Retrieve and Open the Data 1. To download the data, click on the link on the class website for the SPSS syntax file for lab 1. 2. Open the file that you downloaded. 3. In the SPSS Syntax Editor, click on Run All (that means to click on Run in the menu, and then All on the drop down menu). Analyze the Data: Interobserver Reliabilities 4. Each person was observed by two observers. To get the interobserver reliability coefficient, we will correlate the first observer s observation with the second observer s observation. 5. Because these variables are nominally scaled, Cramer s V is the correct measure of association to use (Pearson s r only works with interval and ratio scaled variables) 6. In SPSS, click Analyze Descriptive Statistics Crosstabs 7. Drag the first observer s observation of the person s sex from the box on the left to the Row(s) box on the right 8. Drag the second observer s observation of the person s sex from the box on the left to the Column(s) box on the right 9. Click the Statistics button
10. Check the box to the left of Phi and Cramer s V 11. Click Continue 12. Click OK 13. The output should appear in the SPSS output viewer: The correlation coefficient equals 1.00. The p value (Approx. Sig.) is.000. Because the p value is less than the standard α level (.05), we conclude that these two variables are likely correlated with each other. 14. Repeat the previous steps and calculate the interobserver reliability coefficients for the book carrying styles.
Collapsing Book Carrying Styles to Book Carrying Types 15. We want to change all observations of book carrying Styles A and B into book carrying Type I. Likewise, we want to change Styles C, D and E into Type 2. Since we have no predictions about the other category, we want it to go away. 16. In SPSS, click on Transform Recode Into Different Variables 17. Drag the carrying style as reported by observer 1 variable from the box on the left to the Input Variable -> Output Variable box on the right: 18. In the Output Variable, Name box type a name of the new variable. E.g. booktype
19. Optionally add a label to the output variable. E.g. Book Carrying Type 20. Click Change
21. Click Old and New Values 22. In the Old Values, click in the circle to the left of Range and enter 1 and 2 in the two boxes beneath it. Style A corresponds to the value 1, while Style B corresponds to the value 2 23. In the New Value, Value Box, type 1. This converts Styles A (old value 1) and B (old value 2) into Type 1 (new value 1)
24. Click Add
25. Repeat for Styles C (old value 3), D (old value 4), and E (old value 5) being changed into Type 2 (new value 2) 26. All other values should go away. In the Old Values, click in the circle to the left of All other values 27. In the New Value, click in the circle to the left of System-missing (this tells SPSS to not use these values)
28. Click Add
29. Click Continue 30. Click OK. In the SPSS data editor, there should be a new variable at the right edge of the previous data that corresponds to the variable you just created. Check that it was created correctly. You may need to switch to the Data View to see the data. To do so, click on View Data. Descriptive Statistics and the χ 2 Test of Independence 31. Because all of our data are nominally scaled, and we want to know whether the book carrying type depends on the person s sex, the χ 2 test of independence is the appropriate inferential statistic to use. It tests the null hypothesis that the two variables (sex and book carrying type) are independent of each other (knowing one tells us nothing about the other). H 0 : The two variables are independent of each other H 1 : The two variables are not independent of each other. 32. In SPSS click Analyze Descriptive Statistics Crosstabs 33. Click the Reset button 34. Drag one of the variables (e.g. book carrying type) into the Row(s) box 35. Drag the other variable (e.g. sex of person as reported by observer 1) into the Column(s) box
36. Click the Statistics button 37. We need the χ 2 test, so check the box to the left of Chi-square. 38. APA style mandates that we report an effect size with the Chi-square test. Cramer s V is the appropriate effect size for a χ 2 test. Check the box to the left of Phi and Cramer s V 39. Click Continue 40. Click OK 41. The output appears in the output viewer 42. This tells us that there were 102 females observed using Type 1 to carry books while there were only 12 males using Type 1 to carry books. Conversely, there were 28 females observed using
Type 2 to carry books by there were 143 males using Type 2 to carry books. 43. The value of χ 2 with one degree of freedom is 147.333. The p value (assymp. sig.) is.000. Because the p value is less than the standard α level, we reject the null hypothesis that the two variables are independent. That is, it is likely the case that knowing a person s sex tells us what book carrying type they will use. 44. Cramer s V equals.719. For one degree of freedom, this is considered a large effect (Small effect:.1 V <.3; Medium effect:.3 V <.5; Large effect:.5 V 45. In APA style we would write: A χ 2 test of independence revealed that the person s sex and book carrying style are likely dependent, χ 2 (1) = 147.333, Cramer s V =.719, p =.000, α =.05.