Resources
A machine calibration test evaluates against a left-tailed alternative . A student calculates a sample mean with standard error , yielding . The student looks up in a cumulative normal table, finds the entry , and reports . (a) Explain the directional error in the student's computation of the -value for a left-tailed test. (b) Calculate the correct -value for when . (c) State the correct decision at using standard statistical phrasing.
Answer and steps
Final answer
(a) For a left-tailed test, the -value is the probability of observing a test statistic as small or smaller than observed, ; the student incorrectly reported the cumulative probability while testing for a decrease, ignoring that the sample mean was in the direction opposite to ; (b) Correct -value: ; (c) Since , fail to reject .
Steps
Part (a): Understanding the Directional Error
A -value is the probability, assuming the null hypothesis is true, of observing a test statistic at least as favorable to the alternative hypothesis as the one actually obtained.
-
Direction of the Test:
- The alternative hypothesis is left-tailed: .
- Therefore, evidence against and in favor of occurs when sample values fall far to the left of (negative -scores). The definition of the -value for any left-tailed test is:
-
The Observed Result:
- The observed sample mean is , which gives:
- Because , the sample result lies standard errors in the direction opposite to the alternative hypothesis.
-
The Conceptual Pitfall:
- A common mistake in hypothesis testing is either using an upper tail by default or blindly copying a cumulative table value without checking the direction of the claim. In this scenario, while the cumulative probability numerically matches , reporting a -value near requires recognizing that the data provides virtually no evidence for a decrease (). The student mechanically took the table entry without realizing that the sample fell entirely in the wrong tail for the stated claim.
Part (b): Calculating the Correct -Value
By definition of a left-tailed test:
Substitute :
Part (c): Statistical Decision at
- Comparison: Compare the -value to the significance level :
- Conclusion: Since the -value is greater than , we fail to reject . There is insufficient evidence to conclude that the mean calibration measurement is less than .
Typical errors
01 The student mistakenly believes that a P-value must always be the area in the upper tail or 1 minus the cumulative table probability.
Sample Student Work
Reviewed by ACEpath TeacherStep-by-Step Review
Walk Through It with Your Teacher- Step 1
The student took the cumulative area directly instead of subtracting it from 1. For any test, the P-value is the tail area beyond z, which is 1 - 0.9772.
CorrectFor a left-tailed test (), the -value is the area to the left of the observed test statistic: . Since () lies in the direction opposite to , the test statistic provides virtually no evidence for , and the cumulative probability should not be subtracted from .
You assumed that a -value is always the area in the smaller outer tail. In a left-tailed test, the -value measures the probability of observing a value as small or smaller than observed, which is the left tail , not the right tail.
- Step 2Correct
You subtracted the cumulative probability from . For a left-tailed alternative , the -value is directly the cumulative left-tail area .
- Step 3
Since , reject Ho. There is sufficient evidence that .
CorrectSince , fail to reject . There is insufficient evidence to conclude that the mean calibration measurement is less than .
Using the correct -value of , which is much greater than , we fail to reject . A sample mean of cannot support the claim that the true mean is less than .
WHAT YOU DID WELL
Compliments from Your TeacherYou correctly recalled that a -value smaller than leads to rejecting the null hypothesis.
THE ERROR
Where Your ACEpath Teacher Spotted the Mistakeno
WHY IT HAPPENED
ACEpath Teacher's DiagnosisYou believed that a -value must always represent a small tail area beyond (or ), overlooking that the alternative hypothesis dictates the tail direction regardless of where the sample mean falls.
REASONING PATH
Your ACEpath teacher’s suggested approach- Tail Direction Rule
For a left-tailed test (), the -value is . For a right-tailed test (), the -value is .
Always check the inequality sign in to set the integration/tail direction before calculating the -value.
- Sanity Check for Observed Direction
If the sample statistic deviates in the direction opposite to , the -value must be greater than .
Verify whether the sample statistic supports ; an observed mean greater than the null value cannot provide evidence that the true mean is smaller.
CORRECTED PATH
ACEpath Teacher's DemonstrationIdentify the alternative hypothesis direction: (left-tailed).
Set up the -value probability expression: .
Evaluate using the standard normal cumulative distribution: .
Compare with significance level: fail to reject .
(a) The -value for a left-tailed test is the cumulative left area ; subtracting from wrongly shifts the test to an upper-tail test. (b) . (c) Fail to reject because .
02 The student confuses a left-tailed test with always using the negative z-score tail, calculating P(Z ≤ -2.00).
Sample Student Work
Reviewed by ACEpath TeacherStep-by-Step Review
Walk Through It with Your Teacher- Step 1
For a left-tailed test, the tail must be on the left side, so the student should have looked up instead of .
CorrectFor a left-tailed test with , the -value is defined as . Since gives an observed statistic , the sample lies in the direction opposite to , meaning almost the entire distribution lies to its left.
You cannot change the sign of the observed test statistic. The observed data produced , so the -value must evaluate the probability to the left of , not .
- Step 2Correct
You calculated the tail area using . Evaluating the tail in the direction of from the actual observed statistic gives .
- Step 3
Since , reject . There is sufficient evidence that the mean is less than .
CorrectSince , fail to reject . There is insufficient evidence to conclude that the mean calibration measurement is less than .
Your decision logic followed correctly from your calculated -value of , but because the correct -value is , you must fail to reject .
WHAT YOU DID WELL
Compliments from Your TeacherYou correctly identified that a left-tailed test accumulates probability to the left ().
You correctly looked up the cumulative normal probability for as .
You applied the standard decision rule of comparing your -value to with appropriate statistical language.
THE ERROR
Where Your ACEpath Teacher Spotted the Mistakeno
WHY IT HAPPENED
ACEpath Teacher's DiagnosisYou assumed that because the test is left-tailed, the test statistic used to compute the -value must be placed in the negative tail, confusing the direction of the alternative hypothesis with the location of the observed data.
REASONING PATH
Your ACEpath teacher’s suggested approach- Preserve Observed Values
The test statistic is fixed by the sample data; never negate or alter its sign to match the hypothesized direction.
Keep exactly as calculated from the sample mean .
- Definition of Left-Tailed P-Value
For , the -value is always , regardless of whether is positive or negative.
Shade everything to the left of the observed value to get .
CORRECTED PATH
ACEpath Teacher's DemonstrationIdentify the observed test statistic directly from the data: .
Apply the left-tailed definition: .
Find the cumulative table probability: .
Compare with : since , fail to reject .
(a) The student in the prompt actually reported the correct cumulative probability , but conceptually failed to realize that observing a positive -score provides zero evidence in favor of a decrease; (b) ; (c) Since , fail to reject .
03 The student treats the test as two-tailed by automatically doubling the tail probability.
Sample Student Work
Reviewed by ACEpath TeacherStep-by-Step Review
Walk Through It with Your Teacher- Step 1
The student forgot to double the tail area for the two-tailed test.
CorrectThe test is left-tailed (), not two-tailed. The -value is the probability of obtaining a test statistic as extreme or more extreme in the direction of , which is .
You misidentified the hypothesis test as two-tailed. The problem statement explicitly defines , which is strictly a left-tailed test.
- Step 2Correct
Because you assumed a two-tailed test, you computed the upper-tail probability and doubled it. For a left-tailed test, the -value is the area to the left of the observed test statistic: .
- Step 3
Since , reject . There is sufficient evidence that the mean is less than .
CorrectSince , fail to reject . There is insufficient evidence to conclude that the mean calibration measurement is less than .
Using your calculated -value led to an incorrect decision. Comparing the true -value to shows , so we fail to reject . Notice that an observed sample mean of () cannot provide evidence that the true mean is less than .
WHAT YOU DID WELL
Compliments from Your TeacherYou correctly calculated the upper-tail probability .
You followed the correct statistical decision logic of rejecting the null hypothesis whenever .
THE ERROR
Where Your ACEpath Teacher Spotted the MistakeYou overlooked the inequality sign in and defaulted to a two-tailed calculation.
WHY IT HAPPENED
ACEpath Teacher's DiagnosisYou assumed that tail areas must always be doubled, conflating two-tailed test procedures with directional one-tailed tests. In addition, you did not verify whether the sample evidence () physically supported the claimed decrease before deciding to reject .
REASONING PATH
Your ACEpath teacher’s suggested approach- Check the Alternative Hypothesis Direction
If contains '', it is left-tailed (). If contains '', it is right-tailed (). Only double the tail when contains ''.
Always circle the inequality symbol in before looking up or calculating tail areas.
- Reality-Check the Sample Direction
A sample mean that is greater than () provides no evidence in favor of .
Verify that your observed statistic lies in the direction claimed by . If it lies in the opposite direction, the one-tailed -value will exceed .
CORRECTED PATH
ACEpath Teacher's DemonstrationIdentify the direction of the test from : the test is left-tailed, so the -value is the area to the left of the observed test statistic.
Calculate the -value using the standard normal distribution: .
Compare the -value to : since , do not reject .
(a) The test is left-tailed (), so the -value is ; the observed sample mean was higher than , lying in the direction opposite to . (b) . (c) Since , fail to reject .
Lock it in
How do you determine the tail direction of a -value from the hypotheses?
Look at : is left-tailed (), is right-tailed (), and is two-tailed ().
Similar question
A test of versus yields a test statistic . Find the -value and state the decision at .
Answer
. Since , fail to reject .
A quality inspector tests against . A sample produces with standard error , giving . Compute the -value and state whether this sample provides evidence for .
Answer
. The sample provides no evidence for since is below and .
An experiment tests against . An analyst mistakenly calculates by finding the upper tail of . Explain why this calculation is contradictory and provide the correct -value.
Answer
The test is left-tailed, so . Observing means the sample mean is positive, which directly contradicts a claim that . The correct -value is .