Stats Revision

Stats Revision

1. DESIGNING THE RESEARCH

   • Sampling methods - Stratified Random, cluster, systematic, convenience, self selected
   • Observational or Experimental
   • Independent Groups, Matched Pairs, Repeated measures
   • Nuisance Variables
   • Confounding Factors
   • Randomisation

2. HYPOTHESIS TESTING

   • IS OURS BIGGER THAN YOURS? Difference between two independent groups.
     • FROM SAMPLES TO POPULATIONS. Indépendent Samples t-test & Assumptions
       • Metric Data
       • Normally distributed Populations
       • Equal Variances - Levenes Test
       • Independence of Observations
     • Comparing Distributions - EXPLORE - Percentiles, Descriptives, Histogram, Box Plot
     • Independent Groups Design
   • IS OUR THIS BIGGER THAT OUR THAT? Comparing two measurements of the same thing
     • Comparing Distributions - EXPLORE - Percentiles, Descriptives, Histogram, Box Plot
     • FROM SAMPLES TO POPULATIONS. Paired Samples t-test & Assumptions
       • Metric Data
       • Independence of Observations
       • Normally Distributed Populations
     • Use Repeated Measures Design
   • Single Variable
     • Metric Data
       • DESCRIBING THE DISTRIBUTION. Use the EXPLORE procedure. Summary stats - Mean, Median, sd, quartiles. Histogram & box plot
       • HYPOTHESIS TESTING. Comparing a mean to a reference mean. One sample t-test & 95% confidence interval
     • Categorical Data
       • DESCRIBING THE DISTRIBUTION. Use the FREQUENCIES procedure. Frequency table, percentage bar chat, pie chart
       • HYPOTHESIS TESTING. Comparing a proportion to a reference proportion. Binomial test & 95% confidence interval

3. DOES MORE OF THIS MEAN MORE OF THAT? Relationships between Metric Variables

   • DESCRIBING RELATIONSHIPS. Scatter Plot
     • Direction
     • Form
     • Strength
     • Outliers
   • FROM SAMPLES TO POPULATIONS. Pearson Correlation Coefficient determines the strength of a relationship. 95% Confidence interval for the correlation coefficient.
     • THE REGRESSION EQUATION. Describes the form of the relationship and enables predictions to be made
   • FROM SAMPLES TO POPULATION