Why should you learn anything about your SPSS settings? Well, doing so allows you to

• create better output with less effort;
• speed up a huge variety of tasks;
• accomplish tasks that are otherwise impossible.

So which settings are interesting to learn about? The table below presents our top 10.

Top 10 Most Useful SPSS Settings

Background & Googlesheet for All Settings

The vast majority of settings can be adjusted by navigating to Edit Options as shown below.

Sadly, the options dialog is a bit of a labyrinth and finding your way here can be time consuming. Another major stupidity is that there's no Paste button here either.

For these reasons, using SPSS syntax is a much better option for adjusting your settings. But where to start? Well, first off, we created an overview of all settings in this Googlesheet, partly shown below.

Note that Column F suggests if we think some setting is useful. The remainder of this tutorial demonstrates the 10 most useful settings, ordered from most to least important. All examples use bank-clean.sav, partly shown below.

SET TNUMBERS - Table Numbers

TNUMBERS controls how values are shown in output tables. The syntax below presents a quick demo.

Result

Reporting values and value labels is the recommended setting for screening data. However, for creating “clean” tables to include in your final report, you probably want to show only value labels.

SET TVARS - Table Variables

TVARS controls how variables are shown in output tables. For a quick demo, try the syntax below.

Result

For screening data, showing both variable names and labels is often the way to go. For creating reporting tables, however, we recommend showing only variable labels.

SET OVARS - Outline Variables

OVARS controls how variables are shown in the output outline. The syntax below gives a quick demo.

Result

Note that the OVARS setting ignores the bar charts created with the frequency tables. The same goes for histograms. We believe this to be a tiny -but sometimes annoying- bug in SPSS.

SET TLOOK - Tablelook

Tablelooks or .stt files are plain text files containing styling for tables such as fonts, background colors and borders. The syntax demo below requires that clean-11pt.stt is located in a folder d:/templates on your computer.

Result

Notes

• You can also apply tablelooks to existing tables by using the menu or (faster) OUTPUT MODIFY.
• Sadly, the TLOOK setting ignores the CD command: you always need to specify both the folder and the file name...

SET SIGLESS - Significance Less Than...

SIGLESS (introduced in SPSS 27) controls how small significance levels are shown in output tables. We believe there's a bug regarding this setting: the syntax below ignores it unless you run each line separately.

Result

SET CTEMPLATE - Chart Template

Chart templates or .sgt files are plain text files containing styling for charts such as colors, sizes and layout. The quick demo below requires that barchart-freq-dsort-en-h360.sgt is located in a folder d:/templates on your computer.

Result

Notes

• After creating the charts you need, you probably want to set your CTEMPLATE back to NONE.
• Some commands (GRAPH, GGRAPH) have a /TEMPLATE subcommand for applying chart templates but this doesn't work properly.
• You can also apply chart templates to existing charts from the menu or (faster) OUTPUT MODIFY.

SET LOCALE

In recent SPSS versions, LOCALE is mainly used for specifying which decimal separator to use. It may also affect which character encoding SPSS uses but only if UNICODE = OFF (rarely used anymore).

Result

Quick note: I often prefer to use the DOT and COMMA formats for choosing which decimal separator is used. Doing so also makes converting string to numeric variables independent of your settings as discussed in SPSS Convert String to Numeric Variable. The syntax below, however, simply replicates the previous example using this alternative method.

SET SEED - Random Seed

Setting a random seed allows you to exactly replicate anything random in SPSS: computing random variables or drawing random samples. The example below computes two normally distributed random variables and then exactly replicates both.

Result

Minor note: SET SEED only takes effect if RNG = MC (default). You need to use SET MTINDEX if RNG = MT but I don't think anybody ever uses that.

SET PRINTBACK

PRINTBACK tells SPSS whether to print the syntax you run (either from a syntax window or the menu) into your output window. Doing so may come in handy if you export all output to a single .pdf or WORD file.

Result

SET ODISPLAY - Output Display

For GENLINMIXED and NPTESTS, two output formats are available:

• the dreadful “model viewer” object or
• normal output (pivot) tables.

The syntax below creates both formats for a basic Kruskal-Wallis test.

Result

Note: I usually prefer a completely different command for the Kruskal-Wallis test as discussed in How to Run a Kruskal-Wallis Test in SPSS? This method does not include pairwise comparisons but I'm not a big fan of those anyway.

Related Commands - SHOW, PRESERVE & RESTORE

Before closing off, I should point out that you can use SHOW for looking up settings. Like so, SHOW ALL. shows all settings. These include some that you can't set with SET and some that you can't change at all.

Last but not least, you can undo changes in settings by preceding SET with PRESERVE and succeeding it with RESTORE. I find these commands mostly useful for building SPSS tools that change somebody else’s settings.

Right, I hope you found this tutorial helpful. Let me know if you (dis)agree and/or if I missed anything by throwing a quick comment below. Other than that:

Thanks for reading!