Correlations Flashcards
What is a correlation
- Determining the extent of an association between two co-variables
What are the 2 variables called that we compare in a correlation
Co-variables
What are the words that we CANNOT use when talking about correlations
- linked
- difference
- cause
- effect
- association
- DV / IV
What are the 3 types of correlation
- positive correlation
- negative correlation
- no correlation
What is a positive correlation
- As one variable increases, the other also increases
- likewise for if one decreases
—> results on a graph should resssemble a tick
What is a negative correlation
- as one variable increases, the other decreases
What is no correlation
- The variables aren’t linked
- there is no visible pattern in data
How do you write a Null hypothesis about a correlation
There will be no significant relationship/correlation between ______ and ______
How do you write a directional alternative hypothesis about a correlation
There will be a significant positive/negative relationship/correlation between ______ and ______
How do you write a non-directional alternative hypothesis about a correlation
There will be a significant relationship/correlation between ______ and ______
What type of diagram is a correlation always displayed on
Scatter diagram
You get 4 marks in an exam for drawing a scatter diagram. What are those 4 marks for?
- Title: ‘A scatter diagram to show the relationship between ______ and ______’
- results being plotted correctly
- title (including units) on both x and y axis (e.g. reaction time (seconds))
- an appropriate scale bar
When asked ‘Outline 2 findings from the scatter diagram (2)’ what are the 2 possible findings that you can comment on
- type of relationship (Positive/negative/no relationship)
- strength of relationship (strong/weak/perfect)
- identifying outliers — results that deviates from the trend (NOTE: you get no marks for calling it an anomaly)
Draw what the different scatter diagrams look like (Perfect negative correlation —> perfect positive correlation)
What are the strengths of correlations
- Correlations allow researchers to investigate naturally occurring variables that may be unethical or impractical to test experimentally
- allows researchers to clearly see if there is a relationship between variables
- allows researchers to work out if further investigations is necessary, so could be used as a pilot study to avoid a costly experiment. If there is only a weak relationship found then perhaps an experiment isn’t worth it
- Easy to replicate (assuming the data collection methods are standardised)
What is the weakness to correlations
- correlation cannot be taken to mean causation even if there is a strong association between 2 variables we can’t say one causes the other. There may be other intervening variables that explains the link between the co-variables
- people jump to causal conclusions and research can be misinterpreted so improvement programmes or treatments or policy changes may be incorrectly implemented
- the methods used to collect the data may lack internal/external validity
When asked an exam Q on outlining 1 weakness of a correlation method in a particular study, how should you answer it?
PEE
- Point: e.g. correlation doesn’t = causation because they weren’t manipulated and there could be intervening variables
- Example: make it relevant to the Q. E.g. there are many intervening variables between ______ and ______, such as ______
- Elaborate: This is a weakness because…..
What is a spearman’s Rho
- an inferential statistical test
- it is the only test that can analyse if there is a relationship (note: all other tests find a difference, not relationship)
- it is used for a researcher to state whether the findings are significant or not
- produces an observed value that falls between -1 and +1
What is an observed value
- sometimes referred to as the ‘correlation coefficient’ or ‘r’ or ‘rho’
- it comes from a Spearman’s Rho statistical test calculation
- it falls between -1 and +1
What is a critical value
- the value the observed value is compared to
- it is produced by a critical value table
- to be significant, the observed values has to be greater than or = to the critical value.
—> NOTE: it’s not always given to you in an exam. Although you don’t have to calculate it, you will have to find it
If the relation ship is significant, which hypothesis do we accept/reject
- accept an alternative
- reject a null
“A psychologist found that, after analysing the data with a Spearman’s Rho test, the observed value was 0.56. The critical value is 0.40.”
State whether the psychologist should accept or reject the null hypothesis + why
- The psychologist should reject the null hypothesis
- because the observed value of 0.56 is GREATER THAN the critical value of 0.40 the results ARE SIGNIFICANT
—> NOTE: if you copy these wordings, you will get full marks
“A psychologist found that, after analysing the data with a Spearman’s Rho test, the observed value was -0.78. The critical value is 0.31.”
State whether the psychologist should accept or reject the null hypothesis + why
- The psychologist should reject the null hypothesis
- because the observed value of -0.78 is GREATER THAN the critical value of 0.31 the results ARE SIGNIFICANT
—> NOTE: just ignore the “-“ sign.
