Correlation (research method) Flashcards
Describe Correlation
- to see if there is a relationship between 2 unrelated variables that don’t have a link
- mostly use quantitative data through questionnaire and observations
Define co-variables
- imagine our variables are age (X) and beauty (Y) then the co-variables jobs is to determine the X and Y position of each dot
- scatter graph are most commonly used
- gathered personally by primary data (have to do it themselves) or secondary data (already completed investigations)
Define positive, negative and no correlation
positive = both variables increase and decrease together
negative = 1 variable goes down while the other goes up (vice versa)
No = cant go up or down (the line) as its dotted randomly/everywhere
Define coefficiency and state the measurement uses to understand it
*can add examples for the measurement if need further knowledge to understand it
its simply states the strength and types of agreement between the 2 variables
\+1 = perfect positive correlation e.g. height and weight -1 = perfect negative correlation e.g. driving speed and miles per gallon 0 = no relationship at all e.g. meteor falling when your playing games
0.1-0.3 = very weak/weak positive/negative correlations
0.5 = moderate positive/negative correlations
0.8 = strong positive/negative correlations
1 = perfect positive/negative correlations
0 = no correlations
when using stat test what is the coefficient that determine the significance of the relationship
Spearman’s rho = ordinal data correlations
Pearson = interval data correlations
what is the difference between correlation and experiments
Correlation identifies variables and examines if a relationship is positive, negative or non-existent between said variables.
On the other hand, experiments test the effects of the IV has on the DV
Define curvilinear correlation
obviously not the same as linear relationships but can still have predictable relationship
example:
anxiety and performance do not have a linear relationship. Performing an action can be lowered or heightened if anxiety is too high or too low however its stable when its moderate (middle)
What are the rules when making a hypothesis in correlation
- NEVER PUT DIFFERENCE IN THE HYPOTHESIS (FOR CORRELATIONS)
* needs to be operationalized (how is it measured)
How would you write a 1 tailed directional hypothesis
*use random examples if the answer is shown and diff from yours but operationalized it and gave the right correlation then that fine
The use of phones use measured in hours and exam performance measured in grade/score of the test are positively correlated.
How would you write a 2 tailed directional hypothesis
*use random examples if the answer is shown and diff from yours but operationalized it and gave the right correlation then that fine
there will be a significant correlation between sleep on hours and aggression scores on a questionnaire out of 10
How would you write a null hypothesis
*use random examples if the answer is shown and diff from yours but operationalized it and gave the right correlation then that fine
The size of one’s finger measured in cm and their speed measured in mph has no correlation (relationship)
Strengths
They allow us to investigate otherwise unethical situations by using existing data to see if its reliable such as if smoking causes lung cancer. This is favourable as we can now observer the results on natural occurring variables and see if the results can lead to new revolutionary research.
They control for participant variables. As the measurement comes from the same person therefore it increases the internal validity of the test and can be used to make proper predication if and that there isn’t a participant variable in the first place
They can lead to new research as it compares 2 separate variables to see whether they have any correlation at all. This can open the worlds understanding on humans more as the discover new data that can be seriously beneficial to society an example of this are schizophrenia and dopamine.
Its reliable as its there for any to use and double check themselves
Graphs and correlation are really easy to understand its function and process (its standardised procedure) as most are normally taught this in their experience of secondary school. Therefore, it’s become standardized hence can be easily replicated and compared to other individuals’ findings.
Objective data analysis as its either positive, negative or no correlation which uses quantitative data. This can remove most bias from the data as hard proof stats cant be manipulated as it won’t work.
Correlations are really easy and can be easily done hence its can be identified if its worth furthering the research in addition you can use quantitative data of secondary data to quicken the time taken and if a correlation if found then you can further investigate why there is a correlation between the 2 variables
Weaknesses
They cannot infer cause and effect as it does not mention the reason for what happened and the cause for it. Therefore, won’t be able to fully understand the data so you’re unable to make accurate predictions leading to validity decreasing. In addition to this it could be a 3rd factor (confounding variable) which could have increased in the co-variable
3rd factor could have caused the increase in the co-variable. This is a confounding variable that was failed to be taken into consideration before hand therefore is loses its credibility and its accuracy of the study
There are problems with using secondary data. This will lead to inaccurate and invalid results as it wasn’t being administrated by oneself. This can be severely detrimental to society in addition to it using its credibility. An example of this that its purpose isn’t for always research like attendance aren’t made to predict achievements.
Correlation can be misused as an implication to society as its hard to make a solid conclusion.
