Resources

An exercise physiologist measures the resting heart rates (in beats per minute, bpm\text{bpm}) of 2828 collegiate rowers before and after a 6-week altitude training program. The researcher calculates a sample mean difference of dˉ=−3.8 bpm\bar{d} = -3.8\text{ bpm} with standard deviation sd=2.4 bpms_d = 2.4\text{ bpm}, and drafts the hypotheses as H0:dˉ=0 bpmH_0: \bar{d} = 0\text{ bpm} versus Ha:dˉ<0 bpmH_a: \bar{d} < 0\text{ bpm}. (a) Identify the conceptual error in the researcher's notation. (b) Explain why a paired-difference parameter μd\mu_d must be tested rather than independent two-sample parameters μ1−μ2\mu_1 - \mu_2. (c) State the correct symbolic hypotheses using the parameter μd=μafter−μbefore\mu_d = \mu_{\text{after}} - \mu_{\text{before}}.

Answer and steps

Final answer

(a) Hypotheses must be stated in terms of unknown population parameters (such as μd\mu_d), never calculated sample statistics (such as dˉ\bar{d}), because sample statistics are directly observed and known without testing. (b) The pre-training and post-training observations come from the exact same individuals, creating inherently paired (dependent) bivariate data; defining μd\mu_d isolates within-subject changes and controls for large baseline differences between rowers. (c) H0:μd=0 bpmH_0: \mu_d = 0\text{ bpm} (or μd≥0 bpm\mu_d \ge 0\text{ bpm}) versus Ha:μd<0 bpmH_a: \mu_d < 0\text{ bpm}.

Steps

Part (a): Identifying the Error in Notation

In statistical hypothesis testing, the objective is to draw an inference about an unknown population parameter using evidence collected from a sample.

  • The sample mean difference, dˉ=−3.8 bpm\bar{d} = -3.8\text{ bpm}, is a sample statistic. It is an observed fact from the data, not an unknown quantity. We already know with certainty that dˉ=−3.8 bpm≠0 bpm\bar{d} = -3.8\text{ bpm} \neq 0\text{ bpm}, so stating H0:dˉ=0H_0: \bar{d} = 0 is logically invalid.
  • Hypotheses must always make claims about the true population parameter (here, the true mean difference in the population, denoted by μd\mu_d), never about sample statistics like dˉ\bar{d}, xˉ\bar{x}, or p^\hat{p}.

Part (b): Why Paired Design (μd\mu_d) is Required Instead of Two Independent Samples (μ1−μ2\mu_1 - \mu_2)

An independent two-sample test assumes that the observations in group 1 are completely unrelated to those in group 2. Here, that assumption is violated:

  1. Dependent / Repeated-Measures Structure: The "before" and "after" measurements are taken on the exact same 2828 rowers. Each rower serves as their own control.
  2. Eliminating Inter-Subject Variability: Different athletes have substantially different baseline resting heart rates due to genetics, fitness levels, and physiology. Treating the groups as independent would inflate the overall variance with these large between-subject differences, obscuring the effect of the training.
  3. Focusing on Within-Subject Changes: By defining a single difference score for each individual, di=afteri−beforeid_i = \text{after}_i - \text{before}_i, we convert the paired bivariate data into a single univariate sample of changes. The parameter μd\mu_d directly targets the mean within-subject change attributable to the altitude training program.

Part (c): Stating the Correct Hypotheses

Let μd=μafter−μbefore\mu_d = \mu_{\text{after}} - \mu_{\text{before}} represent the true mean difference in resting heart rate (in bpm\text{bpm}) for all collegiate rowers undergoing this altitude training program.

  • Null Hypothesis (H0H_0): The altitude training program produces no change in mean resting heart rate (or increases it): H0:μd=0 bpm(or H0:μd≥0 bpm)H_0: \mu_d = 0\text{ bpm} \quad (\text{or } H_0: \mu_d \ge 0\text{ bpm})
  • Alternative Hypothesis (HaH_a): The altitude training program reduces mean resting heart rate (a negative change): Ha:μd<0 bpmH_a: \mu_d < 0\text{ bpm}

Typical errors

OverviewACEpath Teacher's Diagnosis

01 The student believes the notation error is the direction of the alternative hypothesis rather than writing hypotheses in terms of a sample statistic.

Sample Student Work

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

    "The error is writing Ha:dˉ<0H_a: \bar{d} < 0 instead of a two-sided test Ha:dˉ≠0H_a: \bar{d} \neq 0."

    Correct

    The notation error is writing hypotheses in terms of the sample statistic dˉ\bar{d} instead of the unknown population parameter μd\mu_d.

    Hypotheses test claims about unknown population parameters, never about observed sample statistics. We already know with certainty that dˉ=−3.8 bpm\bar{d} = -3.8\text{ bpm}, so stating H0:dˉ=0H_0: \bar{d} = 0 is conceptually meaningless.

  2. Step 2

    "The data are paired because each rower is measured before and after, so the two measurements are dependent."

  3. Step 3

    H0:μd=0 bpm  ⟹  Ha:μd<0 bpmH_0: \mu_d = 0\text{ bpm} \implies H_a: \mu_d < 0\text{ bpm}

WHAT YOU DID WELL
Compliments from Your Teacher

You correctly identified that repeated measurements on the same subjects produce dependent, paired data.

You correctly formulated the symbolic hypotheses using the population parameter μd\mu_d in part (c).

THE ERROR
Where Your ACEpath Teacher Spotted the Mistake
Statistic vs. Parameter in Hypotheses

no

WHY IT HAPPENED
ACEpath Teacher's Diagnosis

Confusing observed sample statistics with unobserved population parameters when inspecting statistical hypotheses, leading to focusing on the test's directionality rather than parameter notation.

REASONING PATH
Your ACEpath teacher’s suggested approach
Parameters vs. Statistics in Hypotheses
  • Hypotheses are always statements about unobserved population parameters (e.g., μ\mu, pp, μd\mu_d), never calculated sample statistics (e.g., xˉ\bar{x}, p^\hat{p}, dˉ\bar{d}).

  • When inspecting hypothesis statements, verify that Greek letters representing population parameters are used rather than Latin letters with bars or hats representing sample statistics.

CORRECTED PATH
ACEpath Teacher's Demonstration
  1. Recall the purpose of hypothesis testing: to evaluate claims about an unknown population parameter using sample data as evidence.

  2. Examine the symbol used in H0:dˉ=0H_0: \bar{d} = 0: the bar notation dˉ\bar{d} represents the sample mean difference, which is already known from the sample data (dˉ=−3.8 bpm\bar{d} = -3.8\text{ bpm}).

  3. Conclude that the conceptual error is using the sample statistic dˉ\bar{d} instead of the true population parameter μd\mu_d.

Final answer

The conceptual error is using the sample statistic dˉ\bar{d} instead of the population parameter μd\mu_d in the hypotheses.

NEXT STEPS
Final Takeaways from Your Teacher
  • Review the distinct roles of population parameters (unknown targets of inference) and sample statistics (known evidence calculated from data).

  • Practice identifying whether given symbols (such as μ\mu vs. xˉ\bar{x}, pp vs. p^\hat{p}, μd\mu_d vs. dˉ\bar{d}) belong in hypothesis statements.

OverviewACEpath Teacher's Diagnosis

02 The student reverses the direction of the alternative hypothesis due to confusing reduction in heart rate with an increase in numerical value.

Sample Student Work

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

    The hypotheses are written using the sample statistic dˉ\bar{d} instead of the population parameter μd\mu_d.

  2. Step 2

    The observations come from the same rowers before and after training, so the samples are dependent.

  3. Step 3

    H0:μd=0 bpm  ⟹  Ha:μd>0 bpmH_0: \mu_d = 0\text{ bpm} \implies H_a: \mu_d > 0\text{ bpm}

    Correct

    H0:μd=0 bpm versus Ha:μd<0 bpmH_0: \mu_d = 0\text{ bpm} \text{ versus } H_a: \mu_d < 0\text{ bpm}

    Because the difference is defined as μd=μafter−μbefore\mu_d = \mu_{\text{after}} - \mu_{\text{before}}, a reduction in resting heart rate means the post-training values are lower than baseline, which makes μd<0\mu_d < 0.

WHAT YOU DID WELL
Compliments from Your Teacher

You clearly distinguished between a sample statistic and a population parameter in hypothesis testing notation.

You correctly identified the dependent, paired nature of repeated measurements on the same individuals.

THE ERROR
Where Your ACEpath Teacher Spotted the Mistake
Reversed Inequality Direction in Paired Difference

You set the alternative hypothesis to strictly greater than zero instead of strictly less than zero.

WHY IT HAPPENED
ACEpath Teacher's Diagnosis

You likely associated an improvement in fitness with a positive sign (>0> 0), overlooking that heart rate decreases after training, which yields a negative difference when subtracting baseline from post-training values.

REASONING PATH
Your ACEpath teacher’s suggested approach
Define the Subtraction Order
  • For μd=μafter−μbefore\mu_d = \mu_{\text{after}} - \mu_{\text{before}}, a decrease corresponds to μafter<μbefore\mu_{\text{after}} < \mu_{\text{before}}, so μd<0\mu_d < 0.

  • Always substitute sample values into the difference formula (extafter−extbefore ext{after} - ext{before}) to confirm whether a desired effect produces a positive or negative difference.

CORRECTED PATH
ACEpath Teacher's Demonstration
  1. Identify the research goal: determining if altitude training reduces resting heart rate.

  2. Apply the subtraction definition: μafter<μbefore  ⟹  μafter−μbefore<0\mu_{\text{after}} < \mu_{\text{before}} \implies \mu_{\text{after}} - \mu_{\text{before}} < 0.

  3. Express the research claim in the alternative hypothesis: Ha:μd<0 bpmH_a: \mu_d < 0\text{ bpm}.

Final answer

H0:μd=0 bpm (or μd≥0 bpm) versus Ha:μd<0 bpmH_0: \mu_d = 0\text{ bpm} \text{ (or } \mu_d \ge 0\text{ bpm}) \text{ versus } H_a: \mu_d < 0\text{ bpm}

NEXT STEPS
Final Takeaways from Your Teacher
  • Write out the subtraction order explicitly before setting up the inequality sign.

  • Check that the sign of your sample statistic dˉ=−3.8 bpm\bar{d} = -3.8\text{ bpm} aligns with the direction of your alternative hypothesis.

OverviewACEpath Teacher's Diagnosis

03 The student misinterprets paired data by claiming the two groups have equal sample sizes rather than recognizing the measurements come from the same subjects.

Sample Student Work

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

    Hypotheses must use the population parameter μd\mu_d rather than the sample mean μd\mu_d.

  2. Step 2

    A paired difference is used because both groups have an equal sample size of n=28n = 28.

    Correct

    A paired difference μd\mu_d is required because the before and after measurements are taken on the exact same 2828 rowers (repeated measures / dependent data), which accounts for baseline differences between rowers.

    Equal sample sizes (n1=n2=28n_1 = n_2 = 28) can occur in completely independent two-sample designs. Data is paired because each 'after' observation is paired with the 'before' observation from the exact same individual rower.

  3. Step 3

    H0:μd=0 bpm  ⟹  Ha:μd<0 bpmH_0: \mu_d = 0\text{ bpm} \implies H_a: \mu_d < 0\text{ bpm}

WHAT YOU DID WELL
Compliments from Your Teacher

You correctly recognized that statistical hypotheses test unknown population parameters rather than observed sample statistics.

You accurately formulated the symbolic hypotheses for the paired design in part (c).

THE ERROR
Where Your ACEpath Teacher Spotted the Mistake
Paired vs. Independent Data Misconception
WHY IT HAPPENED
ACEpath Teacher's Diagnosis

You confused equal sample sizes with data dependence. Two independent groups can have identical sample sizes, whereas a paired design specifically requires a one-to-one correspondence between measurements (such as repeated measurements on the same experimental units or matched pairs).

REASONING PATH
Your ACEpath teacher’s suggested approach
Identifying Paired vs. Independent Designs
  • A study is paired if each measurement in one condition is naturally linked to exactly one measurement in the other condition (most commonly, before/after on the same subject).

  • Look for repeated measures on the same subjects or natural matching pairs, not just whether n1=n2n_1 = n_2.

CORRECTED PATH
ACEpath Teacher's Demonstration
  1. Check the study design: were resting heart rates measured on two separate groups of 28 rowers, or on the same 28 rowers twice?

  2. Because the measurements are before and after on the exact same rowers, each rower acts as their own control.

  3. Therefore, the data are inherently dependent (paired), requiring the within-subject difference parameter μd\mu_d.

Final answer

(b) The measurements come from the same 28 rowers before and after training (repeated measures on the same subjects), making the data dependent and requiring a paired parameter μd\mu_d rather than independent two-sample parameters.

NEXT STEPS
Final Takeaways from Your Teacher
  • Review the structural difference between two independent samples and paired (matched/repeated measures) samples.

  • Practice identifying paired designs by asking: 'Does observation 1 in group A have a direct, meaningful link to observation 1 in group B?'

Lock it in

Similar question

01

A quality control engineer inspects 5050 manufactured bolts and calculates a sample mean diameter of xˉ=10.02 mm\bar{x} = 10.02\text{ mm}. The engineer sets up the test as H0:xˉ=10.00 mmH_0: \bar{x} = 10.00\text{ mm} versus Ha:xˉ≠10.00 mmH_a: \bar{x} \neq 10.00\text{ mm}. Explain why this hypothesis setup is conceptually flawed and write the corrected hypotheses.

Answer

The setup is flawed because hypotheses make claims about the unknown population mean μ\mu, not the observed sample mean xˉ\bar{x}, which is already known to be 10.02 mm10.02\text{ mm}. The correct hypotheses are H0:μ=10.00 mmH_0: \mu = 10.00\text{ mm} versus Ha:μ≠10.00 mmH_a: \mu \neq 10.00\text{ mm}.

02

A clinical researcher tests whether a new sleep aid increases sleep duration by recording nightly sleep for 2020 insomniac patients before starting medication and after two weeks of treatment. The researcher defines d=hours after−hours befored = \text{hours after} - \text{hours before} and reports dˉ=+1.2 hours\bar{d} = +1.2\text{ hours}. State the appropriate parameter and write the formal null and alternative hypotheses to test if the medication increases sleep.

Answer

Let μd\mu_d be the true mean difference in sleep duration (after minus before) in the population of insomniacs. Because the researcher wants to test for an increase, the hypotheses are H0:μd=0 hoursH_0: \mu_d = 0\text{ hours} (or μd≤0 hours\mu_d \le 0\text{ hours}) versus Ha:μd>0 hoursH_a: \mu_d > 0\text{ hours}.

03

A researcher conducts a study comparing test scores between a tutoring group (n1=30n_1 = 30) and an independent control group (n2=30n_2 = 30). A junior analyst writes the test as H0:μd=0H_0: \mu_d = 0. Explain why using μd\mu_d is an error here, identify the correct parameter form, and write the appropriate hypotheses to test if the tutoring group scores higher.

Answer

Using μd\mu_d is incorrect because the two groups consist of separate, independent individuals rather than paired or repeated measurements on the same subjects. For independent samples, the parameter is the difference between two population means, μ1−μ2\mu_1 - \mu_2. The correct hypotheses are H0:μ1−μ2=0H_0: \mu_1 - \mu_2 = 0 (or μ1−μ2≤0\mu_1 - \mu_2 \le 0) versus Ha:μ1−μ2>0H_a: \mu_1 - \mu_2 > 0.

Loading workspace…