Lesson 4.2.3.3.1

4.2.3.3.1 Sign test and probability, significance and errors Quiz: AQA Psychology, Unit 2

20 questions

In partnership with Revision Ninja

Lesson 4.2.3.3.1, Sign test and probability, significance and errors: 20 multiple choice questions for the AQA Psychology (7182), Unit 2: Psychology in context, written with Revision Ninja.

Host it live on the board and students join with a game code on their own devices, or revise alone with Free Play. The answers are revealed in the game.

Host this setFree Play

The 20 questions

  1. What type of design is the sign test typically used with?

    • Independent groups with interval data only
    • Repeated measures or matched pairs, where each participant gives two related scores
    • Case studies analysed qualitatively
    • Correlational designs measuring two continuous variables
  2. Why is the sign test described as a non-parametric test?

    • It makes no assumption that the data follow a normal distribution
    • It requires a sample of at least 100 participants
    • It can only be used with ratio level data
    • It always gives a more powerful result than a related t-test
  3. In the sign test, what is recorded for each participant?

    • The exact numerical difference between the two scores
    • The rank of each participant's total score
    • Whether the difference between the two scores is positive, negative or zero
    • The mean of the two conditions for that participant
  4. In the sign test, how are participants with no difference between their two scores treated?

    • They are removed from the analysis, so N is reduced
    • They are counted as positive differences
    • They are counted as negative differences
    • They are split equally between the positive and negative categories
  5. What is a Type I error?

    • Failing to randomise participants to conditions
    • Using the wrong test for the level of measurement
    • Rejecting the null hypothesis when it is actually true, a false positive
    • Accepting the null hypothesis when it is actually false, a false negative
  6. What is a Type II error?

    • Rejecting the null hypothesis when it is actually true
    • Using a one-tailed test when a two-tailed test was predicted
    • Accepting the null hypothesis when a real effect exists, a false negative
    • Setting the significance level at 0.05 rather than 0.01
  7. What significance level is conventionally used in psychology?

    • 0.5 (50%)
    • 0.001 (0.1%)
    • 0.05 (5%)
    • 0.95 (95%)
  8. In a sign test, 10 participants are tested. Eight show an increase and two show a decrease, with no zero differences. What is S, the smaller number of signs?

    • 8
    • 10
    • 5
    • 2
  9. A sign test with N = 10 gives S = 2. The critical value for a two-tailed test at the 0.05 level is 1. Is the result significant?

    • Yes, because S is less than the critical value
    • Yes, because S is greater than the critical value
    • No, because S is less than the critical value
    • No, because S is greater than the critical value
  10. The same data (N = 10, S = 2) are analysed with a one-tailed test at the 0.05 level, where the critical value is 2. What is the conclusion?

    • Significant, because N is 10
    • Not significant, because there were no zero differences
    • Not significant, because S is larger than 2
    • Significant, because S equals the critical value of 2
  11. A study has 12 participants and one shows no change between conditions. What is N for the sign test?

    • 12
    • 1
    • 10
    • 11
  12. Which observed value of S, with N = 10, would be significant on a one-tailed test at the 0.05 level, where the critical value is 2?

    • S = 3
    • S = 5
    • S = 4
    • S = 2
  13. A sign test has 15 non-zero differences, with 11 positive and 4 negative. What is S?

    • 7
    • 15
    • 11
    • 4
  14. A psychologist tests the same students before and after a revision programme, using scores that are ranked, not interval. Which test is suitable?

    • Spearman's rho
    • Sign test
    • Chi-squared test
    • Unrelated t-test
  15. A researcher lowers the significance level from 0.05 to 0.01. What is the effect on the risk of a Type I error?

    • It becomes zero, because the test is now exact
    • It is reduced, because the criterion for significance is stricter
    • It is unchanged, because it depends only on sample size
    • It is increased, because more results will become significant
  16. A sign test shows a significant result, but the effect is later shown not to exist. Which error has occurred?

    • A Type I error
    • A ceiling effect
    • A demand characteristic
    • A Type II error
  17. Why might a researcher choose a two-tailed test rather than a one-tailed test?

    • A two-tailed test allows the researcher to ignore zero differences
    • A two-tailed test is always more powerful than a one-tailed test
    • There is no clear directional prediction, so an effect in either direction should be detectable
    • A one-tailed test always has a higher chance of a Type I error
  18. A sign test with 10 participants gives S = 4. The one-tailed critical value at 0.05 is 2. What is the correct conclusion?

    • Significant, because S is an even number
    • Significant, so the null hypothesis is rejected
    • Not significant, so the null hypothesis is retained
    • Not significant, because the sign test cannot be used with N = 10
  19. Lowering the significance level reduces the risk of one error type but increases the risk of another. Which is the correct trade-off?

    • Lowering alpha increases false positives but reduces false negatives
    • Lowering alpha reduces both Type I and Type II errors equally
    • Lowering alpha has no effect on either type of error
    • Lowering alpha reduces false positives but increases the risk of false negatives
  20. Which is a strength of the sign test?

    • It can only be used with independent groups
    • It uses the actual size of each difference, so it is more precise than a related t-test
    • It is simple to calculate and does not assume normally distributed data
    • It always detects small effects better than parametric tests

All AQA Psychology quizzes