Building customizable and reproducible tables is now even easier in Stata 19. The table command for creating and customizing tables in a single command now allows you to add a title and notes and export tables. It is also easier to create tables from a collection. anova, oneway, and tabulate now work seamlessly with collect. And collect now allows you to query layout specifications and remove results from a collection.

Let's see it work

More powerful single-command table generation

The table command is a flexible tool for creating a tabulation, table of summary statistics, table of regression results, or table of results from other Stata commands. Its options allow you to customize numeric and string formats, add stars, and more. Now you can add and customize a title and notes. You can also export the table directly to Excel, Word, LaTeX, and more. And you can do this all with a single table command.

Here we use data from the Second National Health and Nutrition Examination Survey (NHANES II) (McDowell et al. 1981). We tabulate the hlthstat variable, which contains individuals’ self-reported health status categories. We also specify options for adding a title and notes and for exporting the table to a Word document.

. webuse nhanes2l
(Second National Health and Nutrition Examination Survey)

. table hlthstat,
          statistic(frequency) statistic(percent)
          missing
          title(Table 1)
          note("Source: NHANES II")
          note("Sample includes blank but applicable and missing responses.")
	  export(table.docx)

Table 1


                       
 
  Frequency   Percent
 
 
 
Health status          
 
           
  Excellent            
 
      2,407     23.25
  Very good            
 
      2,591     25.03
  Good                 
 
      2,938     28.38
  Fair                 
 
      1,670     16.13
  Poor                 
 
        729      7.04
  .                    
 
          2      0.02
  Blank but applicable 
 
         14      0.14
  Total                
 
     10,351    100.00

Source: NHANES II
Sample includes blank but applicable and missing responses.
(collection Table exported to file table.docx)

We created a reproducible table and exported it to a Word document in one line of code.

Customizable ANOVA tables

anova and oneway now store their results in the r(ANOVA) matrix. In addition, there is now built-in support to collect these results, so you can build, customize, and publish ANOVA tables in your reports.

Consider the following one-way ANOVA that models the average weight of apples taken from trees treated with varied amounts of fertilizer. Below, we load the data, fit the model, collect the results, apply the new anova collection style, and replay the default layout of our customizable ANOVA table.

. webuse apple
(Apple trees)

. oneway weight treatment

. collect get r(), command(oneway)

. collect style use anova, replace

. collect layout

Collection: default
      Rows: term
   Columns: result[SS df MS F p]
    Tables: cmdset
   Table 1: 3 x 5


               
 
       SS   df        MS   F statistic   Prob>F
 
 
 
Between groups 
 
  5295.54    3   1765.18         21.46   0.0013
 Within groups 
 
   493.59    6     82.27                       
 
 
 
         Total 
 
  5789.14    9    643.24                       

The new command() option of collect get provides command-specific result labels that are similar to those obtained when using the collect prefix.

The anova style can be applied to results collected from oneway or anova to automatically create an ANOVA table. We could use additional collect commands to further customize this table and export it.

Collections with tabulate

The tabulate, tab1, tab2, tabi, and svy: tabulate commands are frequently used for creating tabulations and tables of summary statistics. Beyond simple tabulations, tabulate provides measures of association, and svy:tabulate provides various tests for survey data. You can now easily build tables that you can customize, style, and publish in reports from the results of these commands. Simply add the collect option to instruct Stata to create a new collection that you can customize with the collect suite of commands.

Below, we create a one-way table of frequencies for regions in our census data. We use the sort option to show the table in descending order of frequencies, and we use the collect option to put those frequencies into a collection named Tabulate.

. sysuse census
(1980 Census data by state)

. tabulate region, sort collect


     Census 
 
                                   
     region 
 
      Freq.     Percent        Cum.
 
 
 
      South 
 
         16       32.00       32.00
       West 
 
         13       26.00       58.00
    N Cntrl 
 
         12       24.00       82.00
         NE 
 
          9       18.00      100.00
 
 
 
      Total 
 
         50      100.00            

We type collect to see that we now have the Tabulate collection in memory.

. collect

Collections in memory
Current: Tabulate

 
    Name   No. items
 
Tabulate   14       
 default   34       
 

collect layout shows us the default table layout for the collection that was created by our tabulate command.

. collect layout

 
               Frequency   Percent   Cumulative
 
Census region                                  
  South               16     32.00        32.00
  West                13     26.00        58.00
  N Cntrl             12     24.00        82.00
  NE                   9     18.00       100.00
  Total               50    100.00             
 

If desired, we can further customize this table and then export it to the file type of our choice.

Query the layout

To see the layout specification and the table it creates, you can type

. collect layout

However, you may want to see the layout specification without seeing the table. You can now type

. collect query layout

Remove results from a collection

With the collect prefix and the collect get command, you collect results from commands that post their results in r() and e(). You can even collect custom results you define using expressions.

Now you can remove results from your collection with collect unget.