too little data
Locked
 
|
|
|
|
Show us your syntax and your output. From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Anata Ionescu
|
|
In reply to this post by Anata Ionescu
Dear Anata,
I've no idea! The only thing I can think of, is that the data only look like 3, 5, 2 in SPSS but that there's actually nonvisible decimals. What you see, is not what you get. If you run the syntax below, the data will look like 3, 5, 2 but the correlation will be only .76. Will the real v1 and v2 please stand up now? The last line of syntax will reveal their true nature (in this case at least). P.s. you may have to replace the comma decimal separators by periods. Kind regards, Ruben van den Berg data list free/v1 v2(2f1.0). begin data 2,5 3,49 5,49 4,5 2,49 1,5 end data. cor v1 v2. for v1 v2(f3.2). Date: Tue, 1 Dec 2009 12:00:16 -0800 From: [hidden email] Subject: too little data To: [hidden email]
New Windows 7: Simplify what you do everyday. Find the right PC for you. |
|
|
Hi Anata -
Ruben is thinking exactly what I was. Go into your table and expand the decimal field to see if you've got something like 2.79 instead of 3, etc. Sonia Sonia Brandon, Ph.D. Director, Institutional Research Mesa State College 1100 North Avenue Grand Junction, CO 81501 phone: (970) 248-1884 fax: (970) 248-1812 email: [hidden email] >>> Ruben van den Berg <[hidden email]> 12/1/2009 1:45 PM >>> Dear Anata, I've no idea! The only thing I can think of, is that the data only look like 3, 5, 2 in SPSS but that there's actually nonvisible decimals. What you see, is not what you get. If you run the syntax below, the data will look like 3, 5, 2 but the correlation will be only .76. Will the real v1 and v2 please stand up now? The last line of syntax will reveal their true nature (in this case at least). P.s. you may have to replace the comma decimal separators by periods. Kind regards, Ruben van den Berg data list free/v1 v2(2f1.0). begin data 2,5 3,49 5,49 4,5 2,49 1,5 end data. cor v1 v2. for v1 v2(f3.2). Date: Tue, 1 Dec 2009 12:00:16 -0800 From: [hidden email] Subject: too little data To: [hidden email] Hi! I am not a statistician so you may find my question silly, but I wonder why do I have: Var1: 3 5 2 Var2: 3 5 2, and, with SPSS, I get a correlation of +0.8 instead of +1? _________________________________________________________________ New Windows 7: Simplify what you do everyday. Find the right PC for you. http://windows.microsoft.com/shop ===================== 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 |
|
|
Since variables are columns, if I create a data file within SPSS with:
3 3 5 5 2 2 and run a Pearson correlation, I get the expected 1.0, even with only 3 values per variable. I'm kind of surprised with only 3 rows (too little data), but indeed I got 1.0. Cheers, Necia A. Black, Ph.D. e-mail: [hidden email] 246 Computing Center Web URL:http://www.acsu.buffalo.edu/~black SUNY-Buffalo Buffalo, New York 14260 On Tue, 1 Dec 2009, Sonia Brandon wrote: > Hi Anata - > > Ruben is thinking exactly what I was. Go into your table and expand the decimal field to see if you've got something like 2.79 instead of 3, etc. > > Sonia > > Sonia Brandon, Ph.D. > Director, Institutional Research > Mesa State College > 1100 North Avenue > Grand Junction, CO 81501 > phone: (970) 248-1884 > fax: (970) 248-1812 > email: [hidden email] > > >>>> Ruben van den Berg <[hidden email]> 12/1/2009 1:45 PM >>> > > Dear Anata, > > > > I've no idea! The only thing I can think of, is that the data only look like 3, 5, 2 in SPSS but that there's actually nonvisible decimals. What you see, is not what you get. If you run the syntax below, the data will look like 3, 5, 2 but the correlation will be only .76. Will the real v1 and v2 please stand up now? The last line of syntax will reveal their true nature (in this case at least). > > > > P.s. you may have to replace the comma decimal separators by periods. > > > > Kind regards, > > > > Ruben van den Berg > > > > data list free/v1 v2(2f1.0). > > begin data > > 2,5 3,49 5,49 4,5 2,49 1,5 > > end data. > > cor v1 v2. > > > > for v1 v2(f3.2). > > > > > > > > > > > > Date: Tue, 1 Dec 2009 12:00:16 -0800 > From: [hidden email] > Subject: too little data > To: [hidden email] > > > > > > Hi! > > I am not a statistician so you may find my question silly, but I wonder why do I have: > Var1: 3 5 2 > Var2: 3 5 2, > and, with SPSS, I get a correlation of +0.8 instead of +1? > > > > _________________________________________________________________ > New Windows 7: Simplify what you do everyday. Find the right PC for you. > http://windows.microsoft.com/shop > > ===================== > 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 > > > ===================== 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 |
|
I always understood that the maximum r on a 2 x 2
table was .707, but that was almost 40 years ago when SPSS first appeared
in UK.
|
|
|
r2= 0.5 for 2 by 2 Matrix being maximum 5
Is that game theory that we are talking Some Kind of Prisoner's Dilemma or Just equal sized boxes Graduate Student University of Tennessee, Knoxville. Go Vols! Websites- http://decisionstats.com http://dudeofdata.com http://prayers2go.com Linkedin- www.linkedin.com/in/ajayohri Facebook-www.facebook.com/ajayohri Twitter-www.twitter.com/dudeofdata Quote for the Day- Pablo Picasso - "Computers are useless. They can only give you answers." On Sun, Dec 6, 2009 at 7:37 AM, John F Hall <[hidden email]> wrote:
|
|
In reply to this post by John F Hall
Was that under some set of values or n's? try this example syntax.
data list list /x(f1) y(f1) kount(f4). begin data 1 1 1000 2 2 1000 end data. weight by kount. crosstabs tables = x by y /statistics = corr. CORRELATIONS /VARIABLES=x y /PRINT=TWOTAIL NOSIG /MISSING=PAIRWISE. NONPAR CORR /VARIABLES=x y /PRINT=BOTH TWOTAIL NOSIG /MISSING=PAIRWISE. weight off. compute kount = kount/10. weight by kount. crosstabs tables = x by y /statistics = corr. weight off. compute kount = kount/10. weight by kount. crosstabs tables = x by y /statistics = corr. weight off. compute kount = kount/10. weight by kount. crosstabs tables = x by y /statistics = corr. Art Kendall Social Research Consultants John F Hall wrote: ===================== 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
Art Kendall
Social Research Consultants |
«
Return to SPSSX Discussion
|
1 view|%1 views
| Free forum by Nabble | Edit this page |