Associations between Variables Flashcards
What might correlational data look like?
Correlational designs are useful for variables that are best measured as they exist naturally in the world, or that are impossible or unethical to manipulate
Pearson’s correlation coefficient
A statistical method used to measure similarity or correlation between two data objects
What are parametric tests?
Branch of statistics which leverages models based on a fixed set of parameters
Relies on a set of assumptions, more robust, more powerful
What are non-parametric tests?
A type of statistical analysis that makes minimal assumptions about the underlying distribution of the data being studied
More likely to find a false negative
Explain linearity in regard to variables
If not linear, the result underestimates the degree of the relationship between the variables (false negative)
Check by exploring the shape of your data using a scatterplot, maybe also adding a line of best fit
Homoscedasticity
Is a cigar-shaped distribution
If variance is not equal for all values of the IV, the result overestimates the degree of the relationship between the variables (false positive)
Outliers
If outliers are present, results may falsely suggest there is a correlation (false positive)
Check by exploring your data using a box plot or a scatterplot
What is Spearman’s rho
Compares the relationships between two variables, used for ordinal level or higher
It is a non-parametric test and therefore suitable for non-parametric data
Converts raw scores into ranks, then calculates linear relationships between the ranks
Loss of information = less powerful than parametric test
What is Pearson’s correlation coefficient?
Is the most common way of measuring a linear correlation between two variables
How do we interpret correlation coefficients?
