[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Icon transition complete

From: Bastián Díaz
Subject: Re: Icon transition complete
Date: Mon, 19 Aug 2013 02:19:46 -0700 (PDT)

I have reviewed several documents available on the IBM website 
( and I could not 
find anything that mentions specifically the behavior I describe. However, I 
can tell that this exists (I see it every day when using SPSS).

From what I could notice in reading, SPSS has actually "algorithms" that help 
identify when a variable is good or bad coded and is mainly used by the SPSS 
"statistical coach (and perhaps used indirectly by some statistical tests).

For example:
→ None "string variable" is used for a statistical analysis. By marking a 
variable as type "string", type of measurement automatically switches to 
"nominal" (type "scale" is not 
selectable In 
contrast to previous versions, SPSS marked as "unknown" new variables 
→ The column properties "role" is defining for some analyzes (can block the use 
of a variable and defining between dependent (target role) and independent 
(imput role)).
→ There are preset conditions to help create variables using transformations or 
other automated systems.These conditions identified according to current data, 
if a variable should be coded in a specific way (type, missing values ​​and 
→ In certain analyzes, SPSS does not allow to use variables with missing cases 
(but these have not been established in their properties).

As the examples above, there are several distributed in the documentation and 
that together, give the user a "guided use" when using the SPSS user interface. 
I consider it a good idea to integrate some of PSPPire.

It would be appropriate in PSPPireunable to choose the type of "scale", when 
you set the variable type "string".

Also ask them to consider include the new column "role" in PSPP. This would 
allow a first step, to create databases compatible with analysis in SPSS 

Bastián Díaz

> I didn't know about that.   Perhaps we should have it too.
> Do you have any more information about it?
> Is it described in any manual somewhere that I can lookup ?

reply via email to

[Prev in Thread] Current Thread [Next in Thread]