Extract Diagonals from a Correlations Table
Locked
Extract Diagonals from a Correlations Table
 
|
Dear SPSSers,
I have correlations between ratings at two points in time on 5 items, provided by the syntax below: CORRELATIONS /VARIABLES=a01 to a05 WITH b01 to b05 /PRINT=TWOTAIL NOSIG /MISSING=PAIRWISE . I used OMS to put the table into a new SAV file, which was then cleaned to remove the N and sig values, leaving only the r-values. I would like to create a new variable, the value of which is the the diagonal value in each row. I would use MAX to do this, but the diagonal values are not necessarily the maximum values in each row. Fake sample before and after data are provided below: *** Before . DATA LIST LIST /case_lbl (A4) var1 to var5 . BEGIN DATA "var1" .20 .19 .17 .12 .04 "var2" .09 .16 .01 .03 .14 "var3" .15 .11 .17 .04 .02 "var4" .10 .04 .06 .09 .05 "var5" .04 .17 .01 .11 .30 END DATA . *** After . DATA LIST LIST /case_lbl (A4) var1 to var5 diag . BEGIN DATA "var1" .20 .19 .17 .12 .14 .20 "var2" .09 .16 .01 .03 .14 .16 "var3" .15 .11 .17 .04 .02 .17 "var4" .10 .04 .06 .09 .05 .09 "var5" .04 .17 .01 .11 .30 .30 END DATA . Can anyone help with a reference or solution? I should note that while I have some experience with basic syntax, I have little with macros and scripts, and none with Python. Thank you in advance for your advice. -Jon |
Re: Extract Diagonals from a Correlations Table
|
|
As stated, this syntax will do that for you:
DATA LIST LIST /case_lbl (A4) var1 to var5 . BEGIN DATA "var1" .20 .19 .17 .12 .14 "var2" .09 .16 .01 .03 .14 "var3" .15 .11 .17 .04 .02 "var4" .10 .04 .06 .09 .05 "var5" .04 .17 .01 .11 .30 END DATA . vector v=var1 to var5. compute diag=v($casenum). Use Execute or some procedure to pass the data. You don't say what you want to do after you compute this value so I'll stop here. $casenum is a system variable which is just the case number. -----Original Message----- From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Jon Bernard Sent: Thursday, October 04, 2007 2:23 PM To: [hidden email] Subject: Extract Diagonals from a Correlations Table Dear SPSSers, I have correlations between ratings at two points in time on 5 items, provided by the syntax below: CORRELATIONS /VARIABLES=a01 to a05 WITH b01 to b05 /PRINT=TWOTAIL NOSIG /MISSING=PAIRWISE . I used OMS to put the table into a new SAV file, which was then cleaned to remove the N and sig values, leaving only the r-values. I would like to create a new variable, the value of which is the the diagonal value in each row. I would use MAX to do this, but the diagonal values are not necessarily the maximum values in each row. Fake sample before and after data are provided below: *** Before . DATA LIST LIST /case_lbl (A4) var1 to var5 . BEGIN DATA "var1" .20 .19 .17 .12 .04 "var2" .09 .16 .01 .03 .14 "var3" .15 .11 .17 .04 .02 "var4" .10 .04 .06 .09 .05 "var5" .04 .17 .01 .11 .30 END DATA . *** After . DATA LIST LIST /case_lbl (A4) var1 to var5 diag . BEGIN DATA "var1" .20 .19 .17 .12 .14 .20 "var2" .09 .16 .01 .03 .14 .16 "var3" .15 .11 .17 .04 .02 .17 "var4" .10 .04 .06 .09 .05 .09 "var5" .04 .17 .01 .11 .30 .30 END DATA . Can anyone help with a reference or solution? I should note that while I have some experience with basic syntax, I have little with macros and scripts, and none with Python. Thank you in advance for your advice. -Jon |
|
|
Hi All
I have 3 survey questions each with a before and after answer from two groups (N=393, N=540). I have put this in an SPSS data file. There are seven variables corresponding to the three before and three after survey questions plus one variable as a group coding dummy variable to identify the two groups. The crosstabs option does not allow me to run the computation. I have looked in help and my SPSS books (eg Pallant) but the examples they give a unlike mine. Suggestions appreciated Warm regards/gary |
Re: Extract Diagonals from a Correlations Table
|
|
In reply to this post by ViAnn Beadle
That is perfect, thank you very much, ViAnn!
Jon On 10/4/07, ViAnn Beadle <[hidden email]> wrote: > > As stated, this syntax will do that for you: > > DATA LIST LIST /case_lbl (A4) var1 to var5 . > BEGIN DATA > "var1" .20 .19 .17 .12 .14 > "var2" .09 .16 .01 .03 .14 > "var3" .15 .11 .17 .04 .02 > "var4" .10 .04 .06 .09 .05 > "var5" .04 .17 .01 .11 .30 > END DATA . > vector v=var1 to var5. > compute diag=v($casenum). > > Use Execute or some procedure to pass the data. You don't say what you > want > to do after you compute this value so I'll stop here. > > $casenum is a system variable which is just the case number. > > -----Original Message----- > From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of > Jon > Bernard > Sent: Thursday, October 04, 2007 2:23 PM > To: [hidden email] > Subject: Extract Diagonals from a Correlations Table > > Dear SPSSers, > > I have correlations between ratings at two points in time on 5 items, > provided by the syntax below: > > CORRELATIONS > /VARIABLES=a01 to a05 WITH b01 to b05 > /PRINT=TWOTAIL NOSIG > /MISSING=PAIRWISE . > > I used OMS to put the table into a new SAV file, which was then cleaned to > remove the N and sig values, leaving only the r-values. > > I would like to create a new variable, the value of which is the the > diagonal value in each row. I would use MAX to do this, but the diagonal > values are not necessarily the maximum values in each row. > > Fake sample before and after data are provided below: > > *** Before . > > DATA LIST LIST /case_lbl (A4) var1 to var5 . > BEGIN DATA > "var1" .20 .19 .17 .12 .04 > "var2" .09 .16 .01 .03 .14 > "var3" .15 .11 .17 .04 .02 > "var4" .10 .04 .06 .09 .05 > "var5" .04 .17 .01 .11 .30 > END DATA . > > *** After . > > DATA LIST LIST /case_lbl (A4) var1 to var5 diag . > BEGIN DATA > "var1" .20 .19 .17 .12 .14 .20 > "var2" .09 .16 .01 .03 .14 .16 > "var3" .15 .11 .17 .04 .02 .17 > "var4" .10 .04 .06 .09 .05 .09 > "var5" .04 .17 .01 .11 .30 .30 > END DATA . > > Can anyone help with a reference or solution? I should note that while I > have some experience with basic syntax, I have little with macros and > scripts, and none with Python. > > Thank you in advance for your advice. > > -Jon > > |
|
|
In reply to this post by Gary Oliver
Gary,
Are you interested in a two-way analysis (e.g. pre- versus post- for a specific question), or a three-way analysis (e.g. showing that the pre-post relationship differs between groups, or that the post answer differs between groups after controlling for the pre- response. I am wondering whether your research questions are ones that lend themselves to loglinear analysis and logistic regression. Best, Stephen For personalized and professional consultation in statistics and research design, visit www.statisticsdoc.com -----Original Message----- From: SPSSX(r) Discussion [mailto:[hidden email]]On Behalf Of Gary Oliver Sent: Thursday, October 04, 2007 7:13 PM To: [hidden email] Subject: Chi-Square in Crosstabs Hi All I have 3 survey questions each with a before and after answer from two groups (N=393, N=540). I have put this in an SPSS data file. There are seven variables corresponding to the three before and three after survey questions plus one variable as a group coding dummy variable to identify the two groups. The crosstabs option does not allow me to run the computation. I have looked in help and my SPSS books (eg Pallant) but the examples they give a unlike mine. Suggestions appreciated Warm regards/gary |
|
|
In reply to this post by Gary Oliver
I'm not sure what you mean by the option "does not allow me to run the
computation". You should have two sub-tables, one for each group and a chi-square value for each before and after table. What hypothesis are you testing? IMHO, you have a within-subjects design here for which a chi-square test does not seem to be suitable (note that I am not a statistician and don't even pretend to be one on TV). -----Original Message----- From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Gary Oliver Sent: Thursday, October 04, 2007 5:13 PM To: [hidden email] Subject: Chi-Square in Crosstabs Hi All I have 3 survey questions each with a before and after answer from two groups (N=393, N=540). I have put this in an SPSS data file. There are seven variables corresponding to the three before and three after survey questions plus one variable as a group coding dummy variable to identify the two groups. The crosstabs option does not allow me to run the computation. I have looked in help and my SPSS books (eg Pallant) but the examples they give a unlike mine. Suggestions appreciated Warm regards/gary |
|
|
Gary,
Are you trying to match up cases pre test and post test? As an example, are you looking at the same person's results from the first survey and comparing these to the results to the second survey? Your problem may be the way your data is stored, is the data set up so that you have each person as a case, and each case has the results from the pre and post test stored? Knowing how your data is stored is very important to the types of analysis you can run. On 10/4/07, ViAnn Beadle <[hidden email]> wrote: > > I'm not sure what you mean by the option "does not allow me to run the > computation". You should have two sub-tables, one for each group and a > chi-square value for each before and after table. What hypothesis are you > testing? IMHO, you have a within-subjects design here for which a > chi-square > test does not seem to be suitable (note that I am not a statistician and > don't even pretend to be one on TV). > > -----Original Message----- > From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of > Gary Oliver > Sent: Thursday, October 04, 2007 5:13 PM > To: [hidden email] > Subject: Chi-Square in Crosstabs > > Hi All > > I have 3 survey questions each with a before and after answer from two > groups (N=393, N=540). I have put this in an SPSS data file. There are > seven variables corresponding to the three before and three after survey > questions plus one variable as a group coding dummy variable to identify > the two groups. > > The crosstabs option does not allow me to run the computation. > > I have looked in help and my SPSS books (eg Pallant) but the examples > they give a unlike mine. Suggestions appreciated > > Warm regards/gary > -- Björn Türoque 375 Hudson Street New York, NY 10014-3657 212-366-2000 Some people are just born to rock! |
|
|
Hi Bjorn (sorry I am missing the diacritic characters over the letters)
Thanks for your reply I think you are correct referring to my data store (variables). But I have exactly what you describe: a before and after variable and there are 540 cases However that does not seem to be what SPSS crosstabs wants because the screen dialog box wants both row variables and column variables Warm regards/gary >>-----Original Message----- >>From: SPSSX(r) Discussion [mailto:[hidden email]] >>On Behalf Of Björn Türoque >>Sent: Saturday, October 06, 2007 6:30 AM >>To: [hidden email] >>Subject: Re: Chi-Square in Crosstabs >> >>Gary, >> >>Are you trying to match up cases pre test and post test? As >>an example, are you looking at the same person's results from >>the first survey and comparing these to the results to the >>second survey? Your problem may be the way your data is >>stored, is the data set up so that you have each person as a >>case, and each case has the results from the pre and post test stored? >> >>Knowing how your data is stored is very important to the >>types of analysis you can run. >> >> >> >> >>On 10/4/07, ViAnn Beadle <[hidden email]> wrote: >>> >>> I'm not sure what you mean by the option "does not allow me >>to run the >>> computation". You should have two sub-tables, one for each >>group and a >>> chi-square value for each before and after table. What >>hypothesis are >>> you testing? IMHO, you have a within-subjects design here >>for which a >>> chi-square test does not seem to be suitable (note that I am not a >>> statistician and don't even pretend to be one on TV). >>> >>> -----Original Message----- >>> From: SPSSX(r) Discussion [mailto:[hidden email]] >>On Behalf >>> Of Gary Oliver >>> Sent: Thursday, October 04, 2007 5:13 PM >>> To: [hidden email] >>> Subject: Chi-Square in Crosstabs >>> >>> Hi All >>> >>> I have 3 survey questions each with a before and after >>answer from two >>> groups (N=393, N=540). I have put this in an SPSS data >>file. There are >>> seven variables corresponding to the three before and three after >>> survey questions plus one variable as a group coding dummy >>variable to >>> identify the two groups. >>> >>> The crosstabs option does not allow me to run the computation. >>> >>> I have looked in help and my SPSS books (eg Pallant) but >>the examples >>> they give a unlike mine. Suggestions appreciated >>> >>> Warm regards/gary >>> >> >> >> >>-- >>Björn Türoque >>375 Hudson Street >>New York, NY 10014-3657 >>212-366-2000 >> >>Some people are just born to rock! >> |
|
|
Gary,
What SPSS wants to know when you are ordering a crosstabulation is which variables (columns in your data editor) will appear in the rows of your crosstabulation, and which variables will appear in the columns of your crosstabulation. For instance, if you are ordering a table of gender by hair color, you may have one row for each gender and one column for each hair color, and thus gender is your row variable and hair color your column variable in that particular tabulation. Or you may have it the other way if you like. But both variables occupy columns in your data editor. You can design your table the way you like it best. Moreover, once your table is produced you may easily rotate the table with the pivot table editor (select the table and click on Pivot), putting genders in the columns and hair colors in the rows. Hector -----Original Message----- From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Gary Oliver Sent: 06 October 2007 00:02 To: [hidden email] Subject: Re: Chi-Square in Crosstabs Hi Bjorn (sorry I am missing the diacritic characters over the letters) Thanks for your reply I think you are correct referring to my data store (variables). But I have exactly what you describe: a before and after variable and there are 540 cases However that does not seem to be what SPSS crosstabs wants because the screen dialog box wants both row variables and column variables Warm regards/gary >>-----Original Message----- >>From: SPSSX(r) Discussion [mailto:[hidden email]] >>On Behalf Of Björn Türoque >>Sent: Saturday, October 06, 2007 6:30 AM >>To: [hidden email] >>Subject: Re: Chi-Square in Crosstabs >> >>Gary, >> >>Are you trying to match up cases pre test and post test? As >>an example, are you looking at the same person's results from >>the first survey and comparing these to the results to the >>second survey? Your problem may be the way your data is >>stored, is the data set up so that you have each person as a >>case, and each case has the results from the pre and post test stored? >> >>Knowing how your data is stored is very important to the >>types of analysis you can run. >> >> >> >> >>On 10/4/07, ViAnn Beadle <[hidden email]> wrote: >>> >>> I'm not sure what you mean by the option "does not allow me >>to run the >>> computation". You should have two sub-tables, one for each >>group and a >>> chi-square value for each before and after table. What >>hypothesis are >>> you testing? IMHO, you have a within-subjects design here >>for which a >>> chi-square test does not seem to be suitable (note that I am not a >>> statistician and don't even pretend to be one on TV). >>> >>> -----Original Message----- >>> From: SPSSX(r) Discussion [mailto:[hidden email]] >>On Behalf >>> Of Gary Oliver >>> Sent: Thursday, October 04, 2007 5:13 PM >>> To: [hidden email] >>> Subject: Chi-Square in Crosstabs >>> >>> Hi All >>> >>> I have 3 survey questions each with a before and after >>answer from two >>> groups (N=393, N=540). I have put this in an SPSS data >>file. There are >>> seven variables corresponding to the three before and three after >>> survey questions plus one variable as a group coding dummy >>variable to >>> identify the two groups. >>> >>> The crosstabs option does not allow me to run the computation. >>> >>> I have looked in help and my SPSS books (eg Pallant) but >>the examples >>> they give a unlike mine. Suggestions appreciated >>> >>> Warm regards/gary >>> >> >> >> >>-- >>Björn Türoque >>375 Hudson Street >>New York, NY 10014-3657 >>212-366-2000 >> >>Some people are just born to rock! >> |
«
Return to SPSSX Discussion
|
1 view|%1 views
| Free forum by Nabble | Edit this page |