Subscribe Bookmark



Jun 23, 2011

Columns by Categories in Tabulate

What is "Columns by Categories" ?

Survey data often have sets of questions that have common sets of responses, example, "ye", "no".

Tabulate can summarize the responses for those questions and present the result very compactly,

with the column names as row labels, and the answers as column headers, or vice versa.


You don't even have to do anything special. If the columns are tagged as nominal or ordinal,

dragging and dropping the columns into the "drop zone for rows" will give you the table as shown.

Making a table of "Columns by Categories" in most cases gives a more compact table than

making tables with the columns as separate grouping columns.


For "columns by categories", you are not limitted to counts.

In the combined tables, you can get summary statistics, broken down by the responses.

To do that, just drag the desired analysis columns and drop them onto the body of the table.

[Missing image]

Note that the default statistics is sum.

To change that to the mean, just drag the keyword "Mean" and drop it onto the word "Sum".

Alternatively, you can right-click on the word "Sum",

from the popup menu, select "Mean" from the statistics submenu.


You can further break down the statistics by one or more grouping columns.


To get the above table, I dropped "Gender" on top of the word "Disagree".

And of course, you can ask for the aggregate statistics for the grouping column Gender.

You do that by dropping the word "all" on "Gender".


Grouping columns can go to the side too.

Click and drag "Gender" away from the top of the table and drop it above the first column,

"I am working on my career", you can get a different layout of the same table.


Article Tags