methods: analysing correlational data

Question 1

how do you start analysing a correlational study?

Answer 1

• look for a relationship between two variables
• begin by drawing a scatter diagram to spot any link
• if a link appears, use an inferential test to check significance

Question 2

which statistical test is best for ordinal data in a correlation study?

Answer 2

• spearman’s test is most appropriate for ordinal data in a correlation study

Question 3

what is spearman’s test used for in correlation studies?

Answer 3

• it’s an inferential test to see if results can be generalised to the population
• only used for correlational data
• works with ordinal data or data that can be reduced to ordinal

Question 4

what are the steps in calculating spearman’s rho?

Answer 4

• rank the scores for each of the two variables
• find the difference between the ranks for each pair of scores
• subtract the second rank from the first and record the result
• square the differences to remove negative values
• calculate the total of the squared differences

Question 5

what does the correlation coefficient in spearman’s rho tell us once r is calculated?

Answer 5

• shows how closely the two variables are related
• ranges from +1 to -1
• the closer to +1 or -1, the stronger the relationship

Question 6

what’s an ‘eyeball’ test?

Answer 6

• looking at the data sets to see how they compare is useful, but it’s difficult to draw firm conclusions from
• spearman test needs to be carried out to see if relationship is real for the group

Question 7

how are the results interpreted from the statistical test?

Answer 7

• if both scores rise, there’s a positive correlation. a result of +1 means a perfect positive correlation. there’s a perfect relationship between the two scores
• a result of 0 means that there’s no correlation. the scores don’t relate to each other
• if one score rises and the other falls there’s a negative correlation. there is a perfect relationship between the two scores

Question 8

what does statistical significance mean in research?

Answer 8

• shows if results are likely due to a real relationship, not just chance
• psychologists use a significance level of 5% (p ≤ 0.05)
• means there’s a 95% chance the results reflect a true relationship
• the result from the statistical test is called the observed value

Question 9

how do you know if the result from spearman’s rho is significant?

Answer 9

• compare the observed value to the critical value in the table
• the observed value must exceed the critical value for significance
• critical values are set by statisticians to help judge if results meet the criteria for significance

Question 10

what should you be cautious of when interpreting spearman’s rho results?

Answer 10

• large samples can show significance even with weak correlations
• always check for outliers or extreme scores
• outliers can affect the correlation and the significance of results