Return to Statistics Homepage - Here -            Return to Experimenta Psych Homepage  - Here -
You start with a research question that includes a data collection component to the question
|              |                                            | 
Level of Data?
Nominal
|
Ordinal
    | 
Interval/Ratio
|
Type of Data?
nonparametric
|
            nonparametric
|
parametric
   if not parametric
              |
Options to describe your data?
Discriptive Statistics

Graph
 

Measures
of Central
Tendency




    Measures of Variability
 
 
 
 

 
Looking at Relationships
      (Correlations)
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Looking at relationships(Predictions)
 
 

Inferential Statistics
 

Measuring Differences
 
|
bar
|
|
|
bar
histogram
|
           |                |
/         |              bar
\                   histogram
          frequency polygon
          mode       median                   mean
       mode                  median 
                  mode
      Range                    Range
                Range
        Semi-Interquartile Range     Semi-Interquartile Range
      Standard Deviation
               Z-Scores
Contingency coefficient
           Spearman Rank		 Pearson's 
     Phi-Coefficient      Kendall Rank (small data)

 
         Other correlation techniques where variables are:
Artificial dichotomy and Continuous =                                Biserial Correlation

Widespread artificial dichotomy and Continuous =              Widespread Biserial 
                                                                                          Correlation

True dichotomy and Continuous =                                       Point Biserial 
                                                                                          Correlation

2 or more categories and 2 or more categories =                 Contingency 
                                                                                          Coefficient
 
 

 
 
Linear Regression

 

Chi-square (Goodness-of-fit test) 

 2 groups and 2 conditions/independent samples
(Mann-Whitney U test)		 one group and infinite # of conditions:
    (one sample dependent T-test)
Chi-square (test of independence) 2 groups and 2 conditions/dependent samples
(Wilcoxon signed ranks T- test)		 2 groups and 2 conditions
     ( dependent T-test ) 
  2 groups and 3 or more conditions
         (Kruskal-Wallis Test)		 2 groups and 2 conditions
       ( independent T-test)
3 or more groups and 2 conditions
         (Kurskal-Wallis Test)		 2 groups and 3 or more conditions
        ( one way ANOVA)
2 or more groups and 3 or more conditions: where each group experiences each condition
     (Friedman's Test		 3 or more groups and 2 conditions
        (two way ANOVA)
2 or more groups and 3 or more conditions: where each group experiences each conditon
    (Repeated Measures Anova)