```Date: Thu, 6 Feb 2003 11:09:14 -0500 Reply-To: "Snider-Lotz, Tom" Sender: "SPSSX(r) Discussion" From: "Snider-Lotz, Tom" Subject: Re: Best way to summarize 3 variables? Content-Type: text/plain; charset="iso-8859-1" I generally agree with Carol, except that .60 may be a little steep for a 3-item test. I'd think anything from .20 on up should be OK. Another way to look at the relationships among the items would be a correlation matrix. If all the correlations are positive, then they all are in the same direction. Given the item content, I'd expect the correlations to be positive and pretty high. How many teachers provided data? >-----Original Message----- >From: Carol Albright [mailto:calbright@visi.com] >Sent: Wednesday, February 05, 2003 9:17 PM >To: SPSSX-L@LISTSERV.UGA.EDU >Subject: Re: Best way to summarize 3 variables? > > >Hi, Jessica & Listers, > > Research frequently summarizes groups of Likert-scale >items (i.e. 1-5, >strongly disagree-strongly agree) as a scale, by taking the mean of >non-missing answers. All 3 of your items are in the same >direction, that >is, the higher score indicates more satisfaction with teaching, so you >don't need to reverse code anything. > > Since you only have 3 items, I'd require that all 3 >items be non-missing >in order to compute the scale: > >** Step 1: Check for out-of-range or weird data. >freq tatt1 tatt2 tatt3. > >** Step 2: Assuming data didn't need fixing, compute the scale. >compute tchsatis = mean.3(tatt1, tatt2, tatt3). >var labels tchsatis "Satisfaction with Teaching Scale (mean of >tatt1, tatt2 >& tatt3)". >freq tchsatis/statistics = mean median stddev skewness. > >** Step 3: Check Cronbach's Alpha to make sure the 3 items really hang >together. >RELIABILITY /VARIABLES=tatt1 tatt2 tatt3 > /FORMAT=NOLABELS /SCALE(ALPHA)=ALL/MODEL=ALPHA /SUMMARY=TOTAL . > > >If the alpha is pretty wimpy (uh, .6 or below?), you might >have to leave >them as separate items. > >HTH, >Carol > > > > >At 05:12 PM 2/5/03 -0500, Jessica L. Kenty wrote: >>Hi all, >> >>Here is a basic question for you all. >> >>I have 3 variables that address the same issue. I would like >to summarize >>these variables into 1 variable. I need to be sure first >that they go in >>the same direction & if so, summarize them. What do you all >recommend for >>steps? >> >>They 3 variables & the coding scheme are as follows: >> >>tatt1 Teacher Enjoys Present Teaching Job >>tatt2 Teacher Feels s/he makes a Difference in Children's Lives >>tatt3 Teacher Would Choose Teaching Again >> >>1 Strongly Disagree >>2 Disagree >>3 Neither agree nor disagree >>4 Agree >>5 Strongly Agree >> >> >>Thanks in advance for your insight! >> >>Jessica >> >--------------------------------------------------------------- >---------- >Carol L. Albright, MS | E-Mail : calbright@visi.com >Albright Consulting | Phone : 651/699-7218 >St. Paul, MN 55105 USA | Research data services >http://www.tc.umn.edu/~syzygy >--------------------------------------------------------------- >---------- > ```

Back to: Top of message | Previous page | Main SPSSX-L page