Statistical tests

• What are the 2 classifications of statistical tests


  -parametric

  -non-parametric
• Criteria for parametric tests


  -populations of data drawn should be normally distributed

  -variances of populations should be approximately equal

  -should have at least interval or ratio data

  -should be no extreme scores
• Criteria for non-parametric tests


  -data is skewed

  -variances are not equal

  -can have nominal, ordinal or interval data

  -may be extreme scores
• The types of non-parametric tests


  -Mann Whitney U test

  -Chi Square

  -Binomial Sign test

  -Wilcoxon Signed Ranks test

  -Correlations Spearman's Rho
• How to choose the most appropriate statistical test


  :

  
• Example Q: A researcher tested a new drug for depression by giving one group the new drug and another group the placebo. Ps rate their progress on a self-report scale. what statistical test should be used?


  Mann Whitney

  -difference between a new drug + placebo for depression

  -ordinal data -> progress on a rating scale of 1-10

  -independent measures design - Ps either had the new drug or had the placebo
• Example Q: A researcher wants to investigate if there's a relationship between age and number of wrinkles. Ask Ps 2 Qs - how old they are in years and to estimate how many wrinkles they have. what statistical test should be used?


  Spearman's Rho

  -relationship between age + number of wrinkles

  -interval -> age is in years and number of wrinkles are both known units with regular intervals

  -measure co-variables with no manipulation
• how to rank data


  :

  
• Why do psychologists run stat tests 


  -to check the significance of results

  -must be certain that results of DV are caused by the IV and not just due to chance
• What is the probability level in psychology


  P equal than less to 0.05 meaning we are 95% certain results aren't due to chance