Re: Negative Cronbach's Alpha

[Art Kendall-2]

I should have mentioned that the reversed items are not contiguous with
the others in the scale.
rename variables (CI_4 CI_5 CI_6=
   olditem4 olditem5 olditem6 ).
recode olditem4 olditem5 olditem6
(1=5) (2=4) (3=3) (4=2) (5=1) (else =copy)
into CI_4 CI_5 CI_6 .

VAR LABELS
  CI_4 'I am sometimes dissatisfied with my choice of career fields
(REVERSED)'
  CI_5 'Compared to other areas of my life, my chosen line of work is
not very important to me (REVERSED)'
  CI_6 'Sometimes I wish I had chosen a different career field (REVERSED)'.
* usually undesirable execute command used so the OP can look at position of
* new variables in DATA VIEW to see that ITEMS ARE NOT CONTIGUOUS.
execute.

*WRONG because reversed variables are at the right side of the data view.

RELIABILITY
  /VARIABLES=CI_1 TO CI_8  /SCALE('cs')  ALL/MODEL=ALPHA
  /STATISTICS=DESCRIPTIVE SCALE CORR
  /SUMMARY=TOTAL CORR .

*RIGHT.
RELIABILITY
  /VARIABLES=CI_1 CI_2 CI_3 CI_4 CI_5 CI_6 CI_7 CI_8  /SCALE('cs')
ALL/MODEL=ALPHA
  /STATISTICS=DESCRIPTIVE SCALE CORR
  /SUMMARY=TOTAL CORR .

Art Kendall
Social Research Consultants


Mimi Chan wrote:

> Dear Art,
>
> Thanks for your kindly help in here. I try to re-code the data but
> some of the items still shown in negative alpha eg. -0.041 (CS_MBA)
> and -0.225 (CI). So do you know the reason leading to this problem? I
> have attach my SPSS data file in here. I am so worry that these two
> items will affect my whole analyze.
>
> Regards,
> Mimi
>
>
>
> On 8/6/07, *Art Kendall* <[hidden email]
> <mailto:[hidden email]>> wrote:
>
>     Your items include "(reverse).
>     one way to to this is something like the following untested syntax.
>     (adjust the variable list, the number of values, and the values
>     that are
>     declared missing to fit your situation.)
>
>     rename variables (item4 item5 item6 . . . =
>        olditem4 olditem5 olditem6 . . .).
>     recode olditem4 olditem5 olditem6 ...
>     (1=5) (2=4) (3=3) (4=2) (5=1) (else =copy)
>     into item4 item5 item6 . . ..
>     missing values item4 item5 item6 . . . (lo thru 0).
>
>
>     Art Kendall
>     Social Research Consultants
>
>
>     Mimi wrote:
>     > I am a beginner to SPSS. I input all the data in the SPSS and I
>     try to
>     > run the Cronbach's Alpha, but I have no idea why all the outcome are
>     > negative, like I run the Career Involvement, they give me the
>     outcome
>     > as -0.29. So how can I make the alpha back to a nice outcome? This
>     > will be affect all my analysis? Here are the questions and I am
>     using
>     > the Liker style strongly agree to strongly disagree.
>     >
>     >
>     > I identify strongly with my chosen line of work
>     > My chosen line of work gives me a sense of well-being
>     > I get a sense of pride from my chosen line of work
>     > I am sometimes dissatisfied with my choice of career fields
>     (reverse)
>     > Compared to other areas of my life, my chosen line of work is
>     not very
>     > important to me (reverse)
>     > Sometimes I wish I had chosen a different career field (reverse)
>     > If I were to describe myself to someone, I would probably begin by
>     > stating my line of work
>     > If I were to rank (in importance to me) all the things that I do,
>     > those things related to my line of work would be at or near the top
>     >
>
>

[Dennis Deck]

I would name the recoded items something different like CI_4r
This is more explicit, reduces confusion and chances for error.
When using the same variable names, mistakes would be hard to find.

Dennis Deck
RMC Research

-----Original Message-----
From: Art Kendall [mailto:[hidden email]]
Sent: Thursday, August 09, 2007 5:58 AM
Subject: Re: Negative Cronbach's Alpha

I should have mentioned that the reversed items are not contiguous with
the others in the scale.
rename variables (CI_4 CI_5 CI_6=
   olditem4 olditem5 olditem6 ).
recode olditem4 olditem5 olditem6
(1=5) (2=4) (3=3) (4=2) (5=1) (else =copy) into CI_4 CI_5 CI_6 .

VAR LABELS
  CI_4 'I am sometimes dissatisfied with my choice of career fields
(REVERSED)'
  CI_5 'Compared to other areas of my life, my chosen line of work is
not very important to me (REVERSED)'
  CI_6 'Sometimes I wish I had chosen a different career field
(REVERSED)'.
* usually undesirable execute command used so the OP can look at
position of
* new variables in DATA VIEW to see that ITEMS ARE NOT CONTIGUOUS.
execute.

*WRONG because reversed variables are at the right side of the data
view.

RELIABILITY
  /VARIABLES=CI_1 TO CI_8  /SCALE('cs')  ALL/MODEL=ALPHA
  /STATISTICS=DESCRIPTIVE SCALE CORR
  /SUMMARY=TOTAL CORR .

*RIGHT.
RELIABILITY
  /VARIABLES=CI_1 CI_2 CI_3 CI_4 CI_5 CI_6 CI_7 CI_8  /SCALE('cs')
ALL/MODEL=ALPHA
  /STATISTICS=DESCRIPTIVE SCALE CORR
  /SUMMARY=TOTAL CORR .

Art Kendall
Social Research Consultants


Mimi Chan wrote:

> Dear Art,
>
> Thanks for your kindly help in here. I try to re-code the data but
> some of the items still shown in negative alpha eg. -0.041 (CS_MBA)
> and -0.225 (CI). So do you know the reason leading to this problem? I
> have attach my SPSS data file in here. I am so worry that these two
> items will affect my whole analyze.
>
> Regards,
> Mimi
>
>
>
> On 8/6/07, *Art Kendall* <[hidden email]
> <mailto:[hidden email]>> wrote:
>
>     Your items include "(reverse).
>     one way to to this is something like the following untested

syntax.

>     (adjust the variable list, the number of values, and the values
>     that are
>     declared missing to fit your situation.)
>
>     rename variables (item4 item5 item6 . . . =
>        olditem4 olditem5 olditem6 . . .).
>     recode olditem4 olditem5 olditem6 ...
>     (1=5) (2=4) (3=3) (4=2) (5=1) (else =copy)
>     into item4 item5 item6 . . ..
>     missing values item4 item5 item6 . . . (lo thru 0).
>
>
>     Art Kendall
>     Social Research Consultants
>
>
>     Mimi wrote:
>     > I am a beginner to SPSS. I input all the data in the SPSS and I
>     try to
>     > run the Cronbach's Alpha, but I have no idea why all the outcome

are
>     > negative, like I run the Career Involvement, they give me the
>     outcome
>     > as -0.29. So how can I make the alpha back to a nice outcome?
This

>     > will be affect all my analysis? Here are the questions and I am
>     using
>     > the Liker style strongly agree to strongly disagree.
>     >
>     >
>     > I identify strongly with my chosen line of work
>     > My chosen line of work gives me a sense of well-being
>     > I get a sense of pride from my chosen line of work
>     > I am sometimes dissatisfied with my choice of career fields
>     (reverse)
>     > Compared to other areas of my life, my chosen line of work is
>     not very
>     > important to me (reverse)
>     > Sometimes I wish I had chosen a different career field (reverse)
>     > If I were to describe myself to someone, I would probably begin

by
>     > stating my line of work
>     > If I were to rank (in importance to me) all the things that I
do,
>     > those things related to my line of work would be at or near the
top
>     >
>
>