PSYC 6430: Personal Data Set Assignment Two

Bring your personal data set into SAS and execute this command:

ProcCorr; Run;(or the SPSS or R equivalent)

Yes, you obtained this same correlation matrix in the PDS1 assignment, but now I want you interpret it.

If your data file has more than four variables, use a VAR statement to specify the four desired variables – for example, Proc Corr; Var Sex Group Weight Height; Run;

The output will include a correlation matrix, like that shown below. In each cell of this matrix there appears the correlation coefficient and, below that, the p value. Copy the correlation matrix into a Word document and then, in the upper right portion of the matrix, above the main diagonal of ones, highlight each cellfor which the correlation is significant, that is, p .05. Then prepare a bulleted list in which you interpret each significant correlation, as I have done below. I used the Add.dat data set described on pages 686 & 687 of Howell. If none of your correlations is significant, just write “My data are exceptionally boring.”

Pearson Correlation Coefficients, N = 88
Prob > |r| under H0: Rho=0
socprob / dropout / addsc / iq
socprob
socprob
/ 1.00000
/ 0.43590
<.0001
/ 0.37100
0.0004
/ -0.26516
0.0125
dropout
dropout
/ 0.43590
<.0001
/ 1.00000
/ 0.37969
0.0003
/ -0.30122
0.0043
addsc
addsc
/ 0.37100
0.0004
/ 0.37969
0.0003
/ 1.00000
/ -0.63163
<.0001
iq
iq
/ -0.26516
0.0125
/ -0.30122
0.0043
/ -0.63163
<.0001
/ 1.00000
  • Students with social problems were significantly more likely to dropout of school than were students without social problems,  = .436, p < .001.
  • Having social problems and dropping out of school were coded 0 = no, 1 = yes. Accordingly, the observed positive correlation indicates that those with social problems were more likely to drop out of school.
  • Imagine other dichotomous variables, Gun (coded 0 = do not own a gun, 1 = do own a gun), and Gender (coded 1 = male, 2 = female), with  = -.436. The negative correlation would indicated that being female is negatively associated with owning a gun – that is, men male participants were more likely to own a gun than were female participants.
  • Students with social problems scored significantly higher on the attention deficit battery than did those without social problems, rpb = .371, p < .001.
  • Students with social problems had significantly lower IQ than did those without social problems, rpb = -.265, p = .012.
  • Students who subsequently dropped out of school scored significantly higher on the attention deficit battery than did those who graduated, rpb = .380, p.001.
  • Students who subsequently dropped out of school had significantly lower IQ than did those who graduated, rpb = -.301, p = .004.
  • Students’ scores on the attention deficit battery were significantly, negatively correlated with IQ, r = -.632, p < .001.

When the Pearson r is computed between two dichotomous variables, it is called phi (). Testing its significance the way Proc Corr does it is not quite equivalent to conducting a contingency table analysis with the Person chi-square (2), but we have not yet covered the chi-square statistic.

When the Pearson r is computed between a dichotomous variable and a continuous variable, it is called a point-biserial correlation coefficient (rpb). Testing the significance of a point-biserial correlation coefficient is absolutely equivalent to conducting an independent samplest test.

If one of your variables is categorical with more than two variables, (for example, Ethnicity, coded 1 = Caucasian, 2 = African American, 3 = Hispanic), you cannot interpret its simple correlation with any other variable without making assumptions that are probably unreasonable.

When you report the significant results of a bivariate analysis, you should always emphasize the direction of the significant difference or correlation.

Do not assume that the readers already know what your variables know. Refer to them in descriptive terms, for example, “attention deficit battery” rather than “addsc.”

Likewise, do not assume that the readers already know who or what your subjects are. Describe them in your summary statement. For example, “students” rather than “subjects.”

Before writing your summary statements, read my Thesis Errors document. Do your best not to make any of these errors in documents you present to me. If I detect such errors, my evaluation of your work will be affected.

Create a new thread in the PDS2 Forum in BlackBoard, with your last name as the name of the thread. Enter a very brief description of the most interesting association you found in your data and then attach the Word document. This assignment is due at 5 PM on Thursday of this week.

Professor Karl at the Beach, Cape Cod, right after Woodstock, 1969 (aka “The Mud Hole”):