week 2 correlation and partial correlation Flashcards
What is a monotonic vs a non-monotonic relationship?
In non-monotonic relationships, the relationship changes direction partway through, in monotonic relationships it may not be linear but it never changes direction.
When can you use Pearsons r and spearman’s rho in regards to monotonic relationship and linearity?
If it is monotonic and linear you can use both, if it is not linear you can only use spearman’s rho, if it is not monotonic or linear you cannot use either.
What is effect size?
an objective measure of the magnitude/strength/size of an observed effect.
What is the r-value?
the strength and direction of the relationship between 2 variables.
What are the bounds for effect size of rvalues?
.10-.29 is a small effect size, .30-.49 is medium effect size, .50-1 is large effect size.
the correlation coefficient (r) can tell us what?
the strength and direction of relationship and effect size.
How do you run correlations in SPSS?
check for outliers, plot variables on a scatterplot (linearly), determine the direction of the relationship, determine what test to use, conduct correlation, check confidence intervals, and report findings.
What is power?
the ability of a test to detect an effect, assuming one exists at the population level.
What do partial correlations do?
take a correlation between two variables and makes an adjustment for the fact that they both correlate with a third variable.
What are the assumptions for spearman’s rho?
ordinal, no requirement for normal distribution, copes fairly well with monotonic relationships.
What are the assumptions for Pearson’s r?
continuous data, technically should be normally distributed but does not have to be, must be linear.
scatterplots are useful to provide what?
Clues about r (directionality and the size), Indication of potential assumption violations.
What is the possible range of r-values?
-1 to 1.
What can we derive from an r-value?
effect size, direction of relationship, strength of relationship.
