Review of Michael N. Mitchell’s Create and Export Tables Using Stata

Abstract

In this article, I review Create and Export Tables Using Stata, by Michael N. Mitchell (2025, Stata Press).

1 Introduction

Good tables are excellent ways of sharing information with others. Creating them can involve a mix of automatic and manual steps. Eliminating manual steps can make tables free from transcribing errors and easy to reproduce. Over the years, several community-contributed commands have helped Stata users create, automate, and share various tables. Recently, Stata introduced new commands to create many kinds of customized tables and share them in various formats; these commands are the collect suite of commands, the dtable command, the etable command, and the reimagined table command. They can be used independently or together in many ways.

The collect suite of commands is the underlying engine beneath the dtable, etable, and table commands. The collect commands allow users to create and customize tables in many ways. The dtable command allows users to easily create a table of descriptive statistics, commonly known as a “Table 1”. The etable command allows users to easily create a table of estimation results in a stacked layout. The table command allows users to create a table of frequencies, summary statistics, and results from other Stata commands.

Unfamiliar commands can take time to learn, especially if they are complex. To help one understand how to use these commands, several resources are available. Stata’s official resources relating to these commands are conveniently listed on this FAQ page: https://www.stata.com/support/faqs/reporting/official-resources-available/. This page has links to Stata’s PDF documentation, recordings of past webinars, web training, Stata Blog posts, YouTube videos, feature descriptions, FAQs, and advice on how to reach out for help.

Michael N. Mitchell’s Create and Export Tables Using Stata is another such resource. Being new to these commands, I read this book to test its stated purpose: “to help you create a table of statistical results that you can export as a Word document, a PDF file, an Excel spreadsheet, an HTML file, or a file”.

2 Content

Before describing the content of each chapter, I wish to share this link, which has some details about the book and its contents: https://www.stata.com/bookstore/create-and-export-tablesusing-stata/.

In chapter 1, Mitchell overviews how to create tables in Stata, overviews the book, and introduces making and exporting tables, as well as provides some commands that are used in the book.

In chapter 2, Mitchell shows how to quickly create various popular table layouts, specifically the following:

Tables of summary statistics

Tables of regression output

Tables showing what can be extracted by using the ttest command, sometimes supplemented with measures of effect size from the use of other commands

Tables featuring parts of correlation matrices, sometimes supplemented by means and standard deviations

Tables relating to the modeling of longitudinal data for two groups (specifically a table of marginal means and a table of between-group differences in mean change from baseline at follow-up time points)

There is a step-by-step explanation of how to achieve each table, with relevant Stata output and a picture of the table after it has been successfully exported to a Word document. The tables of summary statistics are often created with a single line of syntax using the dtable command, while the other tables are typically created using 10-20 lines of syntax using the collect suite of commands.

In chapter 3, Mitchell illustrates how to create more tables of summary statistics, with a little more explanation. Once again, the tables are mostly achieved with one use of the dtable command. However, you will likely need to also use some commands in the collect suite. I appreciated Mitchell illustrating this for summarizing a binary variable on just one row, adding row headers, and customizing the order of the table columns.

In chapter 4, Mitchell illustrates how to create more tables of regression output using the collect suite of commands. In section 4.10 (consisting of 73 pages), Mitchell illustrates how to customize regression tables, but much of the content will appeal to readers wanting to customize other tables. I appreciated learning about autolevels and seeing how to precisely control the numeric format of a cell (see Cole [2015] for guidelines on how many decimal places or significant figures to use for a particular cell). I was excited to learn about saving custom styles, which can reduce the amount of syntax and guarantee consistency of style between tables.

In chapter 5, Mitchell gives a deeper explanation of collections, how collect layout assembles tables, and why the syntax in the book works. I agree with Mitchell that it might be helpful to read this chapter first.

In addition, Mitchell provides information on different ways of exporting tables to Word documents and other formats, describes various errors and how to overcome them, and describes and illustrates the etable command occasionally. The table command is mentioned in chapter 1, used for getting means and standard deviations to add to tables featuring parts of correlation matrices, and used to create a table of frequencies, but that is it.

3 Strengths and weaknesses

As a newcomer to Stata’s recent commands to create and export tables, I found this book to be a good introduction to just what is possible in this space. Despite the book being 461 pages, it did not take me long to have a first read of big chunks of it. Throughout the book, concepts and commands are illustrated with simple examples. Many examples are similar, with one example building on the previous one. There is a fair amount of repetition. The pace of the book is gentle, and explanations are clear. Mitchell does not get bogged down in too much detail.

The book is organized such that readers can dive into parts that most interest them. This may be some of the examples in chapter 2 and parts of the following chapters and appendixes. Of course, reading parts that seem less relevant might also be beneficial.

Typing some of Mitchell’s syntax into Stata, I quickly re-created example tables. Then I set about creating some custom tables of my own. Things slowed down! Of course, figuring things out for the first time often takes some experimenting and reading. I reread parts of the book more slowly, understanding more each time I read. I did succeed in finding a way to create the tables I wanted. Using Mitchell’s book together with other official resources, I have started a repository of do-files to create a variety of tables that I see myself (and possibly colleagues) wanting.

I noticed some examples in Mitchell’s book that were similar to some in Stata’s documentation. I did not mind this. Mitchell sometimes takes a slightly different tack to approaching the same kind of problem. Even within the book, Mitchell may attack the same problem from slightly different angles. I found it helpful that Mitchell’s book used similar variables from a similar dataset to Stata’s documentation.

I found it curious that Mitchell did not write much about or much use the table command and that there was no explanation as to why the focus of the book was only on the dtable command and the collect suite of commands (and not also on the table and etable commands). That said, having read Mitchell’s book, I found myself in a better position to read and understand Stata’s documentation associated with the table and etable commands. I say “better position” because sometimes you can feel as if you have entered another world where things seem quite complicated.

I would have also liked to see other things, for example, a mention that the Tables Builder dialog box (db tables) can be helpful simply to see dimensions, levels, labels, layouts, and previews of tables. I would have liked an example of a popular table in clinical epidemiology: a regression table with “univariable” results (from many models) in one column and “multivariable” results in another (as in table 4 of Pillay et al. [2024]). Another example that would have been good to see is a table popular in clinical trials: one with summary statistics by group, supplemented with a column of various effect sizes, for example, adjusted mean differences, risk differences in percent, etc. (as in table 2 of Tomazini et al. [2024]). To be fair, the book did provide some clues to achieve these tables. It just took me some time to figure out a way.

Conclusion

Create and Export Tables Using Stata is a well-written, example-laden introduction to creating and exporting tables. The book shows how to create various popular table layouts, which are mostly achieved by using the dtable command and the collect suite of commands. Only a few examples are shown using the etable command or the reimagined table command. If the type of table you want is covered in this book, it will not take you long to type the syntax. Most likely, you will want other tables too. With this book and other helpful resources (see my introduction), you will be able to achieve those other tables before long.

I recommend the book as a launch pad for users motivated to learn how to automate the production of good tables using Stata’s latest commands.