You are on page 1of 11

COMM 301: Empirical Research in Communication

Statistics: Correlation

Inferential Statistics Pearsons product moment correlation, r Spearmans rho: rs


What is it? Correlation is used to determine if there is a significant relationship between two variables. Two variables are related if knowing the value of one variable allows you to know something about the value of the other variable. More specifically, the Pearsons product moment correlation, r, analyzes if there is a linear relationship between two variables. When is it used? Type of question answered Is there a relationship? Type of data Variables: Two (2) continuous variables Measurement levels: Interval, ratio (for Pearsons r) Ordinal, interval, ratio (for Spearmans rho) What do you need to know? All the above, plus recognition of the SPSS output for the procedure, including scatterplots; linearity assumption r2 (coefficient of determination). How to report? See the specific examples that follow.

COMM 301: Empirical Research in Communication

Statistics: Correlation

Procedures in SPSS Analyze > Correlate > Bivariate Click the two variables into the box. Under Correlation Coefficients, check Pearson and Spearman. Under Test of Significance, check Two-tailed or One-tailed as appropriate. Check Flag significant correlations. Click Paste. Go to the syntax file. Highlight the appropriate section, and click . Graphs > Scatter Choose Simple. Click Define. Click one of the variables into the X Axis box. Click one of the variables into the Y Axis box. Click Paste. Go to the syntax file. Highlight the appropriate section, and click . Analyze > Descriptive Statistics > Frequencies Click the two variables into the box. Click on Statistics Check Skewness and Kurtosis. Choice: Pearson or Spearman? (cf. dont worry about the forumula) Usually, Pearson r

Spearmans rho is appropriate when either of the following two conditions are met: a. One variable is an ordinal scale and the other is an ordinal scale or higher. b. One of the distributions is markedly skewed or has an extreme kurtosis value. Check the variables skewness and kurtosis and if they are less than +/- 3, use Pearson r

Where D= Rank of X Rank of Y (i.e., a Difference between scores).

COMM 301: Empirical Research in Communication

Statistics: Correlation

Scatter plot examples

COMM 301: Empirical Research in Communication

Statistics: Correlation

Correlation assumes the linearity of relationship. If it is violated, you should not use correction. -- Example of violation (curvilinear relationship)

Correlation is very sensitive to outliers. You should check outliers before conducting correlation analyses

COMM 301: Empirical Research in Communication

Statistics: Correlation

Example 1 Research question: Is there a relationship between students existing GPA and their expectations regarding how well they will do in a course? H0: ?

H1: ?

Research design: Survey of 25 students with items measuring age, current GPA, and expectations for a communication course. Expectation is measured by a 7-point agreement question: I expect to do well for this course.

COMM 301: Empirical Research in Communication

Statistics: Correlation

Correlation
Statistics

AGE N Valid Missing Mean Std. Deviation Skewness Std. Error of Skewness Kurtosis Std. Error of Kurtosis 25 0 21.6800 1.84210 -.007 .464 -.698 .902

GPA 25 0 3.1080 .52751 .236 .464 -.965 .902

expectation for COMM301 25 0 4.8000 1.32288 -.070 .464 -.374 .902

expectation for COMM301

2 1 2.0

2.5

3.0

3.5

4.0

4.5

GPA

Correlations
Correlations

COMM 301: Empirical Research in Communication

Statistics: Correlation

expectation for COMM301 expectation Pearson 1 for COMM301 Correlation Sig. (2-tailed) . N 25 GPA Pearson .629(**) Correlation Sig. (2-tailed) .001 N 25 ** Correlation is significant at the 0.01 level (2-tailed).

GPA .629(**) .001 25 1 . 25

Correlations

expectation for COMM301 Correlation Coefficient Sig. (2-tailed) N GPA Correlation Coefficient Sig. (2-tailed) N ** Correlation is significant at the 0.01 level (2-tailed). Spearman's rho expectation for COMM301 1.000 . 25 .621(**) .001 25

GPA .621(**) .001 25 1.000 . 25

Report Analysis revealed that for the 25 students surveyed, there was a significant positive relationship between GPA and expectations (r = 0.63, p < 0.001). As GPA increases, expectation increases. The meaning of r2 (coefficient of determination) As you noticed, the analysis gives a value for r, which is the correlation coefficient. The correlation coefficient ranges from -1 to +1. When you take r and square it, you get r2. r2 is known as the coefficient of determination. This coefficient of determination refers to the amount of variance explained (or the variance two variables share with each other). It is one of the measures of effect size. In the example above, GPA explains (0.63)2 = 0.3969, or approximately 40% of the variance in expectations for Comm301.

COMM 301: Empirical Research in Communication

Statistics: Correlation

Example 2 Research question: Is there a relationship between students age and their expectations regarding how well they will do in a course? H0: ?

H1: ?

Research design: Survey of 25 students with items measuring age, current GPA, and expectations for a communication course. Expectation is measured by a 7-point agreement question: I expect to do well for this course.

COMM 301: Empirical Research in Communication

Statistics: Correlation

Output for Example 2

Correlation
Correlations

expectation for COMM301 expectation for COMM301 AGE Pearson Correlation Sig. (2-tailed) N Pearson Correlation Sig. (2-tailed) N 1 . 25 -.147 .483 25

AGE -.147 .483 25 1 . 25

Graph
8

expectation for COMM301

2 1 18

19

20

21

22

23

24

25

26

AGE

Report How would you write the report?

COMM 301: Empirical Research in Communication

Statistics: Correlation

Cautions Although correlation analysis can tell you if a significant linear relationship exists, it does not analyze other types of relationships well, such as curvilinear relationship. A significant curvilinear relationship can be missed by linear correlation analysis. So what should you do then?

10

COMM 301: Empirical Research in Communication

Statistics: Correlation

11

You might also like