RM5- Correlations Flashcards
what is a correlation?
a form of analysis to see the extent to which two different variables are related
what is measured with correlations?
two covariables
-neither are controlled or manipulated by the researcher
where must both variables being measured come from?
the same participant
what are the two types of correlations?
positive and negative
what is a positive correlation?
as variable 1 increases, variable 2 increases
what is a negative correlation?
as variable 1 increases, variable 2 decreases
how are correlations presented?
using a scatter graph
how does a scatter graph present correlations?
has one covariable on the x-axis and the other on the y-axis
a straight line that best fits the points is plotted
what are some types of correlations
(strength)
perfect, strong, weak
no correlation
what do correlation coefficients measure?
the strength of a linear association between two covariables, producing a value between +1 and -1
what is R= +1
perfect line for a positive correlation
what is R= -1
perfect line for negative correlation
R= 0
no correlation between the variables
what must a hypothesis include?
both variables that the correlation is testing and the word significant
directional hypothesis
they will say if the relationship will be positive or negative
non-directional hypothesis
they will say the relationship between two covariables (difference)
AO3
what do correlations allow us to investigate?
otherwise unethical situations
AO3
what can correlations lead to….
new research, can be used as a starting point before committing to an experimental study
-carried out with relative ease
-therefore decide if new research should be carried out
AO3
how do correlations control participant variables?
participant variables
-both sets of data come from the same participant, no individual differences
-high validity
AO3
do correlations infer causation (cannot establish cause and effect)?
all correlations tell us whether a relationship exists between two co-variables
-they cannot tell us if one variable causes the other, its usefulness is therefore limited
AO3
what validity issues do correlations have?
an untested variable could be impacting upon the relationship
(third variable problem)
-this draws inaccurate conclusions about the existing relationship