Stats - Correlation
Flashcards
2020-12-10
What do you call data with two variables?
Bivariate data.
What is bivariate data?
Data with two variables.
What is correlation?
A measure of the strength and type of a relationship between two variables.
What type of correlation is this?
Strong positive correlation.
What type of correlation is this?
Weak positive correlation.
What type of correlation is this?
Strong negative correlation.
What type of correlation is this?
Weak negative correlation.
What type of correlation is this?
No correlation.
What’s another name for the independent variable?
Explanatory variable.
What’s another name for the dependent variable?
Response variable.
When asked to interpret or describe the correlation between two variables, what must you make sure to do?
State the problem in context.
What does it mean for two variables to have a causal relationship?
Changing one variables causes the other to also change.
What is the general equation found by a linear regression?
What is a linear regression?
A technique used to come up with a model for a linear relationship.
What is the most common type of linear regression?
A Least-Squares Regression.
What does a Least-Squares Regression involve?
Finding a line that minimises the distance squared from each point to the line.
When can you use a linear regression?
When the correlation suggests there’s a linear relationship between two variables.
What is interpolating?
Estimating a value inside the range of data.
What is extrapolating?
Estimating a value outside the range of data.
Which should you use a regression line for: interpolating or extrapolating?
Extrapolating.
2022-01-19
What does $r$ normally represent in stats?
The PMCC (product moment correlation coefficient) for a sample.
What does $\rho$ normally represent in stats?
The PMCC for an entire population.
What is the range of the product moment correlation coefficient (PMCC)?
Between -1 and 1.
What do values of a PMCC $r$ greater than $0.7$ or less than $-0.7$ correspond to?
A strong correlation.
What does a PMCC $r$ value of $0$ correspond to?
Zero correlation.
How can you work out the line of best fit and correlation coefficient for two variables?
Using the $y = a + bx$ option on the calculator.
What are you using a PMCC for when doing a hypothesis test?
Seeing how likely it is that a sample is correlated given the population size.
What is typically the null hypothesis $H _ 0$ for using a PMCC?
When doing a hypothesis test for a PMCC/correlation coefficient, what letter do you normally use and why?
$\rho$ because you’re thinking about the whole population.
Where can you find the critical values for a PMCC/correlation coefficient hypothesis test?
In the formula booklet.