Flow chart Flashcards
What are the assumptions of a parametric test?
- Normality of the data - normal distribution
- Homogeneity of variance - data groups have same variance
- Linearity - data has a linear relationship
- Independence - data is independent
How can you test for normality?
Shapiro-Wilk test Kolmogorov-Smirnov (lots of data) Histograms - analyse the data Box-plots Skewness Kurtosis
When do we use correlation to analyse data?
To test for relationships between data
What is regression?
The assumption that one variable is dependent on the other, e.g. the Y variables are caused by the X variable
What is categorical data?
Data that can be organised into categories, e.g. race and sex
What is continuous data?
Data that can be measured on a continuum or scale
If ratio data ISN’T used then which test is used?
Non-parametric
Which test is used if you want to test frequencies of categorical data between samples?
Chi squared (non parametric) either one way or two way depending on amount of categories
If you are looking for differences between 2 independent samples, which tests are suitable?
Parametric: T-test
Non parametric: Wilcoxon rank sum test
If you are looking for differences between 3 or more samples that are independent, which tests are suitable?
Parametric: One way ANOVA
Non parametric: Krustil Wallis test
If you are looking for differences between paired/related samples, which tests are used?
Parametric: Paired t-test
Non-parametric: Wilcoxon signed rank test
If you are looking at relationships between variables and want to make a prediction, which test is used?
Regression analysis
If you are looking at relationships between variables and want to assess the relationship, which test is used?
Correlation coefficient
If you are looking at the expected outcome of categorical data, which test would you use?
Chi squared goodness of fit
If you are looking at the relationship of categorical data, which test would you use?
Chi squared, two way association
