Sample Student Work

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

    H0:μ=0 vs. Ha:μ>0H_0: \mu = 0 \text{ vs. } H_a: \mu > 0

  2. Step 2

    t=2.806.40=0.4375t = \frac{2.80}{6.40} = 0.4375

    Correct

    t=xˉ−μ0sn=2.806.4016=2.801.60=1.750t = \frac{\bar{x} - \mu_0}{\frac{s}{\sqrt{n}}} = \frac{2.80}{\frac{6.40}{\sqrt{16}}} = \frac{2.80}{1.60} = 1.750

    You divided by the sample standard deviation s=6.40s = 6.40 instead of the standard error of the mean SE=sn=6.4016=1.60SE = \frac{s}{\sqrt{n}} = \frac{6.40}{\sqrt{16}} = 1.60.

  3. Step 3

    Since 0.44<1.753, fail to reject H0; not statistically significant.\text{Since } 0.44 < 1.753\text{, fail to reject } H_0\text{; not statistically significant.}

    Correct

    Since t=1.750<1.753, fail to reject H0 at α=0.05; not statistically significant.\text{Since } t = 1.750 < 1.753\text{, fail to reject } H_0\text{ at } \alpha = 0.05\text{; not statistically significant.}

    Your comparison logic was correct, but it relied on your incorrect tt-value of 0.440.44 rather than t=1.750t = 1.750.

  4. Step 4

    A negative Δ=Xpost−Xpre means post-diet blood pressure decreased, so blood pressure was reduced rather than raised.  ⟹  The correct lower-tail p-value is 1−0.95=0.05.\text{A negative } \Delta = X_{\text{post}} - X_{\text{pre}} \text{ means post-diet blood pressure decreased, so blood pressure was reduced rather than raised.} \implies \text{The correct lower-tail } p\text{-value is } 1 - 0.95 = 0.05.

  5. Step 5

    Because 2.80 mmHg<5.0 mmHg, the observed reduction does not achieve the required threshold and lacks practical significance.\text{Because } 2.80\text{ mmHg} < 5.0\text{ mmHg}\text{, the observed reduction does not achieve the required threshold and lacks practical significance.}

WHAT YOU DID WELL
Compliments from Your Teacher

You stated the correct one-sided hypotheses for blood pressure reduction.

You clearly identified that a negative post-pre difference corresponds to an improvement rather than harm.

You correctly calculated the one-tailed pp-value as 1−0.95=0.051 - 0.95 = 0.05.

You properly evaluated practical significance against the clinical threshold.

THE ERROR
Where Your ACEpath Teacher Spotted the Mistake
Omitted Standard Error Factor

You substituted the raw sample standard deviation ss into the denominator of the tt-statistic instead of the standard error sn\frac{s}{\sqrt{n}}.

WHY IT HAPPENED
ACEpath Teacher's Diagnosis

Confusing the spread of individual observations (ss) with the spread of the sample mean (SE=snSE = \frac{s}{\sqrt{n}}) when standardizing a sample statistic.

REASONING PATH
Your ACEpath teacher’s suggested approach
Standard Deviation vs. Standard Error
  • The denominator of a one-sample tt-test is always the standard error of the mean, SE=snSE = \frac{s}{\sqrt{n}}, not the individual standard deviation ss.

  • Always compute SE=snSE = \frac{s}{\sqrt{n}} first before forming the test statistic ratio.

CORRECTED PATH
ACEpath Teacher's Demonstration
  1. SE=sn=6.4016=6.404=1.60 mmHgSE = \frac{s}{\sqrt{n}} = \frac{6.40}{\sqrt{16}} = \frac{6.40}{4} = 1.60\text{ mmHg}

  2. t=xˉ−μ0SE=2.80−01.60=1.750t = \frac{\bar{x} - \mu_0}{SE} = \frac{2.80 - 0}{1.60} = 1.750

  3. Compare t=1.750 to tcrit=1.753: since 1.750<1.753, fail to reject H0 at α=0.05.\text{Compare } t = 1.750 \text{ to } t_{\text{crit}} = 1.753\text{: since } 1.750 < 1.753\text{, fail to reject } H_0\text{ at } \alpha = 0.05.

Final answer

(a) t=1.750<1.753t = 1.750 < 1.753, fail to reject H0H_0; not statistically significant. (b) Reversing subtraction makes reduction negative (Ha:μ<0H_a: \mu < 0); correct pp-value is ≈0.05\approx 0.05. (c) Because 2.80 mmHg<5.0 mmHg2.80\text{ mmHg} < 5.0\text{ mmHg}, the reduction lacks practical significance.

NEXT STEPS
Final Takeaways from Your Teacher
  • Double-check your test statistic formulas to ensure sample size nn is accounted for via SE=snSE = \frac{s}{\sqrt{n}}.

  • Practice calculating tt-statistics where sample sizes vary to reinforce the role of n\sqrt{n}.

OverviewACEpath Teacher's Diagnosis

02 The student uses degrees of freedom n = 16 to find the standard error or critical value rather than recognizing df = 15, and misinterprets t > 1.75 as significant by confusing borderline rounding.

Sample Student Work

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

    H0:μ=0 vs. Ha:μ>0  ⟹  SE=6.4016=1.60  ⟹  t=2.801.60=1.75H_0: \mu = 0 \text{ vs. } H_a: \mu > 0 \implies SE = \frac{6.40}{\sqrt{16}} = 1.60 \implies t = \frac{2.80}{1.60} = 1.75

  2. Step 2

    1.75≈1.753, so reject H0; the reduction is statistically significant.1.75 \approx 1.753\text{, so reject } H_0\text{; the reduction is statistically significant.}

    Correct

    Because t=1.750<1.753t = 1.750 < 1.753, fail to reject H0H_0; the reduction is not statistically significant at α=0.05\alpha = 0.05.

    In hypothesis testing, the rejection region requires the test statistic to strictly meet or exceed the critical threshold (t≥tcritt \ge t_{\text{crit}} for an upper-tailed test). Approximating 1.750≈1.7531.750 \approx 1.753 crosses the decision boundary; since 1.750<1.7531.750 < 1.753, the result fails to reach statistical significance.

  3. Step 3

    Negative Δ indicates a drop in blood pressure, not harm.  ⟹  The correct one-tailed p-value testing for reduction is 1−0.95=0.05.\text{Negative } \Delta \text{ indicates a drop in blood pressure, not harm.} \implies \text{The correct one-tailed } p\text{-value testing for reduction is } 1 - 0.95 = 0.05.

  4. Step 4

    Since 2.80 mmHg is less than the required 5.0 mmHg reduction, the result lacks practical significance.\text{Since } 2.80\text{ mmHg is less than the required } 5.0\text{ mmHg reduction, the result lacks practical significance.}

WHAT YOU DID WELL
Compliments from Your Teacher

You correctly set up the one-sided null and alternative hypotheses.

You calculated the standard error and test statistic with complete arithmetic accuracy.

You correctly diagnosed the sign reversal trap in part (b) and deduced the proper tail probability.

You accurately distinguished practical significance from statistical significance by comparing the sample mean to the clinical threshold.

THE ERROR
Where Your ACEpath Teacher Spotted the Mistake
Borderline critical value rounding

You treated t=1.75t = 1.75 as approximately equal to 1.7531.753 and used that approximation to reject H0H_0.

WHY IT HAPPENED
ACEpath Teacher's Diagnosis

You relaxed the strict inequality requirement for statistical significance. In critical-value testing, the test statistic must fall strictly inside the critical region (t≥tcritt \ge t_{\text{crit}}). Even a tiny gap where t<tcritt < t_{\text{crit}} means the data remain consistent with H0H_0 at the specified α\alpha level.

REASONING PATH
Your ACEpath teacher’s suggested approach
Strict Rejection Threshold
  • For an upper-tailed test at significance level α\alpha, reject H0H_0 if and only if tcalc≥tcritt_{\text{calc}} \ge t_{\text{crit}} (or p≤αp \le \alpha).

  • Never round or approximate a test statistic across the critical value boundary; preserve all precision when comparing tt to tcritt_{\text{crit}}.

CORRECTED PATH
ACEpath Teacher's Demonstration
  1. Compare the exact test statistic to the critical value: t=1.750t = 1.750 and tcrit=1.753t_{\text{crit}} = 1.753.

  2. Observe that 1.750<1.7531.750 < 1.753, which places tt outside the rejection region.

  3. Conclude that we fail to reject H0H_0 at α=0.05\alpha = 0.05.

Final answer

(a) t=1.750<1.753t = 1.750 < 1.753, fail to reject H0H_0; the reduction is not statistically significant. (b) A negative change signifies improvement under Δ=Xpost−Xpre\Delta = X_{\text{post}} - X_{\text{pre}}; the correct one-tailed pp-value is 1−0.95=0.051 - 0.95 = 0.05 (p≈0.0503p \approx 0.0503). (c) Because 2.80 mmHg<5.0 mmHg2.80\text{ mmHg} < 5.0\text{ mmHg}, the result lacks practical significance.

NEXT STEPS
Final Takeaways from Your Teacher
  • Always compare calculated test statistics to critical values using strict inequalities (t≥tcritt \ge t_{\text{crit}}) rather than rounding near the boundary.

  • Practice interpreting borderline pp-values and test statistics to reinforce that 'almost significant' is formally 'not statistically significant.'

OverviewACEpath Teacher's Diagnosis

03 The student conflates statistical significance with practical significance, believing that practical significance depends solely on a small p-value rather than the magnitude of the effect.

Sample Student Work

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

    H0:μ=0 vs. Ha:μ>0  ⟹  SE=6.4016=1.60  ⟹  t=2.801.60=1.75  ⟹  t=1.75<1.753, so not statistically significant.H_0: \mu = 0 \text{ vs. } H_a: \mu > 0 \implies SE = \frac{6.40}{\sqrt{16}} = 1.60 \implies t = \frac{2.80}{1.60} = 1.75 \implies t = 1.75 < 1.753\text{, so not statistically significant.}

  2. Step 2

    A negative difference under Xpost−XpreX_{\text{post}} - X_{\text{pre}} indicates blood pressure dropped, meaning the diet reduced blood pressure. The appropriate one-tailed pp-value is 0.050.05.

  3. Step 3

    Practical significance requires statistical significance first, so because the test was not significant at [crossed out: α=0.01\alpha = 0.01], it cannot be practically significant regardless of the 5.0 mmHg5.0\text{ mmHg} cutoff.

    Correct

    Compare the observed sample mean reduction directly to the clinical benchmark: xˉ=2.80 mmHg<5.0 mmHg\bar{x} = 2.80\text{ mmHg} < 5.0\text{ mmHg}, meaning the effect fails to reach the threshold for practical significance.

    You incorrectly stated that practical significance requires statistical significance as a prerequisite. Practical significance is an independent evaluation of whether the magnitude of an observed effect (ar{x} = 2.80\text{ mmHg}) is meaningful in a real-world clinical context relative to the benchmark (5.0 mmHg5.0\text{ mmHg}).

WHAT YOU DID WELL
Compliments from Your Teacher

You correctly formulated the one-sided null and alternative hypotheses.

You accurately calculated the standard error and test statistic.

You correctly compared the test statistic to the critical value to conclude that the result was not statistically significant.

You clearly identified how reversing the subtraction order inverted the sign of the effect in part (b).

THE ERROR
Where Your ACEpath Teacher Spotted the Mistake
Conflating statistical significance with practical significance

no

WHY IT HAPPENED
ACEpath Teacher's Diagnosis

You treated practical significance as a hierarchical downstream outcome of a hypothesis test rather than an independent domain-specific assessment of effect size.

REASONING PATH
Your ACEpath teacher’s suggested approach
Evaluate effect magnitude directly
  • Practical significance measures whether an effect size is large enough to matter in practice, regardless of sample size or hypothesis test outcomes.

  • Always compare the point estimate (or confidence interval) directly against the practical threshold value.

CORRECTED PATH
ACEpath Teacher's Demonstration
  1. Identify the practical/clinical threshold: Δpractical=5.0 mmHg\Delta_{\text{practical}} = 5.0\text{ mmHg}.

  2. Compare the observed sample reduction xˉ=2.80 mmHg\bar{x} = 2.80\text{ mmHg} to the clinical threshold: 2.80 mmHg<5.0 mmHg2.80\text{ mmHg} < 5.0\text{ mmHg}.

  3. State whether the observed reduction meets the standard of practical importance.

Final answer

Because the observed mean reduction of xˉ=2.80 mmHg\bar{x} = 2.80\text{ mmHg} is less than the required clinical threshold of 5.0 mmHg5.0\text{ mmHg}, the result is not practically significant.

NEXT STEPS
Final Takeaways from Your Teacher
  • Review the distinction between statistical significance (pp-values and sampling error) and practical significance (real-world effect size).

  • Practice using confidence intervals to report both statistical uncertainty and clinical relevance simultaneously.

Lock it in

Similar question

01

A researcher tests a new study technique on n=25n = 25 students. The mean test score gain is xˉ=4.5\bar{x} = 4.5 points with sample standard deviation s=7.5s = 7.5 points. Calculate the standard error of the mean and the test statistic tt for H0:μ=0H_0: \mu = 0.

Answer

SE=sn=7.525=7.55=1.5SE = \frac{s}{\sqrt{n}} = \frac{7.5}{\sqrt{25}} = \frac{7.5}{5} = 1.5. Then t=4.5−01.5=3.00t = \frac{4.5 - 0}{1.5} = 3.00.

02

In a matched-pairs experiment with n=9n = 9 subjects, the mean reduction in cholesterol is xˉ=6.0 mg/dL\bar{x} = 6.0\text{ mg/dL} with s=9.0 mg/dLs = 9.0\text{ mg/dL}. A clinical benchmark requires a reduction of at least 10.0 mg/dL10.0\text{ mg/dL} to be considered practically significant. (1) Find the test statistic tt for H0:μ=0H_0: \mu = 0 vs. Ha:μ>0H_a: \mu > 0. (2) Does the result meet practical significance?

Answer

(1) SE=9.09=3.0SE = \frac{9.0}{\sqrt{9}} = 3.0; t=6.0−03.0=2.00t = \frac{6.0 - 0}{3.0} = 2.00. (2) No; the observed sample mean of 6.0 mg/dL6.0\text{ mg/dL} is below the clinical threshold of 10.0 mg/dL10.0\text{ mg/dL}, so it lacks practical significance.

03

An analyst tests H0:μD=0H_0: \mu_D = 0 versus Ha:μD<0H_a: \mu_D < 0 where D=Post−PreD = \text{Post} - \text{Pre}. The sample size is n=64n = 64, dˉ=−1.80\bar{d} = -1.80, and sD=4.80s_D = 4.80. (1) Calculate the test statistic tt. (2) If a colleague mistakenly evaluated the right-tail probability and got p=0.9987p = 0.9987, what is the correct one-tailed pp-value?

Answer

(1) SE=4.8064=4.808=0.60SE = \frac{4.80}{\sqrt{64}} = \frac{4.80}{8} = 0.60; t=−1.80−00.60=−3.00t = \frac{-1.80 - 0}{0.60} = -3.00. (2) Because the test is left-tailed, the correct pp-value is P(T63≤−3.00)=1−0.9987=0.0013P(T_{63} \le -3.00) = 1 - 0.9987 = 0.0013.

Loading workspace…