correlation between two measures setting it up

[Talino Bruno]

I am planning on doing a correlation between two measures from different
informants assessing one individual. So I have a teacher evaluation and a
studnet self-evaluation and would like to see how closely they are
associated. My problem is how to set that up in SPSS. Do I need to input
the item scores from each measure then run a correlation? I assume so.
Should the subjects' id be recored down the column and the individual items
recorded on the row. If so, can I put the score of both measures on one row
if they correspond to one subject?

=====================
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

[Maguin, Eugene]

Talino,

>>Do I need to input the item scores from each measure then run a
correlation?

Sounds like your measure(s) are multiple item scales. So, you either have to
input the item scores, sum or average the items to get a total or mean for
each measure that you then correlate. Or, you input the scale totals or
means and correlate those numbers.

>>Should the subjects' id be recored down the column and the individual
items
recorded on the row. If so, can I put the score of both measures on one row
if they correspond to one subject?

Yes to both questions. So suppose one scale had 3 items (S1I1-S1I3) and the
other scale had five items (S2I1-S2I5). Your data file would look like

ID S1I1 S1I2 S1I3 S2I1 S2I2 S2I3 S2I4 S2I5

Gene Maguin

=====================
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

[Clive Downs]

Hi Taliono,

If I understand your question accurately, you wanted to know how to arrange
your data so as to examine the correlation between two raters.

I hope the following syntax will show how to do it. The syntax includes in-
line data, but the format shows how you need to set up your data. I hope I
have understood your question, and that this helps.

* ------------------------------------------------------------.
* SECTION 1 create example data set.
*   S001 is the ID, tev is teacher evaluation, sev is student evaluation.
*    eg S001 is ID of first student, 23 is their teacher evaluation, 24 .
*    is their self evaluation.
*-----------------------------------------------------.

DATA LIST FREE/ ID(A4) tev sev.
BEGIN DATA

S001 23 24
S002 32 33
S003 35 29
S004 37 24
S005 21 24
S006 35 38
S007 43 24
S008 32 12

END DATA.

FILE HANDLE evals /NAME = 'H:\spss\evals.sav'.
SAVE OUTFILE = evals.
EXE.

* SECTION 2 run correlation.
*-------------------------------------.
* run Pearson correlation.
CORRELATIONS
  /VARIABLES=tev sev
  /PRINT=TWOTAIL NOSIG
  /STATISTICS DESCRIPTIVES
  /MISSING=PAIRWISE .

=====================
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