Hi,
I just tried ChatGpt for creating PSPP syntax to see if it can create
simple syntax and does not create hallucinations.
My question was basic:
syntax pspp for a ctables procedure, with SocialTrust as a dependent
scale variable and country and education as independent variables.
The result seems reasonable, but my question was also quite basic. It
seems that AI might at least be a learning tool to find your way through
the large manual (Thank you, Ben, for the manual !). For getting
reasonable results, ChatGpt obliges you to ask clear questions, and this
intellectual discipline is perhaps the most important result of its use.
- ftr
And the result produced by ChatGpt was:
Given that *|SocialTrust|* is a *scale (continuous) dependent variable*
and you want to examine it across the categories of *|country|* and
*|education|*, the correct PSPP syntax using the *|CTABLES|* procedure
to display descriptive statistics (e.g., *mean*) is as follows:
CTABLES
/TABLE country [C] BY education [C]
/SUMMARY VARIABLES=SocialTrust
STATISTICS=MEAN STDDEV MIN MAX N
FORMAT=DEFAULT.
🔍 Explanation:
*
|/TABLE country [C] BY education [C]|: Treats both |country| and
|education| as *categorical* grouping variables.
*
|/SUMMARY VARIABLES=SocialTrust|: Specifies that |SocialTrust| is
the *dependent scale variable*.
*
|STATISTICS=MEAN STDDEV MIN MAX N|: Requests summary statistics —
*mean*, *standard deviation*, *minimum*, *maximum*, and *count (N)*.
*
|FORMAT=DEFAULT|: Uses the default format for table output.
Optional: For cleaner output (only means):
CTABLES
/TABLE country [C] BY education [C]
/SUMMARY VARIABLES=SocialTrust
STATISTICS=MEAN.
------------------------------------------------------------------------
