Inferential Statistical Tests

• Tells you whether your results are significant or not
• Allows you to find out whether experimental or correlation findings are significant
• Start by checking whether your results are significant at the 0.05 level
• Experiment - Want to see the IV produce real effects on the DV
• Correlation - Want the co-variable to have a relationship that is strong enough to be confident that it isn't by chance
• Math test names-

  1. Sign test
  2. Wilcoxon
  3. Mann-Whitney
  4. Related t-test
  5. Unrelated t-test
  6. Chi-squared
  7. Spearman's rho
  8. Pearson's r
• Nominal -

  Independent group - Chi-squared

  Repeated measures/matched pairs - Sign test
• Ordinal-

  Independent group - Mann-Whitney

  Repeated measures/matched pairs - Wilcoxon
• Interval/ratio -

  Independent group - Unrelated t-test

  Repeated measures/matched pairs - Related t-test
• Correlation -

  Ordinal and/or Interval/ratio - Spearman's rho

  Interval/ratio - Pearson's r
• Which test you decide to choose depends on 3 things:

  1. Are you looking for a relationship (correlation) or a difference (experiment)
  2. Is it a MP, IG or RMs design
  3. Is the data from the study nominal, ordinal or interval/ratio
• If the results are statistically significant them p<0.05 (probability is less than 0.05)
• If the results are non-significant then p>0.05 (probability is more than 0.05)
• p<0.05 - probability is less to be due to chance and more to the researcher's manipulation of the IV - alternative hypothesis is supported and the null hypothesis is rejected
• p>0.05 - probability is less to be due to the researchers manipulation of the IV and more due to chance - null hypothesis is supported and the alternative hypothesis is rejected