Re: confidence interval for geo mean?

Posted by David Marso on Nov 09, 2011; 3:54pm
URL: http://spssx-discussion.165.s1.nabble.com/confidence-interval-for-geo-mean-tp4975675p4978107.html

GET FILE='C:\Program Files\SPSS\1991 U.S. General Social Survey.sav'.
MEANS TABLES=age  /CELLS GEOMETRIC  .
COMPUTE No_Break=1.
COMPUTE L_age=LN(age).
AGGREGATE OUTFILE * / BREAK No_Break / M_LAge=MEAN(L_age) /SdLAge=SD(L_Age)/ N=N.
COMPUTE #Half=IDF.T(.975,N-1)*SdLAge/SQRT(N).
COMPUTE LCL=EXP(M_Lage-#Half).
COMPUTE UCL=EXP(M_Lage+#Half).
COMPUTE GMean=EXP(M_Lage).
exe.

Ian Martin-2 wrote

I can't find any way to get back transformed values for a confidence interval for the geometric mean.  Surely I don't need to log transform all the data and do the stats and then back transform the results?

regards,
Ian

Ian D. Martin, Ph.D.

Tsuji Laboratory
University of Waterloo
Dept. of Environment & Resource Studies

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