Standards for the formatting of tables or the uniform use of abbreviations are becoming more and more established. Whether for a term paper, a doctoral dissertation or for publication in a scientific journal, the style and format of a paper must meet a certain requirement. The standards from the American Psychological Association (APA) are widely used for this, and not just in psychology. We at STATWORX also like to use this method, because tables formatted according to APA are visually appealing. But how exactly do my results tables have to be designed, and can I make my work easier somehow?? Exactly to these questions you get an answer here.

## Overview of the APA Standard

The APA publishes a guideline for the publication of scientific articles, the so-called Publication Manual, at irregular intervals of several years. There are guidelines for almost every question about the structure of a scientific paper, writing style, citation or the comprehensible presentation of empirical results as a table. Currently, the sixth edition from 2010 is valid, which is also the basis for all information in this article.

## General guidelines for tables

If results are to be presented, the first question is whether tables are necessary at all. The Publication Manual advises to be selective and to keep three aspects in mind. On the one hand, it can be very exhausting for the reader if many tables have to be worked through. In addition, it becomes more difficult to follow a text mentally if it is interrupted too often by tables. Finally, certain information (such as the result of a single analysis of variance) can easily be integrated into the text.

The same principle applies to statistics in general: good planning is half the battle. Tables at any price should therefore be avoided. If they really offer an added value, they must be well integrated into the text. But be careful: The Publication Manual clearly states that tables should be understandable on their own. Abbreviations must therefore also be explained there.

In addition, tables should be labeled with numbers in the order in which they are mentioned in the text. The labels should be Table , Table and Table and no suffixes like Table and should be used. In addition, each table must be referred to in the text. These references should be made unambiguous, for example with "as shown in table, …". References such as "see above table" or "the table on page " should not be used. Tables in the appendix should additionally be assigned to the respective section with capital letters. Table would be accordingly the first table in the appendix .

## Tables according to APA standards

Let us now look at how tables should look in detail according to APA standards. Since pictures speak louder than words, here is an example of how a simple table looks according to APA standards.

According to the standards, as few horizontal lines as possible are used and no vertical lines at all. Clarity was created more by free rows and columns between individual aspects. The font chosen is Times New Roman with a font size of pt, which is recommended in principle. However, the font should not differ from the rest of the text. Both the title of the table and the respective abbreviations and the word "Notes" are written in italics.

The general notes for the table appear first below the table, including the definition of abbreviations. Individual notes and explanations of abbreviations end with a period and are thus separated from each other. Copyright information is also part of the general remarks and this should of course always be indicated.

In a separate line below the general remarks follow the specific remarks. They are symbolized with small letters and assigned in order from left to right and from top to bottom. Starting at the top left. Several specific notes are written directly one after the other and again separated by a period.

At the very end of the notes are those on the significance level. They also start on a new line and are separated by a period. Indications smaller than "*** " (not shown in the example) should not be used. If it is necessary to distinguish between one-sided and two-sided p-values, the * symbol is used for two-sided and another symbol for the one-sided p-values. For example: "* , two-sided. † , one-sided.". Since the specification of an exact p-value is always more informative than the division into less than or, it is recommended to specify p-values directly in the table up to the second or third decimal place.

In addition to the annotations, confidence intervals are also plotted in the table. In the past, often only the point estimates were reported together with the p-value, but nowadays it is recommended to also report the confidence intervals. The corresponding confidence level is explicitly stated in each case (here ), which should definitely be done. Usually, the same confidence level is also used for all calculations of an article. However, since each table should be understandable on its own, the confidence level must be clearly marked in each case. The representation in square brackets chosen in the example is not the only option here. It is also possible to show the lower and upper limits of the confidence interval each in a separate column. This is illustrated again in the table below.

## Special tables of results

Depending on the statistical method used, there are of course other ratios that must be presented to the reader. Therefore, in the following you will find different tables again, each of which is intended to illustrate a different aspect. Hopefully, this gives you an idea of how a table according to the APA standard generally looks best.

Let’s start with a table that includes descriptive analyses. The values are presented separately for men and women. A special feature is the value for Cronbach’s, which is reported without leading, because the value cannot be greater than. In addition, an explanation of the range of values is included, namely how wide it is theoretically and how wide it actually was with the data. Finally, it should be mentioned that in the general remarks the reason for the different sample sizes is specified.

Next, a table is presented that contains all relevant information for a t-test with dependent samples. It is crucial that both the mean and the standard deviation for each time point are given. But also the degrees of freedom of the t-test (here ) as well as the confidence interval and the effect size (Cohen’s *d*) are shown and belong to such a summary table.

Finally, a table for a linear regression, which is probably the most frequently used statistical method. The table contains different models, so it is a stepwise regression. In the first step only the control variables are included and in the second step additionally the relevant predictors. In addition to the unstandardized regression coefficients, the first model already includes this and the associated F-value. Together with the sample size specified in the general notes, this information is sufficient to evaluate the first model. For the second model the same information can be found in the table, but additionally the confidence interval as well as the and . The confidence intervals are important to see the estimated range in which the true value of the regression coefficients is likely to lie. The and is used to assess whether the inclusion of the relevant predictors *additionally* to the control variables significantly explain variance from the dependent variable. Taking all this information together, we get a rounded picture of how the predictors relate to the dependent variable.

## Tables according to APA in R

To create tables according to APA manually would be worth the effort of a clear presentation, but in times of statistical programs one can (fortunately) save this work, too. Well, let’s say at least facilitate.

In the program R there is a rather new package with the unique name "apaTables". Thus, one can almost effortlessly create tables in Word according to the APA guidelines. Tables for analyses of variance (with and without repeated measures), effect sizes (Cohen’s and ), correlations and regressions are currently possible. You can even have bootstrap confidence intervals generated for the regression.

As an example, let’s create a correlation table below:

The apa command.cor.table() calculates the bivariate Pearson correlation for all variables in the data set attitude and presents it together with the corresponding confidence interval in a table according to APA guidelines. This is shown in the file "Correlation.doc" and then contains the following table:

The table already looks very good, but two points have to be changed manually in any case. A paragraph has to be inserted in front of the *-data for the p-values, because according to the APA standard these notes should be in a separate line. Also, the numbers must be centered in all cells. That being said, in this (created in a few seconds!) Table considers many APA standards. For example, the designations *M* for mean value and *SD* for standard deviation written in italics, which should always be done if latin letters are used as abbreviation for statistical parameters. In addition, the decimal point in front of the correlations and the p-value is always omitted, which is the standard for numbers that cannot be larger than.

## Tables according to APA in SPSS

Also in the program SPSS you can make your work easier and output tables directly to APA. To do this, one must go to **SPSS> Settings..** in the new window that opens **Pivot tables** select. There you can set the general template for result tables.

However, the "APA_TimesRoman_12pt" template used here is not implemented by default in SPSS. It must be created manually, which has the advantage that new APA standards or special, additional requirements can be implemented manually. You can download the template shown in the example here. You then have to add it to the "Looks" folder of SPSS and then select it as default in SPSS. The folder "Looks" can be found where you have installed SPSS. Otherwise, the reworking of an already existing template is documented many times on the Internet.

## Summary

Creating tables correctly formatted according to APA used to be a lot of work, but with the help of modern statistical programs this has already been greatly reduced. They are definitely worth the effort, because if done correctly, they are not only good to look at, but also good to read. If you want to know even more about APA standards or need help preparing your statistical results, we at STATWORX are of course happy to help (contact us directly).