Questions and lessons can't be printed or saved as a PDF.
What does the correlation coefficient, chosen for this data type, say about the link between two variables?
More, if you want it
What does the patient's doctor want to know about the patient's two scores?Tap a term: its definition, what you can do with it, and how often the exam asks it.
Definition: a link in which both variables rise together.
You can: Tell a positive correlation from a negative correlation by the sign of r.
Definition: a link in which one variable rises as the other falls.
You can: Judge a negative correlation's strength by the size of r, ignoring the sign.
Definition: the test for two continuous, roughly normally distributed variables.
You can: Tell Pearson's correlation from Spearman's rank correlation by continuous, roughly normal scores.
Definition: the test for ordinal or non-normally distributed data.
You can: Tell Spearman's rank correlation from Pearson's correlation by ordinal or skewed scores.
Definition: the test for two binary variables.
You can: Tell the phi coefficient from Spearman's rank correlation by two yes or no variables.
Each card is one term. The grey line is what the patient says, which is what the question stem gives.
Summary
A correlation coefficient is a number that always lies between -1 and +1.
A is a number that always lies between -1 and +1. The letter r stands for the coefficient.
A coefficient of -0.9 is stronger than +0.7. The value -1 is the strongest possible.
Test-retest reliability comes from giving the same test again to a group and correlating the results. Odd-numbered test items can be correlated with even-numbered items to check internal consistency. In reliability work, a higher coefficient means a better test instrument.
Dr Okafor reads depression trials in a journal club. Dr Okafor says, "One trial shows r = -0.9, and the other shows r = +0.7."
Candidates often treat a negative r as weak and the bigger positive number as stronger. The size of r, ignoring the sign, gives the strength of the link.
The sign of r gives only the direction. So the -0.9 trial shows the stronger link.
Mnemonic
Pearson's smooth bell, Spearman's ranked queue, phi's yes-or-no flip. Cue: strongest association
Pearson's correlation suits smooth, bell-shaped continuous scores. Scores that stand in a ranked queue, meaning ordinal or skewed data, suit Spearman's rank correlation. The phi coefficient suits a variable that flips between two answers, yes or no.
5 more sections on Correlation coefficients: Pearson's, Spearman's rank, phi and more
Log in free to read the rest: the full mental model, exam recognition, and the look-alikes that trip people up. Your daily session is free, always.
mrcpsych.co.uk · An independent study tool · not affiliated with the Royal College of Psychiatrists