Two-Sample T Test and Paired T Test
A decision-and-workflow guide for two-sample and paired t tests, covering method selection, conditions, mathematics, calculator evidence, and contextual reporting.
Method at a Glance: Two Sample t-Test
Two-sample t inference analyzes independent groups with a Welch standard error, while paired t inference analyzes one sample of within-pair differences.
Procedure Workflow
- Identify the data structure and parameter before selecting two sample t test; the name of a calculator menu is not method evidence.
- State the hypotheses or estimation target for two sample t test using population notation and the order defined by the question.
- Verify the design, independence, and approximation conditions that specifically justify two sample t test rather than reciting every condition learned in the course.
- Compute the statistic, standard error, interval, or p-value for two sample t test with defined symbols, guard digits, and an independent arithmetic check.
- Interpret two sample t test in the population and units named by the problem, then limit causation and generalization to what the collection design supports.
Procedure Formulas and Notation
Two-sample t statistic
Two-sample t statistic in Two Sample t-Test: Keep the population mean, sample mean, sample standard deviation, and standard error distinct, including their original measurement units.
Matched-pairs t statistic
Matched-pairs t statistic in Two Sample t-Test: This expression belongs specifically to two-sample and paired t tests; define every symbol and apply the scope rule for independent groups, difference scores, Welch inference, and design selection before calculation.
Independent two-sample t test
Decision
For Independent two-sample t test in two sample t test, A constructed sample from an online-course completion sample gives , , and . Test two-sided.
Independent two-sample t test result in two sample t test: The statistic is t=1.429 with df=24; p=0.1660.
Interpretation and validity
Independent two-sample t test interpretation for two sample t test: The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation.
Condition evidence for two sample t test and Independent two-sample t test: Use a random sample or randomized design, independence, and a distribution without problematic skew or outliers for the sample size.
Conditions
Decision
For Conditions in two sample t test, A constructed paired study for a city bus arrival investigation gives , , and . Test two-sided.
Conditions result in two sample t test: The paired statistic is t=2.984 with df=20; p=0.0073.
Interpretation and validity
Conditions interpretation for two sample t test: The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation.
Condition evidence for two sample t test and Conditions: Differences must be independent across pairs and approximately normal or supported by adequate sample size without severe outliers.
Welch test
Decision
For Welch test in two sample t test, A constructed sample from a website response-time study gives , , and . Test two-sided.
Welch test result in two sample t test: The statistic is t=2.705 with df=26; p=0.0119.
Interpretation and validity
Welch test interpretation for two sample t test: The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation.
Condition evidence for two sample t test and Welch test: Use a random sample or randomized design, independence, and a distribution without problematic skew or outliers for the sample size.
Paired t test
Decision
For Paired t test in two sample t test, A constructed paired study for a school library checkout study gives , , and . Test two-sided.
Paired t test result in two sample t test: The paired statistic is t=2.616 with df=22; p=0.0158.
Interpretation and validity
Paired t test interpretation for two sample t test: The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation.
Condition evidence for two sample t test and Paired t test: Differences must be independent across pairs and approximately normal or supported by adequate sample size without severe outliers.
Difference scores
Decision
For Difference scores in two sample t test, A constructed sample from a campus dining survey gives , , and . Test two-sided.
Difference scores result in two sample t test: The statistic is t=1.628 with df=28; p=0.1147.
Interpretation and validity
Difference scores interpretation for two sample t test: The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation.
Condition evidence for two sample t test and Difference scores: Use a random sample or randomized design, independence, and a distribution without problematic skew or outliers for the sample size.
Calculator
Decision
For Calculator in two sample t test, A constructed paired study for a tutoring-program evaluation gives , , and . Test two-sided.
Calculator result in two sample t test: The paired statistic is t=3.404 with df=24; p=0.0023.
Interpretation and validity
Calculator interpretation for two sample t test: The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation.
Condition evidence for two sample t test and Calculator: Differences must be independent across pairs and approximately normal or supported by adequate sample size without severe outliers.
Conclusions
Decision
For Conclusions in two sample t test, A constructed sample from a tutoring-program evaluation gives , , and . Test two-sided.
Conclusions result in two sample t test: The statistic is t=2.404 with df=30; p=0.0226.
Interpretation and validity
Conclusions interpretation for two sample t test: The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation.
Condition evidence for two sample t test and Conclusions: Use a random sample or randomized design, independence, and a distribution without problematic skew or outliers for the sample size.
Choosing the correct design
Decision
For Choosing the correct design in two sample t test, A constructed paired study for a package-delivery sample gives , , and . Test two-sided.
Choosing the correct design result in two sample t test: The paired statistic is t=3.897 with df=26; p=0.0006.
Interpretation and validity
Choosing the correct design interpretation for two sample t test: The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation.
Condition evidence for two sample t test and Choosing the correct design: Differences must be independent across pairs and approximately normal or supported by adequate sample size without severe outliers.
Procedure Practice and Full Solutions
Every question in Two-Sample T Test and Paired T Test is newly written from the revised framework and the logic visible in public College Board materials. Constructed numerical settings are identified as instructional scenarios and are never represented as measurements from a real population. No released or secure question wording is reproduced.
Easy Practice
Easy 1: Paired t test
Question P67-Easy-1. A constructed sample from a greenhouse germination experiment gives , , and . Test two-sided.
Worked solution and validity check
Worked solution P67-Easy-1. The statistic is t=2.535 with df=24; p=0.0182. Interpretation: The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation. Validity: Use a random sample or randomized design, independence, and a distribution without problematic skew or outliers for the sample size. Error to reject: The t procedure accounts for estimating sigma with s; substituting a z distribution ignores that uncertainty.
Easy 2: Difference scores
Question P67-Easy-2. A constructed paired study for a tutoring-program evaluation gives , , and . Test two-sided.
Worked solution and validity check
Worked solution P67-Easy-2. The paired statistic is t=3.541 with df=20; p=0.0021. Interpretation: The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation. Validity: Differences must be independent across pairs and approximately normal or supported by adequate sample size without severe outliers. Error to reject: The analysis uses one sample of differences, not two independent samples.
Easy 3: Calculator
Question P67-Easy-3. A constructed sample from a reading-speed investigation gives , , and . Test two-sided.
Worked solution and validity check
Worked solution P67-Easy-3. The statistic is t=2.425 with df=26; p=0.0226. Interpretation: The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation. Validity: Use a random sample or randomized design, independence, and a distribution without problematic skew or outliers for the sample size. Error to reject: The t procedure accounts for estimating sigma with s; substituting a z distribution ignores that uncertainty.
Easy 4: Conclusions
Question P67-Easy-4. A constructed paired study for a manufacturing fill-volume check gives , , and . Test two-sided.
Worked solution and validity check
Worked solution P67-Easy-4. The paired statistic is t=3.681 with df=22; p=0.0013. Interpretation: The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation. Validity: Differences must be independent across pairs and approximately normal or supported by adequate sample size without severe outliers. Error to reject: The analysis uses one sample of differences, not two independent samples.
Easy 5: Choosing the correct design
Question P67-Easy-5. A constructed sample from a recycling-behavior survey gives , , and . Test two-sided.
Worked solution and validity check
Worked solution P67-Easy-5. The statistic is t=2.662 with df=28; p=0.0127. Interpretation: The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation. Validity: Use a random sample or randomized design, independence, and a distribution without problematic skew or outliers for the sample size. Error to reject: The t procedure accounts for estimating sigma with s; substituting a z distribution ignores that uncertainty.
Easy 6: Independent two-sample t test
Question P67-Easy-6. A constructed paired study for a seedling-growth comparison gives , , and . Test two-sided.
Worked solution and validity check
Worked solution P67-Easy-6. The paired statistic is t=2.449 with df=24; p=0.0220. Interpretation: The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation. Validity: Differences must be independent across pairs and approximately normal or supported by adequate sample size without severe outliers. Error to reject: The analysis uses one sample of differences, not two independent samples.
Easy 7: Conditions
Question P67-Easy-7. A constructed sample from a city bus arrival investigation gives , , and . Test two-sided.
Worked solution and validity check
Worked solution P67-Easy-7. The statistic is t=2.007 with df=30; p=0.0538. Interpretation: The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation. Validity: Use a random sample or randomized design, independence, and a distribution without problematic skew or outliers for the sample size. Error to reject: The t procedure accounts for estimating sigma with s; substituting a z distribution ignores that uncertainty.
Easy 8: Welch test
Question P67-Easy-8. A constructed paired study for a manufacturing fill-volume check gives , , and . Test two-sided.
Worked solution and validity check
Worked solution P67-Easy-8. The paired statistic is t=3.792 with df=26; p=0.0008. Interpretation: The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation. Validity: Differences must be independent across pairs and approximately normal or supported by adequate sample size without severe outliers. Error to reject: The analysis uses one sample of differences, not two independent samples.
Easy 9: Paired t test
Question P67-Easy-9. A constructed sample from a city bus arrival investigation gives , , and . Test two-sided.
Worked solution and validity check
Worked solution P67-Easy-9. The statistic is t=2.667 with df=32; p=0.0119. Interpretation: The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation. Validity: Use a random sample or randomized design, independence, and a distribution without problematic skew or outliers for the sample size. Error to reject: The t procedure accounts for estimating sigma with s; substituting a z distribution ignores that uncertainty.
Easy 10: Difference scores
Question P67-Easy-10. A constructed paired study for a water-filtration experiment gives , , and . Test two-sided.
Worked solution and validity check
Worked solution P67-Easy-10. The paired statistic is t=1.948 with df=28; p=0.0615. Interpretation: The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation. Validity: Differences must be independent across pairs and approximately normal or supported by adequate sample size without severe outliers. Error to reject: The analysis uses one sample of differences, not two independent samples.
Easy 11: Calculator
Question P67-Easy-11. A constructed sample from a website response-time study gives , , and . Test two-sided.
Worked solution and validity check
Worked solution P67-Easy-11. The statistic is t=3.295 with df=34; p=0.0023. Interpretation: The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation. Validity: Use a random sample or randomized design, independence, and a distribution without problematic skew or outliers for the sample size. Error to reject: The t procedure accounts for estimating sigma with s; substituting a z distribution ignores that uncertainty.
Easy 12: Conclusions
Question P67-Easy-12. A constructed paired study for a quality-control inspection gives , , and . Test two-sided.
Worked solution and validity check
Worked solution P67-Easy-12. The paired statistic is t=1.979 with df=20; p=0.0618. Interpretation: The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation. Validity: Differences must be independent across pairs and approximately normal or supported by adequate sample size without severe outliers. Error to reject: The analysis uses one sample of differences, not two independent samples.
Easy 13: Choosing the correct design
Question P67-Easy-13. A constructed sample from a website response-time study gives , , and . Test two-sided.
Worked solution and validity check
Worked solution P67-Easy-13. The statistic is t=3.388 with df=36; p=0.0017. Interpretation: The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation. Validity: Use a random sample or randomized design, independence, and a distribution without problematic skew or outliers for the sample size. Error to reject: The t procedure accounts for estimating sigma with s; substituting a z distribution ignores that uncertainty.
Easy 14: Independent two-sample t test
Question P67-Easy-14. A constructed paired study for a seedling-growth comparison gives , , and . Test two-sided.
Worked solution and validity check
Worked solution P67-Easy-14. The paired statistic is t=3.328 with df=22; p=0.0031. Interpretation: The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation. Validity: Differences must be independent across pairs and approximately normal or supported by adequate sample size without severe outliers. Error to reject: The analysis uses one sample of differences, not two independent samples.
Easy 15: Conditions
Question P67-Easy-15. A constructed sample from a public-parks visitor survey gives , , and . Test two-sided.
Worked solution and validity check
Worked solution P67-Easy-15. The statistic is t=3.291 with df=38; p=0.0022. Interpretation: The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation. Validity: Use a random sample or randomized design, independence, and a distribution without problematic skew or outliers for the sample size. Error to reject: The t procedure accounts for estimating sigma with s; substituting a z distribution ignores that uncertainty.
Easy 16: Welch test
Question P67-Easy-16. A constructed paired study for a school library checkout study gives , , and . Test two-sided.
Worked solution and validity check
Worked solution P67-Easy-16. The paired statistic is t=4.286 with df=24; p=0.0003. Interpretation: The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation. Validity: Differences must be independent across pairs and approximately normal or supported by adequate sample size without severe outliers. Error to reject: The analysis uses one sample of differences, not two independent samples.
Easy 17: Paired t test
Question P67-Easy-17. A constructed sample from a quality-control inspection gives , , and . Test two-sided.
Worked solution and validity check
Worked solution P67-Easy-17. The statistic is t=1.500 with df=25; p=0.1462. Interpretation: The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation. Validity: Use a random sample or randomized design, independence, and a distribution without problematic skew or outliers for the sample size. Error to reject: The t procedure accounts for estimating sigma with s; substituting a z distribution ignores that uncertainty.
Easy 18: Difference scores
Question P67-Easy-18. A constructed paired study for a commuter route study gives , , and . Test two-sided.
Worked solution and validity check
Worked solution P67-Easy-18. The paired statistic is t=3.929 with df=26; p=0.0006. Interpretation: The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation. Validity: Differences must be independent across pairs and approximately normal or supported by adequate sample size without severe outliers. Error to reject: The analysis uses one sample of differences, not two independent samples.
Easy 19: Calculator
Question P67-Easy-19. A constructed sample from a tutoring-program evaluation gives , , and . Test two-sided.
Worked solution and validity check
Worked solution P67-Easy-19. The statistic is t=2.772 with df=27; p=0.0100. Interpretation: The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation. Validity: Use a random sample or randomized design, independence, and a distribution without problematic skew or outliers for the sample size. Error to reject: The t procedure accounts for estimating sigma with s; substituting a z distribution ignores that uncertainty.
Tough Practice
Tough 1: Difference scores
Question P67-Tough-1. A constructed sample from a water-filtration experiment gives , , and . Test two-sided.
Worked solution and validity check
Worked solution P67-Tough-1. The statistic is t=2.000 with df=24; p=0.0569. Interpretation: The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation. Validity: Use a random sample or randomized design, independence, and a distribution without problematic skew or outliers for the sample size. Error to reject: The t procedure accounts for estimating sigma with s; substituting a z distribution ignores that uncertainty.
Tough 2: Calculator
Question P67-Tough-2. A constructed paired study for a package-delivery sample gives , , and . Test two-sided.
Worked solution and validity check
Worked solution P67-Tough-2. The paired statistic is t=2.245 with df=20; p=0.0363. Interpretation: The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation. Validity: Differences must be independent across pairs and approximately normal or supported by adequate sample size without severe outliers. Error to reject: The analysis uses one sample of differences, not two independent samples.
Tough 3: Conclusions
Question P67-Tough-3. A constructed sample from a school library checkout study gives , , and . Test two-sided.
Worked solution and validity check
Worked solution P67-Tough-3. The statistic is t=1.856 with df=26; p=0.0749. Interpretation: The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation. Validity: Use a random sample or randomized design, independence, and a distribution without problematic skew or outliers for the sample size. Error to reject: The t procedure accounts for estimating sigma with s; substituting a z distribution ignores that uncertainty.
Tough 4: Choosing the correct design
Question P67-Tough-4. A constructed paired study for a commuter route study gives , , and . Test two-sided.
Worked solution and validity check
Worked solution P67-Tough-4. The paired statistic is t=4.165 with df=22; p=0.0004. Interpretation: The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation. Validity: Differences must be independent across pairs and approximately normal or supported by adequate sample size without severe outliers. Error to reject: The analysis uses one sample of differences, not two independent samples.
Tough 5: Independent two-sample t test
Question P67-Tough-5. A constructed sample from a battery-life laboratory trial gives , , and . Test two-sided.
Worked solution and validity check
Worked solution P67-Tough-5. The statistic is t=2.513 with df=28; p=0.0180. Interpretation: The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation. Validity: Use a random sample or randomized design, independence, and a distribution without problematic skew or outliers for the sample size. Error to reject: The t procedure accounts for estimating sigma with s; substituting a z distribution ignores that uncertainty.
Tough 6: Conditions
Question P67-Tough-6. A constructed paired study for a tutoring-program evaluation gives , , and . Test two-sided.
Worked solution and validity check
Worked solution P67-Tough-6. The paired statistic is t=2.340 with df=24; p=0.0279. Interpretation: The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation. Validity: Differences must be independent across pairs and approximately normal or supported by adequate sample size without severe outliers. Error to reject: The analysis uses one sample of differences, not two independent samples.
Tough 7: Welch test
Question P67-Tough-7. A constructed sample from a seedling-growth comparison gives , , and . Test two-sided.
Worked solution and validity check
Worked solution P67-Tough-7. The statistic is t=1.701 with df=30; p=0.0992. Interpretation: The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation. Validity: Use a random sample or randomized design, independence, and a distribution without problematic skew or outliers for the sample size. Error to reject: The t procedure accounts for estimating sigma with s; substituting a z distribution ignores that uncertainty.
Tough 8: Paired t test
Question P67-Tough-8. A constructed paired study for a package-delivery sample gives , , and . Test two-sided.
Worked solution and validity check
Worked solution P67-Tough-8. The paired statistic is t=1.807 with df=26; p=0.0823. Interpretation: The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation. Validity: Differences must be independent across pairs and approximately normal or supported by adequate sample size without severe outliers. Error to reject: The analysis uses one sample of differences, not two independent samples.
Tough 9: Difference scores
Question P67-Tough-9. A constructed sample from an online-course completion sample gives , , and . Test two-sided.
Worked solution and validity check
Worked solution P67-Tough-9. The statistic is t=1.501 with df=32; p=0.1431. Interpretation: The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation. Validity: Use a random sample or randomized design, independence, and a distribution without problematic skew or outliers for the sample size. Error to reject: The t procedure accounts for estimating sigma with s; substituting a z distribution ignores that uncertainty.
Tough 10: Calculator
Question P67-Tough-10. A constructed paired study for a package-delivery sample gives , , and . Test two-sided.
Worked solution and validity check
Worked solution P67-Tough-10. The paired statistic is t=4.039 with df=28; p=0.0004. Interpretation: The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation. Validity: Differences must be independent across pairs and approximately normal or supported by adequate sample size without severe outliers. Error to reject: The analysis uses one sample of differences, not two independent samples.
Tough 11: Conclusions
Question P67-Tough-11. A constructed sample from a commuter route study gives , , and . Test two-sided.
Worked solution and validity check
Worked solution P67-Tough-11. The statistic is t=1.690 with df=34; p=0.1001. Interpretation: The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation. Validity: Use a random sample or randomized design, independence, and a distribution without problematic skew or outliers for the sample size. Error to reject: The t procedure accounts for estimating sigma with s; substituting a z distribution ignores that uncertainty.
Tough 12: Choosing the correct design
Question P67-Tough-12. A constructed paired study for a greenhouse germination experiment gives , , and . Test two-sided.
Worked solution and validity check
Worked solution P67-Tough-12. The paired statistic is t=3.946 with df=20; p=0.0008. Interpretation: The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation. Validity: Differences must be independent across pairs and approximately normal or supported by adequate sample size without severe outliers. Error to reject: The analysis uses one sample of differences, not two independent samples.
Tough 13: Independent two-sample t test
Question P67-Tough-13. A constructed sample from a city bus arrival investigation gives , , and . Test two-sided.
Worked solution and validity check
Worked solution P67-Tough-13. The statistic is t=2.481 with df=36; p=0.0179. Interpretation: The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation. Validity: Use a random sample or randomized design, independence, and a distribution without problematic skew or outliers for the sample size. Error to reject: The t procedure accounts for estimating sigma with s; substituting a z distribution ignores that uncertainty.
Tough 14: Conditions
Question P67-Tough-14. A constructed paired study for a commuter route study gives , , and . Test two-sided.
Worked solution and validity check
Worked solution P67-Tough-14. The paired statistic is t=4.165 with df=22; p=0.0004. Interpretation: The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation. Validity: Differences must be independent across pairs and approximately normal or supported by adequate sample size without severe outliers. Error to reject: The analysis uses one sample of differences, not two independent samples.
Tough 15: Welch test
Question P67-Tough-15. A constructed sample from a quality-control inspection gives , , and . Test two-sided.
Worked solution and validity check
Worked solution P67-Tough-15. The statistic is t=1.619 with df=38; p=0.1137. Interpretation: The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation. Validity: Use a random sample or randomized design, independence, and a distribution without problematic skew or outliers for the sample size. Error to reject: The t procedure accounts for estimating sigma with s; substituting a z distribution ignores that uncertainty.
Tough 16: Paired t test
Question P67-Tough-16. A constructed paired study for a city bus arrival investigation gives , , and . Test two-sided.
Worked solution and validity check
Worked solution P67-Tough-16. The paired statistic is t=2.973 with df=24; p=0.0066. Interpretation: The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation. Validity: Differences must be independent across pairs and approximately normal or supported by adequate sample size without severe outliers. Error to reject: The analysis uses one sample of differences, not two independent samples.
Tough 17: Difference scores
Question P67-Tough-17. A constructed sample from a seedling-growth comparison gives , , and . Test two-sided.
Worked solution and validity check
Worked solution P67-Tough-17. The statistic is t=2.914 with df=25; p=0.0074. Interpretation: The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation. Validity: Use a random sample or randomized design, independence, and a distribution without problematic skew or outliers for the sample size. Error to reject: The t procedure accounts for estimating sigma with s; substituting a z distribution ignores that uncertainty.
Tough 18: Calculator
Question P67-Tough-18. A constructed paired study for a water-filtration experiment gives , , and . Test two-sided.
Worked solution and validity check
Worked solution P67-Tough-18. The paired statistic is t=2.985 with df=26; p=0.0061. Interpretation: The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation. Validity: Differences must be independent across pairs and approximately normal or supported by adequate sample size without severe outliers. Error to reject: The analysis uses one sample of differences, not two independent samples.
Tough 19: Conclusions
Question P67-Tough-19. A constructed sample from a battery-life laboratory trial gives , , and . Test two-sided.
Worked solution and validity check
Worked solution P67-Tough-19. The statistic is t=2.612 with df=27; p=0.0145. Interpretation: The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation. Validity: Use a random sample or randomized design, independence, and a distribution without problematic skew or outliers for the sample size. Error to reject: The t procedure accounts for estimating sigma with s; substituting a z distribution ignores that uncertainty.
Toughest Practice
Toughest 1: Conclusions
Question P67-Toughest-1. A constructed sample from a reading-speed investigation gives , , and . Test two-sided.
Worked solution and validity check
Worked solution P67-Toughest-1. The statistic is t=2.403 with df=24; p=0.0244. Interpretation: The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation. Validity: Use a random sample or randomized design, independence, and a distribution without problematic skew or outliers for the sample size. Error to reject: The t procedure accounts for estimating sigma with s; substituting a z distribution ignores that uncertainty.
Toughest 2: Choosing the correct design
Question P67-Toughest-2. A constructed paired study for a quality-control inspection gives , , and . Test two-sided.
Worked solution and validity check
Worked solution P67-Toughest-2. The paired statistic is t=3.273 with df=20; p=0.0038. Interpretation: The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation. Validity: Differences must be independent across pairs and approximately normal or supported by adequate sample size without severe outliers. Error to reject: The analysis uses one sample of differences, not two independent samples.
Toughest 3: Independent two-sample t test
Question P67-Toughest-3. A constructed sample from a battery-life laboratory trial gives , , and . Test two-sided.
Worked solution and validity check
Worked solution P67-Toughest-3. The statistic is t=1.386 with df=26; p=0.1776. Interpretation: The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation. Validity: Use a random sample or randomized design, independence, and a distribution without problematic skew or outliers for the sample size. Error to reject: The t procedure accounts for estimating sigma with s; substituting a z distribution ignores that uncertainty.
Toughest 4: Conditions
Question P67-Toughest-4. A constructed paired study for a manufacturing fill-volume check gives , , and . Test two-sided.
Worked solution and validity check
Worked solution P67-Toughest-4. The paired statistic is t=1.798 with df=22; p=0.0858. Interpretation: The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation. Validity: Differences must be independent across pairs and approximately normal or supported by adequate sample size without severe outliers. Error to reject: The analysis uses one sample of differences, not two independent samples.
Toughest 5: Welch test
Question P67-Toughest-5. A constructed sample from a campus dining survey gives , , and . Test two-sided.
Worked solution and validity check
Worked solution P67-Toughest-5. The statistic is t=1.488 with df=28; p=0.1479. Interpretation: The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation. Validity: Use a random sample or randomized design, independence, and a distribution without problematic skew or outliers for the sample size. Error to reject: The t procedure accounts for estimating sigma with s; substituting a z distribution ignores that uncertainty.
Toughest 6: Paired t test
Question P67-Toughest-6. A constructed paired study for a public-parks visitor survey gives , , and . Test two-sided.
Worked solution and validity check
Worked solution P67-Toughest-6. The paired statistic is t=2.368 with df=24; p=0.0263. Interpretation: The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation. Validity: Differences must be independent across pairs and approximately normal or supported by adequate sample size without severe outliers. Error to reject: The analysis uses one sample of differences, not two independent samples.
Toughest 7: Difference scores
Question P67-Toughest-7. A constructed sample from an online-course completion sample gives , , and . Test two-sided.
Worked solution and validity check
Worked solution P67-Toughest-7. The statistic is t=1.683 with df=30; p=0.1027. Interpretation: The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation. Validity: Use a random sample or randomized design, independence, and a distribution without problematic skew or outliers for the sample size. Error to reject: The t procedure accounts for estimating sigma with s; substituting a z distribution ignores that uncertainty.
Toughest 8: Calculator
Question P67-Toughest-8. A constructed paired study for a battery-life laboratory trial gives , , and . Test two-sided.
Worked solution and validity check
Worked solution P67-Toughest-8. The paired statistic is t=1.949 with df=26; p=0.0622. Interpretation: The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation. Validity: Differences must be independent across pairs and approximately normal or supported by adequate sample size without severe outliers. Error to reject: The analysis uses one sample of differences, not two independent samples.
Toughest 9: Conclusions
Question P67-Toughest-9. A constructed sample from a recycling-behavior survey gives , , and . Test two-sided.
Worked solution and validity check
Worked solution P67-Toughest-9. The statistic is t=1.783 with df=32; p=0.0841. Interpretation: The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation. Validity: Use a random sample or randomized design, independence, and a distribution without problematic skew or outliers for the sample size. Error to reject: The t procedure accounts for estimating sigma with s; substituting a z distribution ignores that uncertainty.
Toughest 10: Choosing the correct design
Question P67-Toughest-10. A constructed paired study for a manufacturing fill-volume check gives , , and . Test two-sided.
Worked solution and validity check
Worked solution P67-Toughest-10. The paired statistic is t=3.629 with df=28; p=0.0011. Interpretation: The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation. Validity: Differences must be independent across pairs and approximately normal or supported by adequate sample size without severe outliers. Error to reject: The analysis uses one sample of differences, not two independent samples.
Toughest 11: Independent two-sample t test
Question P67-Toughest-11. A constructed sample from a manufacturing fill-volume check gives , , and . Test two-sided.
Worked solution and validity check
Worked solution P67-Toughest-11. The statistic is t=3.485 with df=34; p=0.0014. Interpretation: The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation. Validity: Use a random sample or randomized design, independence, and a distribution without problematic skew or outliers for the sample size. Error to reject: The t procedure accounts for estimating sigma with s; substituting a z distribution ignores that uncertainty.
Toughest 12: Conditions
Question P67-Toughest-12. A constructed paired study for a recycling-behavior survey gives , , and . Test two-sided.
Worked solution and validity check
Worked solution P67-Toughest-12. The paired statistic is t=2.106 with df=20; p=0.0481. Interpretation: The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation. Validity: Differences must be independent across pairs and approximately normal or supported by adequate sample size without severe outliers. Error to reject: The analysis uses one sample of differences, not two independent samples.
Toughest 13: Welch test
Question P67-Toughest-13. A constructed sample from a city bus arrival investigation gives , , and . Test two-sided.
Worked solution and validity check
Worked solution P67-Toughest-13. The statistic is t=1.632 with df=36; p=0.1114. Interpretation: The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation. Validity: Use a random sample or randomized design, independence, and a distribution without problematic skew or outliers for the sample size. Error to reject: The t procedure accounts for estimating sigma with s; substituting a z distribution ignores that uncertainty.
Toughest 14: Paired t test
Question P67-Toughest-14. A constructed paired study for a recycling-behavior survey gives , , and . Test two-sided.
Worked solution and validity check
Worked solution P67-Toughest-14. The paired statistic is t=2.838 with df=22; p=0.0096. Interpretation: The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation. Validity: Differences must be independent across pairs and approximately normal or supported by adequate sample size without severe outliers. Error to reject: The analysis uses one sample of differences, not two independent samples.
Toughest 15: Difference scores
Question P67-Toughest-15. A constructed sample from a manufacturing fill-volume check gives , , and . Test two-sided.
Worked solution and validity check
Worked solution P67-Toughest-15. The statistic is t=2.082 with df=38; p=0.0442. Interpretation: The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation. Validity: Use a random sample or randomized design, independence, and a distribution without problematic skew or outliers for the sample size. Error to reject: The t procedure accounts for estimating sigma with s; substituting a z distribution ignores that uncertainty.
Toughest 16: Calculator
Question P67-Toughest-16. A constructed paired study for a recycling-behavior survey gives , , and . Test two-sided.
Worked solution and validity check
Worked solution P67-Toughest-16. The paired statistic is t=2.283 with df=24; p=0.0316. Interpretation: The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation. Validity: Differences must be independent across pairs and approximately normal or supported by adequate sample size without severe outliers. Error to reject: The analysis uses one sample of differences, not two independent samples.
Toughest 17: Conclusions
Question P67-Toughest-17. A constructed sample from a commuter route study gives , , and . Test two-sided.
Worked solution and validity check
Worked solution P67-Toughest-17. The statistic is t=2.234 with df=25; p=0.0346. Interpretation: The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation. Validity: Use a random sample or randomized design, independence, and a distribution without problematic skew or outliers for the sample size. Error to reject: The t procedure accounts for estimating sigma with s; substituting a z distribution ignores that uncertainty.
Toughest 18: Choosing the correct design
Question P67-Toughest-18. A constructed paired study for a greenhouse germination experiment gives , , and . Test two-sided.
Worked solution and validity check
Worked solution P67-Toughest-18. The paired statistic is t=2.103 with df=26; p=0.0453. Interpretation: The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation. Validity: Differences must be independent across pairs and approximately normal or supported by adequate sample size without severe outliers. Error to reject: The analysis uses one sample of differences, not two independent samples.
Toughest 19: Independent two-sample t test
Question P67-Toughest-19. A constructed sample from a water-filtration experiment gives , , and . Test two-sided.
Worked solution and validity check
Worked solution P67-Toughest-19. The statistic is t=1.724 with df=27; p=0.0961. Interpretation: The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation. Validity: Use a random sample or randomized design, independence, and a distribution without problematic skew or outliers for the sample size. Error to reject: The t procedure accounts for estimating sigma with s; substituting a z distribution ignores that uncertainty.
AP Response and Publication Checklist
| Audit point | Required evidence for two sample t test |
|---|---|
| Scope | Never analyze paired observations as independent samples. |
| Method or source | Two-sample t inference analyzes independent groups with a Welch standard error, while paired t inference analyzes one sample of within-pair differences. |
| Calculation | |
| Interpretation | The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation. |
| Validity | Use a random sample or randomized design, independence, and a distribution without problematic skew or outliers for the sample size. |
| Correction | The t procedure accounts for estimating sigma with s; substituting a z distribution ignores that uncertainty. |
Frequently Asked Questions
How does independent two-sample t test work in two sample t test?
Answer for two sample t test and Independent two-sample t test. The statistic is t=2.088 with df=34; p=0.0444. The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation. The required validity evidence is: Use a random sample or randomized design, independence, and a distribution without problematic skew or outliers for the sample size.
How does conditions work in two sample t test?
Answer for two sample t test and Conditions. The paired statistic is t=2.106 with df=20; p=0.0481. The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation. The required validity evidence is: Differences must be independent across pairs and approximately normal or supported by adequate sample size without severe outliers.
How does welch test work in two sample t test?
Answer for two sample t test and Welch test. The statistic is t=2.281 with df=36; p=0.0286. The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation. The required validity evidence is: Use a random sample or randomized design, independence, and a distribution without problematic skew or outliers for the sample size.
How does paired t test work in two sample t test?
Answer for two sample t test and Paired t test. The paired statistic is t=2.055 with df=22; p=0.0519. The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation. The required validity evidence is: Differences must be independent across pairs and approximately normal or supported by adequate sample size without severe outliers.
How does difference scores work in two sample t test?
Answer for two sample t test and Difference scores. The statistic is t=3.052 with df=38; p=0.0041. The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation. The required validity evidence is: Use a random sample or randomized design, independence, and a distribution without problematic skew or outliers for the sample size.
How does calculator work in two sample t test?
Answer for two sample t test and Calculator. The paired statistic is t=4.324 with df=24; p=0.0002. The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation. The required validity evidence is: Differences must be independent across pairs and approximately normal or supported by adequate sample size without severe outliers.
How does hypothesis for paired t test connect to Two Sample t-Test?
hypothesis for paired t test within two sample t test. The statistic is t=2.101 with df=29; p=0.0445. The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation. For Independent two-sample t test, the controlling scope is: Never analyze paired observations as independent samples.
How does hypothesis testing paired t test connect to Two Sample t-Test?
hypothesis testing paired t test within two sample t test. The paired statistic is t=2.291 with df=20; p=0.0329. The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation. For Conditions, the controlling scope is: Never analyze paired observations as independent samples.
How does paired t test hypothesis connect to Two Sample t-Test?
paired t test hypothesis within two sample t test. The statistic is t=2.351 with df=31; p=0.0253. The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation. For Welch test, the controlling scope is: Never analyze paired observations as independent samples.
How does two sample t test hypothesis connect to Two Sample t-Test?
two sample t test hypothesis within two sample t test. The paired statistic is t=4.130 with df=22; p=0.0004. The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation. For Paired t test, the controlling scope is: Never analyze paired observations as independent samples.
How does hypothesis for two sample t test connect to Two Sample t-Test?
hypothesis for two sample t test within two sample t test. The statistic is t=2.841 with df=33; p=0.0077. The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation. For Difference scores, the controlling scope is: Never analyze paired observations as independent samples.
How does hypothesis paired t test connect to Two Sample t-Test?
hypothesis paired t test within two sample t test. The paired statistic is t=2.619 with df=24; p=0.0150. The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation. For Calculator, the controlling scope is: Never analyze paired observations as independent samples.
How does hypothesis of paired t test connect to Two Sample t-Test?
hypothesis of paired t test within two sample t test. The statistic is t=2.447 with df=35; p=0.0196. The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation. For Conclusions, the controlling scope is: Never analyze paired observations as independent samples.
How does paired t test hypothesis example connect to Two Sample t-Test?
paired t test hypothesis example within two sample t test. The paired statistic is t=3.248 with df=26; p=0.0032. The p-value quantifies evidence against the null mean; the design determines whether the result generalizes or supports causation. For Choosing the correct design, the controlling scope is: Never analyze paired observations as independent samples.
Sources
Administrative and curricular statements in Two-Sample T Test and Paired T Test were checked on July 18, 2026. The linked College Board pages control any later policy change; all instructional datasets in original questions are explicitly constructed rather than attributed to a real study.
Two Sample t-Test Conclusion
Two-sample t inference analyzes independent groups with a Welch standard error, while paired t inference analyzes one sample of within-pair differences. Mastery of two sample t test therefore requires the exact evidence, mathematics, interpretation, and scope developed in this guide, while preserving this boundary: Never analyze paired observations as independent samples.