22 Feb Cross Tabulations, the workhorse of Data Reporting
Data tables are the workhorse of data reporting for the vast majority of polls and surveys. It’s also the go to data format analysts and market researchers still use as the starting point to analyse and report data. It’s not hard to see why. Anyone used to working with spreadsheets, and using Excel, will be more or less comfortable with using data tables, even if you are not a data scientist. And for researchers, the humble data table provides a convenient place to dive in and review large amounts of data.
What are data tables?
Data tables are any set of data laid out in tabular format. Typically, the attributes (questions and answer responses) are set out in rows, and in columns with the columns showing key breaks in the data, such as gender or age.
Here is a basic example of a cross tabulation. We are showing the answer to a question, cross tabbed by gender.
How do we read a table? Most people instinctively make sense of a data table. Questions and answers are shown in the rows, and sub groups of data for comparisons are in the columns. The reader simply looks down the columns. So 66.2% of total sample say they save regularly. The next column indicates 67.3% of males save, as opposed to 65.1% of females. It is usually a good idea to show the sample sizes in the columns.
Why do we use tables and cross tabs?
Standard tables provide an easy way to review a lot of data and provides the basis to analyse and report. Tables and cross tables also underpin analytical hypothesis. We might have a hunch that results may vary by subgroup, such as age and gender. The cross tab is a quick way to look at key possible variations by demographics or crossed by other factors.
Better, simpler tables
To simplify the table we could de-clutter it further, by reducing the amount of detail. Below, we have stripped out the % in each cell, displayed the data as a significant figure without one decimal place and removed the sample size data. This should make it easier and quicker to view and understand the data. These changes are reducing clutter which arguably were taking up “brain bandwidth”. The columns now have 70% fewer characters for the brain to process, so we should be able to read and process quicker with this simpler format.
In tables where we have many answer codes or rating scales there can be merit in combining some of the categories. For example, the 5 options Very good, good, fair, poor, very poor becomes 3 with top and bottom box summaries, as below:
Of course, table format is also a matter of taste. Some people prefer to see as much data as possible. It’s also the case that if you are only receiving data outputs as a data table, there is a logic in accessing the data as one large bundle of data.
Standard Tables and Online Reporting
The output below is typical of the core reporting many clients opt for: the standard data table, which shows all the questions and key breaks (e.g. age, gender, region) in one block or cube of data.
Because the tables are hosted online in a secure reporting area, clients can access the data online. Although we can include a lot of data in an online hosted standard data table, and this may look like an intimidating block of data, compared to the previous version of the data table (published on paper in a huge stack of pagers), the modern version of the data table is very accessible. Compared to an excel export, the online version is superior.
Online reporting offers the ability to perform DIY analysis, such as filtering the view of the data outputs, to see fewer subgroups.
Having deselected certain age groups, we can now see fewer columns and easily compare certain subgroups of interest.
This is one of the benefits of accessing data in the key findings area of our data portal. You can also download tables in excel or pdf format to read and share offline. If charts have been created, they would also be available to read and download.
This is a quick introduction to data tables. We will return to the subject in the future to look at reporting options and how best to access and report them.
Further customisation options for cross tabs
Other ways to enhance the reporting of tables include:
- Stats tests
- Average scores
- Conditional formatting (e.g. to highlight different cells by statistical significance)
From Data Tables to Enhanced Tables and beyond (Data Visualisation)
So, we have established that data tables are the core of data reporting, and that there are different ways to customise and present tables. Highly customised, enhanced tables require more time to create, but in principle offer many possibilities to reduce complexity of data and uncover key insights in data.
Summary Data Reports
While data tables offer the ability to detail large amounts of data, they can take time to sift through to pick out the story. One approach is to combine aspects of cross tabs with extracts of data in chart form. Summary Data Reports are not a full written report, but a topline of key aspects of a data set. They typically will report on summary (Net data) and some sub groups (e.g. age and gender) and will include some summary analysis. They can be accessed as a powerpoint, pdf or hosted online.
Surveygoo offer a range reporting formats, including customised cross tabs, charts, and online reporting tool such as interactive reports, dashboards and infographics. We will look at some of these formats in future blogs, but in the meantime, do get in touch if you want to explore the alternative reporting formats.