Correlations Flashcards
what is correlational anaylsis
determine whether there is a significant linear relationship between 2 variables
what is a positive correlational analysis
as one variable increases so does the other
what is negative correlational analysis
as one variable increases, the other decreases
when you have 2 conditions, if the data is parametric, what correlation do you use
pearson
when you have 2 conditions, if the data is non-parametric, what correlation do you use
spearman
what are the 5 things you look out for with data
are there any obvious patterns is the pattern what you predicted do scores vary a lot are there any anomalies are scored clustered/spread out
what are 2 positives of using a correlation
shows a relationship
can make predictions from them
what is one negative of using a correlation
cannot draw conclusions
what 3 things does data have to have to make it parametric
interval/ratio scale
equality of homogeneity of variance
normal distribution
what are 3 positives of having parametric data
more powerful test
more likely to demonstrate of significant result
depends on parametric assumptions
what 3 things does data have to have to make it non-parametric
ordinal/nominal data
any variance
any distribution
what is a positive of having non parametric data
more tolerant of outsiders
what is a negative of having non parametric data
less powerful
what is a correlation coefficient
tells us the strength and direction of the relationship
during a correlation coefficient, what does the numbers go from
-1—–+1
in a correlation coefficient what does -1 mean
perfect negative corelation
in a correlation coefficient what does +1 mean
perfect positive corrlation
in a correlation coefficient what does 0 mean
no relationship
what can r tell us about the correlation?
the direction of the relationship-showed by - and +
strength of the relationship
what is a one tailed hypothesis
predicting the direction of the relationship
what is a 2 tailed hypothesis
predicting a relationship with no direction
what is the level of significance we accept
5%
0.05
what degress of freedom do we do
N-2
if r is larger than the critical value, then what do we do
the result is significant at p<0.5
we can be 95% certain the result did not happen by chance
we accept the experimental hypothesis and reject the null hypothesis
if r is smaller than the critical value, then what do we do
the result is non-significant
we are not certain enough that the result didnt happen by chance
we have to reject the experimental hypothesis and accept the null hypothesis
how do we write the formula for a significant result
r(degrees of freedom) =r value, p<0.05