Anatomy of a Crosstab
Are you new to market research or do you need a refresher course in basic survey tabulation principles? The following guide will help you better understand how to read and interpret the results of your survey project.
A crosstabulation (or "crosstab") is a table displaying survey results. It resembles a spreadsheet with multiple rows and columns, typically where the rows are tabulated survey responses and the columns are subgroups for analysis. Survey data can be analyzed in a variety of ways, but basic crosstabs remain a powerful tool in exploring your data. The best way to describe a crosstab and its components is by looking at an example.
Key to below Crosstab:
Typical banner points are demographic breakdowns or other classification questions like overall satisfaction levels. The banner points are the individual columns which comprise the banner.
Title lines are typically either the actual question text or a paraphrased description of the table contents. Any bases or filters applied to the table or other descriptions may also be labeled here.
The banner (or header) consists of the subgroups for data analysis. Usually all tables are run by a common banner of up to approximately 20 banner points.
The stub contains the table rows. These rows typically include the questionnaire responses choices, plus any additional summary rows, nets or statistic rows.
The summary rows shown here are Total (all respondents) and No Answer (those missing from the table or who did not answer the question). Also sometimes shown here are Total Answering, Total Mentions, or other summaries.
The table cells may be frequency counts, row or column percepts, or other data values. Each cell in this table shows the frequency count and vertical (column) percentage (based on total respondents).
This table shows that 182 respondents or 36% of the total sample gave a rating of excellent.
This table cell shows that 70 respondents or 47% of those aged 35+ gave a rating of excellent.
Nets are groups of similar answers combined in the stub. Shown here are the combined high (top-2) and low (bottom-2) responses to the rating scale. A table for an open-ended question may show nets for likes, dislikes, or other answer groupings.
Where applicable, descriptive statistics such as the mean (average), median, standard deviation and standard error are often included on tables for analysis purposes.
Having a solid understanding of the basic elements of a crosstab will promote better communication with your tabulation partner and help you to get the most from your survey data. There are many options not described here which are also available when analyzing survey results. The next StarTips article will help round out some of the more advanced concepts of survey tabulation and develop specifications (a "tab plan") for your data analysis.