Question

Now please fill your homework to a blank A4 paper. Do not forget to write

1- What is a correlation in statistics?

2- What is the purpose of correlation?

3- What are differences between Spearman and Pearson correlation?

4- Why pearson and Spearmen correlation can differ?

5- What do you mean by rank correlation?

6a- What are the assumptions of pearson’s correlation coefficient itself?

6b- What are the assumptions of statistical significance-test for correlations?

(see the file PSY214-03.ppt)

7- What is homoscedasticity assumption? Please explain with a simple drawing.

8- Please check correlations between (I_V and D_V1, I_V and D_V2, D_V1 and D_V2) and report in APA style.

9- With your results (by cleaning outliers and not cleaning) explain the effect of outliers on Pearson, Spearman and Kendal's coefficients.

10-Please write SPSS commands and option selections for correlation analysis.

If you have problem with data cleaning please use the data sets in drive noted “outliers cleaned”

Answer 1

1. Correlation in statistics is an assessment that measures the quantity of the direction and the strength, of two variables that we analyze, to fluctuate together.

2. The purpose of correlation is to find a relationship between two variables so that one can predict how they will behave in the future. Correlation coefficient helps in putting a value to the relationship between two variables. This prediction is very important in the fields like health care, government sector and also in business. In business it helps in forming business plans and budgets.

3. Pearson's correlation coefficient is used when the variables under consideration are linearly related ( that is they move together at a constant rate and we can graphically represent them as a straight line) whereas spearman correlation coefficient can be used with ordinal, interval or ratio variables. Pearson's correlation coefficient depends on the values of the variables but superman's is based on tha rank that we assign to each variable. Pearson's do not work well when there are outliers in the data but Spearman can be used in this case as well.

4. Pearson's coefficient and Spearman's coefficient, both measures the relationship between the variables. But they are not the same. Pearson's measures the linear relationship between the variables and finds how well a straight line can describe the relationship whereas spearman's measures the rank order of two variables and it does not worry about where exactly these points are.