Correlation Flashcards
Words not to use in correlation studies
Difference Effect Affect Change IV/DV Experimental design
Research Aim and Questions
- What is a research question?
- What is a research aim?
A research question is a question related to what is being investigated, the statement must end with a ‘?’
A research aim is a statement about the purpose of the study
What kind of descriptive statistics is used for correlation?
Scatter graphs are used o plot findings for correlation graphs
A correlation shows the relationship as well as the strength and direction of the relationship
Types of correlation
What happens to co-variables for each type of correlation?
Positive correlation
One co-variable increases = the other co-variable increases
Negative correlation
One co-variable increases = the other co-variable decreases
No correlation
No trend or relationship between the two co-variables
Key points
how many sets of information needs to be collected from each participant? Why?
What type of data needs to be collected?
What level of data needs to be collected?What type of questions need to be avioded? why? (what type of data will they give)
For a correlation study, two sets of information needs to be collected from each participant as it will then allow the researcher to compare the collected information to see if there is a relationship
The data collected has to be at least ordinal level
Types of questions to avoid
- Open questions = qualitative data
- Closed questions = nominal data
What does operationalise mean?
Operational variables allows researchers to ensure that the things being investigated are testable and measurable.
What is the difference between experiments and correlations?
Experiments VS Correlation IV/DV VS Co-variables Control and experimental groups affect or difference Any level of data Bar charts or Pie charts
Key points
How many sets of information needs to be collected from each participant? Why?
What type of data needs to be collected?
What level of data needs to be collected?What type of questions need to be avoided? Why? (what type of data will they give)
For a correlation study, two sets of information needs to be collected from each participant as it will then allow the researcher to compare the collected information to see if there is a relationship
The data collected has to be at least ordinal level
Types of questions to avoid
- Open questions = qualitative data
- Closed questions = nominal data
Correlational hypothesis
Do correlation studies have alternate and null hypotheses?
what is the difference between experimental and correlational hypotheses?
As with experiments, correlational studies also have alternate and null hypotheses however there is a difference.
differences
co-variables not IV/DV
Relationship not difference
Points to remember
What can not be inferred from a correlation?
What needs to be carried out to find the relationship between co-variables?
What is the difference between findings and conclusions?
Correlations can not show cause and effect it only shows the relationship
Statistical tests are required to determine the correlation, a scatter graph alone will not signify a relationship
Findings refer to the data collected e.g. the mean, median or mode
Conclusions refer to what the findings mean e.g. is there a positive, negative or no correlation between the two co-variables, how strong is the relationship
How to know if there is a true correlation
What needs to be carried out to find if there is a real correlation?
what will this provide?
What does this value provide?
To see if there is a real correlation, inferentail statsictal tests like Spearmen Rho need to be carried out
These tests will provide a correlational coeffiecient (a number value for the relationship) that will indicate how strong the relationship is between the two co-variables and the relationship
Correlation coefficient Range? What is the correlation coefficient for: Positive correlation negative correlation No correlation? What makes a correlation coefficient stronger?
Ranges from -1 through 0 to +1
Perfect negative correlation has a correlation coefficient of -1
Perfect positive correlation has a coefficient of +1
No correlation = coefficient of 0
The further away the coefficient is from 0, the stronger the correlation will be