Question about Cronbach's Alpha

classic Classic list List threaded Threaded
4 messages Options
Reply | Threaded
Open this post in threaded view
|

Question about Cronbach's Alpha

Kreischer,Resha M

First off, what is the correct pronunciation for Cronbach? J

 

I have utilized SPSS to analyze the reliability of survey subscales by going to Scale/Reliability Analysis/Alpha. Below is the resulting output. Is this enough analysis to determine the reliability of a survey or do I need to do more? If so, what? What is the best way to describe what the alpha value means in lay person terms?

 

 

 

                                                  Parent Survey

Subscale

Number of Items

Coefficient Alpha

Communicating

7

.843

Parenting

5

.822

Student Learning

8

.915

Volunteering

5

.814

School Decision Making

5

.862

Collaborating with Community

5

.818

School Climate

7

.818

·  Coefficient values range between 0 and 1.

 

 

 

Confidentiality Notice: This message may contain confidential information intended for the person(s) named above. If you have received this message in error, you are prohibited from disseminating or copying this message and requested to notify the sender immediately. Failure to do so may result in violations of federal and Maryland law.

 

 

Reply | Threaded
Open this post in threaded view
|

Re: Question about Cronbach's Alpha

Maguin, Eugene

I think it’s pronounced as if it were spelled Crawnbock but see what others say.

 

The table shows the internal consistency of the scale. That’s not the same as the imaginary situation where everybody rates the items and then does so immediately again but have NO memory of having rated the previous item (the metaphor of a scale).

 

It would also be interesting to see the item-total correlations and/or the ‘alpha is item deleted’ numbers but with that few of items and those alphas, I’d bet the item-totals and item-deleted alphas both will be uniformly high and have little variation.

 

The key to talking about it to a lay person is how they go about seeing what somebody, like their kid, has been doing while they were gone. They would ask multiple, more-or-less closely related questions from different points of view about a particular topic. They’re looking for consistency in the answers. Variation increases curiosity and skepticism. The formula (look this up on Wikipedia) is the ratio of the  K*rbar/(1 + (K-1) *rbar), where K is the number of items and rbar is the average correlation between scale items. Correlation measures consistency in responding. The higher the consistency, the higher the alpha value. (Note that asking more items also makes alpha increase—so long as the average correlation does not change.) So the key idea to communicate is that of consistency in responding by the test takers.

 

Gene Maguin

 

 

 

 

From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Kreischer,Resha M
Sent: Wednesday, May 30, 2012 2:02 PM
To: [hidden email]
Subject: Question about Cronbach's Alpha

 

First off, what is the correct pronunciation for Cronbach? J

 

I have utilized SPSS to analyze the reliability of survey subscales by going to Scale/Reliability Analysis/Alpha. Below is the resulting output. Is this enough analysis to determine the reliability of a survey or do I need to do more? If so, what? What is the best way to describe what the alpha value means in lay person terms?

 

 

 

                                                  Parent Survey

Subscale

Number of Items

Coefficient Alpha

Communicating

7

.843

Parenting

5

.822

Student Learning

8

.915

Volunteering

5

.814

School Decision Making

5

.862

Collaborating with Community

5

.818

School Climate

7

.818

·  Coefficient values range between 0 and 1.

 

 

 

Confidentiality Notice: This message may contain confidential information intended for the person(s) named above. If you have received this message in error, you are prohibited from disseminating or copying this message and requested to notify the sender immediately. Failure to do so may result in violations of federal and Maryland law.

 

 

Reply | Threaded
Open this post in threaded view
|

Re: Question about Cronbach's Alpha

Bruce Weaver
Administrator
I'll give this a try and see if the special characters come through okay.

The way I have usually heard it and say it:   krōn-bǎk
The way it probably ought to be pronounced:  krōn-bäKH

Pronunciation Key:  http://www.thefreedictionary.com/_/pk.htm 

In case the special characters don't come through:

The way I have usually heard it and say it:   crone-back
The way it probably ought to be pronounced:  crone-Bach (with the "ch" pronounced as in the German word "ich")


Maguin, Eugene wrote
I think it's pronounced as if it were spelled Crawnbock but see what others say.
--
Bruce Weaver
bweaver@lakeheadu.ca
http://sites.google.com/a/lakeheadu.ca/bweaver/

"When all else fails, RTFM."

NOTE: My Hotmail account is not monitored regularly.
To send me an e-mail, please use the address shown above.
Reply | Threaded
Open this post in threaded view
|

Re: Question about Cronbach's Alpha

Beckstead, Jason
In reply to this post by Kreischer,Resha M

Cronbach's coefficient is a form of the intraclass correlation coefficient (ICC). It is basically telling you the proportion of variance in your two-way table of data (subjects x items) that is due to individual differences among the subjects. Although a form of alpha can be calculated from a correlation matrix, in its original form it is calculated from a covariance matrix.

 

So, if you have k=10 items and a bunch of subjects who have answered all items, first we calculate a covariance matrix among the 10 items, which has the item variances in its diagonal. To get alpha, first sum of the diagonal elements (variances) and divide this sum by the sum of all the elements in the matrix. Second, this ratio is subtracted from 1.0. Third, multiple by a weighting factor of k over k-1 or 10/9.

 

hth.

 

_____________________________________________________________

 Jason W. Beckstead, Ph.D.       

  Associate Professor/Quantitative Methodologist

  University of South Florida College of Nursing

  12901 Bruce B. Downs Blvd., MDC22, Tampa, FL 33612, USA

  Statistical Editor, International Journal of Nursing Studies

  phone: +1.813.974.7667  fax: +1.813.974.5418          

  personal website:  http://personal.health.usf.edu/jbeckste/

  International Journal of Nursing Studies  http://www.elsevier.com/ijns

 

 

 

From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Kreischer,Resha M
Sent: Wednesday, May 30, 2012 2:02 PM
To: [hidden email]
Subject: Question about Cronbach's Alpha

 

First off, what is the correct pronunciation for Cronbach? J

 

I have utilized SPSS to analyze the reliability of survey subscales by going to Scale/Reliability Analysis/Alpha. Below is the resulting output. Is this enough analysis to determine the reliability of a survey or do I need to do more? If so, what? What is the best way to describe what the alpha value means in lay person terms?

 

 

 

                                                  Parent Survey

Subscale

Number of Items

Coefficient Alpha

Communicating

7

.843

Parenting

5

.822

Student Learning

8

.915

Volunteering

5

.814

School Decision Making

5

.862

Collaborating with Community

5

.818

School Climate

7

.818

·  Coefficient values range between 0 and 1.

 

 

 

Confidentiality Notice: This message may contain confidential information intended for the person(s) named above. If you have received this message in error, you are prohibited from disseminating or copying this message and requested to notify the sender immediately. Failure to do so may result in violations of federal and Maryland law.