SPSS AnswerNet: Result
Solution ID: 100000632
Product: SPSS Base
Title:
Using date variables in TABLES
Description:
Q.
I have a date variable where I'd like to get the mean and standard
deviation in days. When I run the table, I get the mean, but the
standard deviation isn't coming out correctly. I get either all
asterisks, or some really high number. How can I get the standard
deviation out of the date variable?
A.
The following shows how to get a table that has the mean date and
standard deviation in days from the mean date.
* Here is a sample data file. Suppose we want to get
* the mean date and standard deviation in days of the
* DT variable by the VARA variable.
DATA LIST fixed / dt 1-8 (adate) vara 10.
BEGIN DATA
06/11/50 1
11/17/53 1
01/19/56 2
08/26/59 2
END DATA.
* The first thing we do is to use AGGREGATE to compute
* a variable that represents the standard deviation of
* the DT variable. We can then incorporate that value
* back to our original data file.
AGGREGATE outfile=temp.sav / break=vara
/sddt=sd(dt).
MATCH FILES file=* / table=temp.sav / by vara.
* Since the new variable (SDDT) is stored as
* of seconds, we can use a COMPUTE command to translate
* this value into days by using the CTIME.DAYS function.
COMPUTE sddt=trunc(ctime.days(sddt)).
EXECUTE.
FORMATS sddt (f5.0).
* Some cosmetic work here.
VARIABLE LABELS dt 'Mean Date'
sddt 'Standard Deviation in Days' .
* Now we will compute the table. We will use DT and
* SDDT. Although SDDT represents a standard deviation,
* we will use the MEAN statistics to represent it.
* This is because the standard deviation is already
* computed in the dataset, and we just want to display
* the value in the dataset. We can use the MEAN statistic
* to do that.
TABLES observation=dt sddt
/table=dt+sddt by vara
/statistics mean(dt'') mean(sddt'').