SPSS tutorials

BASICS DATA ANALYSIS T-TEST ANOVA CHI-SQUARE TEST

Creating APA Style Frequency Tables in SPSS

The most basic table in statistics is probably a simple frequency distribution. Sadly, basic frequency tables from SPSS are monstrous. On top of that, they don't meet APA recommendations.
So how to create better frequency tables -preferably fast? This tutorial shows a cool trick for doing just that! We'll use bank_clean.sav throughout, part of which is shown below.

SPSS Bank Sav Data View

Why Basic SPSS Frequency Tables Suck

So let's take a close look at some basic frequency tables. We'll create some by running the syntax below.

*Show only value labels and variable labels in output tables.

set
tnumbers labels
tvars labels.

*Standard SPSS frequencies tables.

frequencies educ marit.

Result

SPSS Basic Frequency Table

For just inspecting your data, this may do. However, these tables are not suitable for reporting:

Now, the styling is easily fixed with a table template. After applying it, our table looks much better as shown below.

SPSS Frequency Table Styled

Unfortunately, FREQUENCIES has no options for avoiding the other issues we just mentioned. So let's try something completely different.

Frequencies in MEANS Tables

Ok, this'll sound crazy but -really- do give it a go. First off, we'll create a new variable holding zeroes for all cases. Next, we'll run a minimal MEANS table for our constant by our target variables. Let's run the syntax below and see what happens.

*Add new variable to data with same value for all cases.

compute constant = 0.

*Basic means tables constant over educ and marit.

means constant by educ marit.

Result

SPSS Basic Means Table

Our table looks stupid. Obviously, all means and standard deviations are 0.000. However, we do have clean and simple frequencies but we don't have the corresponding percentages. Yet.

MEANS without Means

Now the trick is that MEANS allows us to choose which columns we want in which order. Therefore, we can have MEANS without means or standard deviations but with frequencies and percentages. This results in nice and clean frequency tables in the APA recommended format.

*Set variable label for constant.

variable labels constant 'Table ## ...'.

*Run means table but show only frequencies and percentages.

means constant by educ marit
/cells count npct.

*Optionally: prettify tables.

output modify
/select tables
/table tabletitle = ' '
/tablecells select = ['% of Total N'] applyto = columnheader replace = '%'
/tablecells select = ['Total'] applyto = row style = bold.

Result

SPSS APA Frequencies Table No Missing Values

First note that we set part of the desired table title as a variable label for our constant. If you don't want that, try and run variable labels constant ' '. After doing so, the title consists of a single space so there seems to be no title at all.
Also note that we prettified our tables with OUTPUT MODIFY which requires SPSS version 22 or higher. For keeping it simple, we processed all tables in the output window. If you don't want that, adding one or two lines to OUTPUT MODIFY restricts the modifications to a precise selection of tables.
If you're on SPSS 21 or lower, the Ctrl + H shortkey -either in the output window or after exporting to WORD- may help in removing or replacing text.

Including User Missing Values

By default, our approach includes only valid values. However, including user missing values is easily done by just adding a single line to the syntax.

*Same APA frequencies tables but include user missings.

means constant by educ marit
/cells count npct
/missing include.

Including System Missing Values

Very few SPSS procedures can include system missing values. However, that's easily solved: we'll use a simple RECODE to change them to some huge number. We then give that a value label and set it as missing.
Quick tip: don't use very small numbers such as -9999 for this. Small numbers often end up as the first -rather than the last- rows in your tables.

*Recode system missing into huge value.

recode educ marit (sysmis = 999999999).

*Add value label.

add value labels educ marit 999999999 '(Question Skipped)'.

*Set 999999999 as user missing.

missing values educ marit (999999999,7).

*Run APA frequencies with user and system missing values.

means constant by educ marit
/cells count npct
/missing include.

Result

SPSS APA Frequency Table With User And System Missing Values

APA Frequency Tables from CTABLES

For me, creating frequency tables like we just discussed is the preferred option. It's fast and simple. However, an alternative is using CTABLES but this requires a license for the custom tables option.
CTABLES can create a single frequency table for multiple variables in one go. The syntax below presents a minimal example for doing so.

*APA frequency table for multiple variables.

ctables
/table (educ + jtype) [count 'n' colpct.count '%'].

Result

SPSS APA Frequency Tables Via Ctables

So that'll do for our frequency tables. I hope you found this tutorial helpful! Let me know by throwing in a comment below.

Thanks for reading!

Previous tutorial: Simple Overview Statistical Comparison Tests

Next tutorial: SPSS Bar Charts Tutorial

Let me know what you think!

*Required field. Your comment will show up after approval from a moderator.

This tutorial has 2 comments