SPSS Probabilities or Correlation of Yes-No

[ncanfield]

Hello,

 

I am needing some help in which direction to go with a dataset (n=58) of 17 variables with possible responses yes(1), no(0), and missing(666).  The dataset is a screening form of the life history of individuals incarcerated locally.  I’d like to know the probabilities/correlation (%) of a response yes to a question correlating a yes response on another question.  Is it a simple bivariate correlation command?  The one I ran came up with several significant correlations from the Pearson Correlation Coefficients, but then how do I determine correlation of a yes instead of no?  I may be over thinking this or missing something. 

 

Thanks,

 

Nathan R. Canfield
Evaluation Senior Associate
Region 3 Behavioral Health Services
P.O. Box 2555
Kearney, NE 68848-2555
E-mail: [hidden email]
Phone: [phone] ext. 231
Fax: [phone]
Web: www.region3.net

 

---

CONFIDENTIALITY NOTICE: This communication and any files or attachments transmitted with it may contain information that is confidential, privileged and exempt from disclosure under applicable law. It is intended solely for the use of the intended recipient. If you are not the intended recipient, you are hereby notified that any unauthorized review, use, disclosure, dissemination, or copying of this communication is strictly prohibited. If you have received this communication in error, please notify the sender by reply e-mail and destroy all copies of the original message. Additionally, we will take the appropriate action to avoid sending you an unintended e-mail in the future. Thank you for your cooperation.

[Bruce Weaver]

Nathan Canfield wrote

Hello,

I am needing some help in which direction to go with a dataset (n=58) of 17 variables with possible responses yes(1), no(0), and missing(666).  The dataset is a screening form of the life history of individuals incarcerated locally.  I'd like to know the probabilities/correlation (%) of a response yes to a question correlating a yes response on another question.  Is it a simple bivariate correlation command?  The one I ran came up with several significant correlations from the Pearson Correlation Coefficients, but then how do I determine correlation of a yes instead of no?  I may be over thinking this or missing something.

Thanks,

You say you'd like to know "the probabilities/correlation (%) of a response yes to a question correlating a yes response on another question."  Later, you ask, "how do I determine correlation of a yes instead of no?"

This leaves me quite uncertain as to what you really want.  If you want the correlation between two questions, then there is only one correlation, not separate correlations for Yes and No.  

But I wonder if what you're really trying to ask for is:

1) the proportion (or %) who responded Yes to item 2 given that they responded Yes to item 1; and
2) the proportion (or %) who responded Yes to item 2 given that they responded No to item 1.

If you want the latter, use CROSSTABS, and ask for the row percentages (or column percentages, depending on how you've entered the variables).

If it really is the correlation you want, it is fine to compute Pearson's r computed on two dichotomous variables--this is equivalent to what used to be called the "phi coefficient"; and phi-squared = Pearson's chi-square computed on the same pair of dichotomous variables.  

[Arthur Burke]

Nathan ... Look at the McNemar option on CrossTabs.

 

The following may be useful ...

 

http://faculty.chass.ncsu.edu/garson/PA765/mcnemar.htm#mcnemar

Art

 

---

From: Nathan Canfield [mailto:[hidden email]]
Sent: Thursday, June 25, 2009 9:55 AM
To: [hidden email]
Subject: SPSS Probabilities or Correlation of Yes-No

Hello,

 

I am needing some help in which direction to go with a dataset (n=58) of 17 variables with possible responses yes(1), no(0), and missing(666).  The dataset is a screening form of the life history of individuals incarcerated locally.  I’d like to know the probabilities/correlation (%) of a response yes to a question correlating a yes response on another question.  Is it a simple bivariate correlation command?  The one I ran came up with several significant correlations from the Pearson Correlation Coefficients, but then how do I determine correlation of a yes instead of no?  I may be over thinking this or missing something. 

 

Thanks,

 

Nathan R. Canfield
Evaluation Senior Associate
Region 3 Behavioral Health Services
P.O. Box 2555
Kearney, NE 68848-2555
E-mail: [hidden email]
Phone: [phone] ext. 231
Fax: [phone]
Web: www.region3.net

 

---

CONFIDENTIALITY NOTICE: This communication and any files or attachments transmitted with it may contain information that is confidential, privileged and exempt from disclosure under applicable law. It is intended solely for the use of the intended recipient. If you are not the intended recipient, you are hereby notified that any unauthorized review, use, disclosure, dissemination, or copying of this communication is strictly prohibited. If you have received this communication in error, please notify the sender by reply e-mail and destroy all copies of the original message. Additionally, we will take the appropriate action to avoid sending you an unintended e-mail in the future. Thank you for your cooperation.

[Art Kendall]

I am not sure what you are asking.

Cases with either variable missing will not be used in the correlation calculation.
If you have a positive correlation there is a tendency for both variables to have the same value. When the first is yes the other is yes and vice versa. The two variables tend to be high (1) together and low (0)  together.
If you have a negative correlation there is a tendency for the 2 variables to have different values. When the first is yes the other is no and vice versa. When one variable is high the other is low.

Art Kendall
Social Research Consultants

Nathan Canfield wrote:

 

I am needing some help in which direction to go with a dataset (n=58) of 17 variables with possible responses yes(1), no(0), and missing(666).  The dataset is a screening form of the life history of individuals incarcerated locally.  I’d like to know the probabilities/correlation (%) of a response yes to a question correlating a yes response on another question.  Is it a simple bivariate correlation command?  The one I ran came up with several significant correlations from the Pearson Correlation Coefficients, but then how do I determine correlation of a yes instead of no?  I may be over thinking this or missing something. 

 

Thanks,

 

Nathan R. Canfield
Evaluation Senior Associate
Region 3 Behavioral Health Services
P.O. Box 2555
Kearney, NE 68848-2555
E-mail: [hidden email]
Phone: [phone] ext. 231
Fax: [phone]
Web: www.region3.net

 

---

CONFIDENTIALITY NOTICE: This communication and any files or attachments transmitted with it may contain information that is confidential, privileged and exempt from disclosure under applicable law. It is intended solely for the use of the intended recipient. If you are not the intended recipient, you are hereby notified that any unauthorized review, use, disclosure, dissemination, or copying of this communication is strictly prohibited. If you have received this communication in error, please notify the sender by reply e-mail and destroy all copies of the original message. Additionally, we will take the appropriate action to avoid sending you an unintended e-mail in the future. Thank you for your cooperation.

===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD