Crosstabs and banner tables

Award winning crosstabs and banner tables

There's a good chance you found us while searching for crosstabs software or banner tables software. That's how we became famous. StatPac is known as one of the best "tabs packages" available. It does more than most other crosstabs software but it's especially well-known for its ease in creating complex banner tables. 

 

Crosstabs is a way to breakdown responses to a question by another question. It's often used to show the answers to a question for each subgroup (stratum) of a demographic variable. The idea is to be able to identify differences between the different subgroups (e.g., are there differences between males and females in how they answered a question?).  Banners usually refers to the ability to breakdown a question by several other questions on the same page. It shows multiple crosstab tables simultaneously.

 

Banners is, without a doubt, the most valuable technique used by survey researchers because it reveals differences and relationships between variables. StatPac can show significance testing between percents and means in the table itself. It's the most efficient way to compare responses between different demographic groups. Decision makers use the results set goals and priorities. Resources are often allocated based on the information revealed in the banner tables.

 

All crosstabs and banner tables are not created equally

The ease (or difficulty) in creating the tables is the distinguishing factor among the various software packages. Most survey software programs have some crosstab capabilities, but they're all across the board in what they actually produce and how easy they are to create.

All crosstabs and banners programs are not the same

 

Some software packages will only crosstab one variable by one other variable. Others take you through a tedious menu to build each row and column of the table. StatPac's banners command makes it simple to generate hundreds of tables with a single statement. If you do a lot of surveys, this will add up to a huge time savings.

 

This page will show you how easy it is to use the banners command to create perfectly formatted crosstab tables. Options commands are shown to illustrate how they can be used to modify the appearance of the tables. The tables are exactly what StatPac created. They haven't been modified or doctored in any way to make them look better.

 

The format for the banners command is:

         Banners [stub-questions] By [banner-questions]

The stub refers to the questions down the side of the page and the banners are the questions going across the top of the page.

 

Example 1 - Creating several crosstab tables with a single command

 

The following command would produce five crosstab tables (one page for each stub variable). The stub would be questions 1 to question 5 and the banner across the top of the page would contain a total column and a variable named year.

        Banners Q1-Q5 By Total Year

See the report

Example 2 - Creating a crosstabs table with several questions in the banner

 

Questions can be identified with a variable number (V1 for variable 1, V2 for variable 2, etc.) or by using a variable name. Variable names can be anything that helps you quickly identify a question (e.g., Q1, Q2, Q3, Age, Gender, Income). V numbers and variable names can be used interchangeably in designing analyses. For example V25 might be the same as Gender.

 

An options command can be added to any procedure to control the analysis. Options can be selected from a menu or just typed.  One of the most common options is for paper orientation (portrait or landscape). When creating banner tables with many questions in the banners, you'll probably want to switch orientation to the landscape mode (OR=L) to fit more banners across the top of the page.

 

This crosstabs table has four questions in the banner. We use the orientation option to switch to the landscape mode (OR=L) so that all the banner questions can be displayed on one page.

 

     Banners Favorite By Total Register Gender Education Employed
     Options OR=L

 

See the report

Example 3 - How multiple options are used to control the survey analysis

 

Options are very powerful. They control how the analysis is performed (i.e., all the behind the scenes decisions and processes) and how the table is formatted on the page. You'll frequently use several options for a single analysis.

 

Options are often used to determine what will be displayed. For example, surveys often contain 1 to 5 Likert scale questions asking respondents to rate their level of agreement or disagreement to a statement. Mean averages for these kinds of questions can be printed by setting the automatic mean option to yes (AM=Y).

 

Another option can be used to turn on significance testing. Significance testing is used to highlight  differences between the columns in the banners (e.g., are males significantly different from females?).  Use the  significance tests option (ST=T) to turn on t-tests between percents and means. Statistically significant differences will be shown as upper and lower case letters in the table.

 

     Banners Prediction By Total Industry

     Options OR=L AM=Y ST=T

 

See the report

Example 4 - How many banner table columns can be shown on one page?

 

One of the most common questions we're asked about our crosstabs software is how many columns can appear in the banner. There's no set answer to this question because our software lets you set the font sizes, spacing, and column widths. These can be set individually, or even easier, by using the zoom factor option. The ZF option lets you reduce the size of everything in a table by a percent. For example, ZF=85 will reduce everything in the table to 85 percent of its original size. As you reduce the zoom factor, font sizes will become smaller, but you'll be able to fit a lot of banner points on one page. You can keep adding banner questions until the font size becomes too small to easily read.

 

This example is similar to the previous example except there are several variables in the banner (across the top of the page). The zoom factor is used to reduce the table to 70% of its normal size, so the entire table fits on one page (ZF=70).

 

     Banners Recreation By Total Years Children Age Gender Lake Distance Area

     Options AM=Y ST=T OR=L ZF=70

 

See the report

Example 5 - Options can be used to accommodate non-categorical data

 

Banner crosstabs are usually used for categorical data (i.e., the stub variable has response categories). However, sometimes you might want to do crosstabs when the stub variable is not categorical. In the following example, there are three parametric (non-categorical) variables on the stub. The data for those variables does not have categories, so we'll display means and standard deviations instead of counts and percents. The stub type option is set to parametric, which essentially means non-categorical (TY=P). Significance testing (ST=T) is used to identify statistically significant differences between the means.

 

     Banners Sales Wages Profit By Total Region

     Options TY=P ST=T

 

See the report

 

All the options in our crosstabs software can be selected from a menu

Use options to control the crosstabs and banner tables There are literally hundreds of options in the software. Fortunately, you'll only use a handful on a regular basis and memorizing them is quite easy. But while you're learning them, a menu is available to make it even easier. Just find the option you want to change and make the change in the menu. The options menu for banner crosstabs looks like this:

 

Select analysis options from a menu

 

(read more about StatPac survey software)