Resources

A clinical research report evaluates an experimental allergy medication and states the following result: 'Using a two-tailed one-proportion zz-test with a calculated test statistic of z=−2.48z = -2.48 and a PP-value of 0.01320.0132, we reject the null hypothesis at the α=0.05\alpha = 0.05 significance level and conclude that the true proportion of patients reporting daytime drowsiness significantly differs from the historical benchmark of 12%12\%.' Reverse-engineer the study's formal statistical setup by: (a) defining the population parameter pp, (b) stating the exact symbolic null and alternative hypotheses that yielded this test, and (c) determining whether the observed sample proportion p^\hat{p} was strictly greater than or strictly less than 0.120.12.

Answer and steps

Final answer

(a) pp is the true population proportion of all patients taking the experimental medication who experience daytime drowsiness. (b) H0:p=0.12H_0: p = 0.12 versus Ha:p≠0.12H_a: p \ne 0.12. (c) Because the test statistic z=p^−p0p0(1−p0)/n=−2.48z = \frac{\hat{p} - p_0}{\sqrt{p_0(1-p_0)/n}} = -2.48 is negative, the observed sample proportion p^\hat{p} was strictly less than 0.120.12.

Steps

Part (a): Defining the Population Parameter

In statistical inference for a single proportion, the parameter represents the true, unknown proportion in the target population rather than the sample statistic.

Based on the study context, the benchmark refers to the rate of daytime drowsiness under the new treatment:

  • Let pp be the true population proportion of all patients taking the experimental allergy medication who experience daytime drowsiness.

Common Pitfall: Do not define pp as the proportion of patients in the sample who experienced drowsiness. That quantity is the sample statistic p^\hat{p}, which is known and observed, not the unknown population parameter pp.


Part (b): Formulating the Null and Alternative Hypotheses

  1. Null Hypothesis (H0H_0): The null hypothesis represents the claim of "no difference" or conformity to the established benchmark. The historical benchmark given is 12%12\%, or 0.120.12: H0:p=0.12H_0: p = 0.12
  2. Alternative Hypothesis (HaH_a): The report states that researchers conducted a two-tailed test to determine if the rate significantly differs from 12%12\%. A two-tailed test evaluates differences in either direction (greater than or less than): Ha:p≠0.12H_a: p \ne 0.12

Part (c): Comparing p^\hat{p} to the Hypothesized Value

The test statistic for a one-proportion zz-test is calculated using: z=p^−p0p0(1−p0)nz = \frac{\hat{p} - p_0}{\sqrt{\frac{p_0(1 - p_0)}{n}}}

where:

  • p0=0.12p_0 = 0.12 is the null value,
  • p^\hat{p} is the sample proportion,
  • nn is the sample size, and
  • the standard error in the denominator, p0(1−p0)n\sqrt{\frac{p_0(1 - p_0)}{n}}, is always strictly positive.

Since the calculated test statistic is negative: z=−2.48<0z = -2.48 < 0

it follows that the numerator must also be negative: p^−0.12<0  ⟹  p^<0.12\hat{p} - 0.12 < 0 \implies \hat{p} < 0.12

Therefore, the observed sample proportion p^\hat{p} was strictly less than 0.120.12.

Typical errors

OverviewACEpath Teacher's Diagnosis

01 A student mistakenly sets up a one-tailed alternative hypothesis based on the negative sign of the calculated test statistic instead of adhering to the report's two-tailed test description.

Sample Student Work

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

    Let p=true proportion of patients on the medication who report daytime drowsiness.p = \text{true proportion of patients on the medication who report daytime drowsiness.}

  2. Step 2

    H0:p=0.12H_0: p = 0.12 \newline Ha:p<0.12H_a: p < 0.12

    Correct

    H0:p=0.12H_0: p = 0.12 \newline Ha:p≠0.12H_a: p \ne 0.12

    You correctly identified the null hypothesis, but you wrote a one-sided alternative hypothesis (p<0.12p < 0.12). Because the report explicitly stated that researchers conducted a "two-tailed" test to determine if the rate "significantly differs," the alternative hypothesis must be two-sided: Ha:p≠0.12H_a: p \ne 0.12.

  3. Step 3

    z=p^−0.12SE=−2.48  ⟹  p^−0.12<0  ⟹  p^ is strictly less than 0.12z = \frac{\hat{p} - 0.12}{\text{SE}} = -2.48 \implies \hat{p} - 0.12 < 0 \implies \hat{p} \text{ is strictly less than } 0.12

WHAT YOU DID WELL
Compliments from Your Teacher

You correctly identified and defined the population parameter pp in context rather than confusing it with a sample statistic.

You set the null hypothesis benchmark value accurately to 0.120.12.

You correctly related the sign of the test statistic to the direction of the deviation of p^\hat{p} from the null value.

THE ERROR
Where Your ACEpath Teacher Spotted the Mistake
One-Tailed vs. Two-Tailed Hypothesis Formulation

You let the post-data negative test statistic (z=−2.48z = -2.48) dictate the direction of your alternative hypothesis instead of following the research design specified prior to collecting data.

WHY IT HAPPENED
ACEpath Teacher's Diagnosis

Hypotheses must represent the research question established before data collection rather than being modified post-hoc by the observed sample result. Keywords such as "two-tailed" and "significantly differs" always correspond to a two-sided inequality (p≠p0p \ne p_0), regardless of whether the calculated zz-value turns out positive or negative.

REASONING PATH
Your ACEpath teacher’s suggested approach
Research Question Directionality
  • An alternative hypothesis reflects the planned inquiry: "differs from" or "two-tailed" requires ≠\ne, while directional claims like "greater than" or "less than" require >> or <<.

  • Look for research design keywords in the prompt before looking at any sample statistics or zz-scores.

Hypotheses Are A Priori
  • Hypotheses are stated before observing sample data, so an observed negative sample result or test statistic cannot turn a two-tailed test into a one-tailed alternative hypothesis.

  • Never use the sign of zz or p^\hat{p} to choose the inequality symbol in HaH_a when the study protocol specifies a two-tailed test.

CORRECTED PATH
ACEpath Teacher's Demonstration
  1. Identify the study design from the text: "Using a two-tailed one-proportion zz-test... conclude that the true proportion... significantly differs from the historical benchmark of 12%12\%."

  2. Express "differs from 0.120.12" symbolically as p≠0.12p \ne 0.12.

  3. Combine with the baseline benchmark: H0:p=0.12H_0: p = 0.12 versus Ha:p≠0.12H_a: p \ne 0.12.

Final answer

(a) pp is the true population proportion of all patients taking the experimental medication who experience daytime drowsiness. (b) H0:p=0.12H_0: p = 0.12 versus Ha:p≠0.12H_a: p \ne 0.12. (c) Because the test statistic z=−2.48<0z = -2.48 < 0, the observed sample proportion p^\hat{p} was strictly less than 0.120.12.

NEXT STEPS
Final Takeaways from Your Teacher
  • Practice picking out key design phrasing ("differs from", "is different", "two-tailed") vs. directional phrasing ("decreased", "improved", "exceeds").

  • Remember that sample data (zz, xˉ\bar{x}, p^\hat{p}) evaluate the hypotheses, but they never determine the direction of the hypotheses themselves.

OverviewACEpath Teacher's Diagnosis

02 A student defines the parameter in terms of the sample statistic rather than the unknown population proportion.

Sample Student Work

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

    Let p=proportion of patients in the study sample who report daytime drowsiness.p = \text{proportion of patients in the study sample who report daytime drowsiness.}

    Correct

    Let pp be the true population proportion of all patients taking the experimental medication who experience daytime drowsiness.

    You defined pp as the proportion of patients in the study sample, which is the sample statistic p^\hat{p}. Statistical hypotheses and population parameters always describe the entire target population, not the observed sample.

  2. Step 2

    H0:p=0.12  ⟹  Ha:p≠0.12H_0: p = 0.12 \implies H_a: p \ne 0.12

  3. Step 3

    z=p^−0.12SE=−2.48<0  ⟹  p^<0.12  ⟹  p^ is strictly less than 0.12z = \frac{\hat{p} - 0.12}{SE} = -2.48 < 0 \implies \hat{p} < 0.12 \implies \hat{p} \text{ is strictly less than } 0.12

WHAT YOU DID WELL
Compliments from Your Teacher

You correctly stated both the null and two-tailed alternative hypotheses using proper parameter notation.

You connected the sign of the test statistic to the difference between the sample statistic and the null value.

THE ERROR
Where Your ACEpath Teacher Spotted the Mistake
Parameter Defined as Sample Statistic

no

WHY IT HAPPENED
ACEpath Teacher's Diagnosis

Confusing the unknown population parameter pp (which hypotheses are written about) with the observed sample statistic p^\hat{p} (which is computed directly from data).

REASONING PATH
Your ACEpath teacher’s suggested approach
Parameter vs. Statistic Scope
  • Parameters describe the entire underlying population or general process; statistics describe only the collected sample.

  • Always use words like 'true population proportion of all...' when defining pp, and reserve 'in the sample' exclusively for p^\hat{p}.

CORRECTED PATH
ACEpath Teacher's Demonstration
  1. Identify the target group that the researchers wish to draw conclusions about: all current and future patients who might take this experimental medication.

  2. Define pp as: 'p=the true population proportion of all patients taking the experimental medication who report daytime drowsiness.p = \text{the true population proportion of all patients taking the experimental medication who report daytime drowsiness.}'

Final answer

(a) pp is the true population proportion of all patients taking the experimental allergy medication who experience daytime drowsiness. (b) H0:p=0.12H_0: p = 0.12 vs. Ha:p≠0.12H_a: p \ne 0.12. (c) p^\hat{p} is strictly less than 0.120.12.

NEXT STEPS
Final Takeaways from Your Teacher
  • Practice writing formal parameter definitions by explicitly including the words 'true population proportion' or 'all individuals in the population'.

  • Review why hypothesis tests never make statements about p^\hat{p}, since sample statistics are already known with certainty once data is gathered.

OverviewACEpath Teacher's Diagnosis

03 A student confuses the formula for the z-test statistic by inverting the order of subtraction as (p₀ - p̂), concluding that p̂ was strictly greater than 0.12.

Sample Student Work

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

    pp is the true proportion of patients taking the medication who report daytime drowsiness.

  2. Step 2

    H0:p=0.12  ⟹  Ha:p≠0.12H_0: p = 0.12 \implies H_a: p \ne 0.12

  3. Step 3

    z=0.12−p^SE=−2.48z = \frac{0.12 - \hat{p}}{\text{SE}} = -2.48

    Correct

    z=p^−0.12SE=−2.48z = \frac{\hat{p} - 0.12}{\text{SE}} = -2.48

    You inverted the order of subtraction in the zz-test formula. A standardized test statistic is always calculated as observed statistic−null valuestandard error\frac{\text{observed statistic} - \text{null value}}{\text{standard error}}, which gives p^−0.12SE\frac{\hat{p} - 0.12}{\text{SE}} rather than 0.12−p^SE\frac{0.12 - \hat{p}}{\text{SE}}.

  4. Step 4

    0.12−p^<0  ⟹  p^0.12 - \hat{p} < 0 \implies \hat{p} is strictly greater than 0.120.12

    Correct

    p^−0.12<0  ⟹  p^\hat{p} - 0.12 < 0 \implies \hat{p} is strictly less than 0.120.12

    Your inequality reasoning followed logically from your formula in Step 3, but because the numerator was inverted, it led to the opposite conclusion.

WHAT YOU DID WELL
Compliments from Your Teacher

You correctly identified that the population parameter pp represents the true proportion in the entire target population rather than the sample.

You accurately formulated the two-tailed null and alternative hypotheses using the historical rate of 0.120.12.

You correctly recognized that the sign of the zz-score is determined entirely by the sign of its numerator because the standard error is strictly positive.

THE ERROR
Where Your ACEpath Teacher Spotted the Mistake
Inverted Test Statistic Numerator

Inverting the order of subtraction in the numerator of the standardized zz-statistic as (p0−p^)(p_0 - \hat{p}) instead of (p^−p0)(\hat{p} - p_0).

WHY IT HAPPENED
ACEpath Teacher's Diagnosis

Conflating the baseline/null value with the observed sample statistic when building the standardized distance from the null. Remembering that test statistics measure 'how far the observed sample lies from the hypothesized center' establishes the natural subtraction order: observation−center\text{observation} - \text{center}.

REASONING PATH
Your ACEpath teacher’s suggested approach
Standardized Score Template
  • Every standard score measures distance from the reference center: z=observed−hypothesizedSEz = \frac{\text{observed} - \text{hypothesized}}{\text{SE}}.

  • Write down the general word formula z=statistic−parameterSEz = \frac{\text{statistic} - \text{parameter}}{\text{SE}} before substituting numbers.

Sign Check on Test Statistics
  • A negative zz-statistic (z<0z < 0) indicates that the observed sample fell below the hypothesized benchmark, while a positive zz-statistic (z>0z > 0) indicates the sample fell above it.

  • Check that the sign of zz matches the physical direction: here, z=−2.48<0z = -2.48 < 0 requires p^<p0\hat{p} < p_0.

CORRECTED PATH
ACEpath Teacher's Demonstration
  1. Recall the formula for a one-proportion zz-test: z=p^−p0SEz = \frac{\hat{p} - p_0}{\text{SE}} where SE=p0(1−p0)n>0\text{SE} = \sqrt{\frac{p_0(1 - p_0)}{n}} > 0.

  2. Substitute the given values: p^−0.12SE=−2.48\frac{\hat{p} - 0.12}{\text{SE}} = -2.48.

  3. Because the denominator SE\text{SE} is strictly positive, the numerator must share the negative sign of zz: p^−0.12<0\hat{p} - 0.12 < 0.

  4. Add 0.120.12 to both sides to obtain p^<0.12\hat{p} < 0.12.

Final answer

The observed sample proportion p^\hat{p} was strictly less than 0.120.12.

NEXT STEPS
Final Takeaways from Your Teacher
  • Memorize the anchor structure: z=observed−expectedSEz = \frac{\text{observed} - \text{expected}}{\text{SE}}, so the sample always comes first.

  • Whenever you calculate or interpret a negative test statistic, immediately verify that the sample result is smaller than the hypothesized benchmark.

Lock it in

Similar question

01

A quality control manager inspects ball bearings to see if their mean diameter differs from the standard specification of 5.00 mm5.00\text{ mm}. A sample produces a test statistic of t=−3.12t = -3.12. State the null and alternative hypotheses for this test.

Answer

H0:μ=5.00H_0: \mu = 5.00 versus Ha:μ≠5.00H_a: \mu \ne 5.00. Even though the sample mean was smaller than 5.005.00 (t<0t < 0), the manager's inquiry was whether the diameter simply "differs from" the specification, which requires a two-sided alternative.

02

An educational researcher conducts a study reported as: "A one-sample zz-test for a proportion yielded z=+1.98z = +1.98 with a two-sided PP-value of 0.04770.0477, showing that the percentage of students passing the exam differs significantly from the state baseline of 70%70\%." (a) State H0H_0 and HaH_a. (b) Was the sample pass rate p^\hat{p} greater than or less than 0.700.70?

Answer

(a) H0:p=0.70H_0: p = 0.70 versus Ha:p≠0.70H_a: p \ne 0.70, because the test is explicitly two-sided ("differs significantly"). (b) Because z=p^−0.70SE=+1.98>0z = \frac{\hat{p} - 0.70}{\text{SE}} = +1.98 > 0, the sample proportion p^\hat{p} was strictly greater than 0.700.70.

03

A marketing analyst wants to know if a website redesign changed the historical 8%8\% conversion rate. After collecting data, they find p^=0.065\hat{p} = 0.065 and z=−1.85z = -1.85. A colleague writes Ha:p<0.08H_a: p < 0.08. Explain why the colleague's alternative hypothesis is incorrect based on the research goal, and provide the correct pair of hypotheses.

Answer

The research question asked if the redesign "changed" the conversion rate (a non-directional, two-sided goal formulated before observing the data). Basing the inequality on the observed drop to 6.5%6.5\% is a post-hoc error. The correct setup is H0:p=0.08H_0: p = 0.08 versus Ha:p≠0.08H_a: p \ne 0.08.

Loading workspace…