There are many ways to look at and take advantage of data. One of the most important is to segment it properly. That’s where cross tabulation comes in.
It makes the data you have more useful while ensuring you’re taking feedback about things from the right person. In this guide, you’ll learn what cross tabulation is, see a few examples, and figure out when to use it for optimal impact.
Cross tabulation defined
In its simplest form, cross tabulation is a research analysis method that directly compares and evaluates the relationships between two data groupings.
The reason why this is important is that there are many nuances to data – especially when it comes to behavioral patterns or opinions. Survey data that’s presented in aggregate may miss these insights because you’re looking at an average.
For example, if you have multiple age groups in your pool of respondents you may find that 75% of them trust online reviews. Using cross tabulation, it may come to light that 98% of those aged 24 – 34 trust online reviews, 60% of those aged 18 – 24, and only 40% of those aged 35 – 44 trust online reviews. If your main demographic group are those aged 35 – 44 then the aggregate data may lead you astray.
When to use Cross tabulation
Since cross tabulation is a quantitative research analysis method, two things need to be true before it can be used:
- There have to be independent variable
- Those variables have to be quantifiable
In a more practical sense, you wouldn’t be able to use cross tabulation with a long answer or short answer question. Conversely, you’d be able to use it with almost any type of multiple choice question (yes or no, rating scale, choose one, etc.).
Let’s go through a few examples.
Let’s say you serve multiple geographic regions and you send out a survey for product feedback. One question asks about the annual income of the respondent. Another question asks how they feel about the price of your product on a scale of 1 – 10. There are 1,000 total respondents.
From responses, it seems that most people are happy with the pricing but this is an aggregate sentiment. You can use cross tabulation to compare how each income bracket feels about the price of your product.
You may see that those in a lower income bracket think your price is too high. That’s to be expected but what if people in a higher income bracket also feel the same way? That’s an insight worth digging into more deeply.
The applications of cross tabulation are nearly inexhaustible and the better you’re able to use it, the more impactful your decisions will become. Keep in mind that it’s best used when comparing data that seems unrelated on the surface like gender and product liability for products that don’t favor either gender.
More uses of cross tabulation
- Customer churn surveys – Correlation between pricing plans and churn rate or feature loss and churn rate, etc.
- Employee evaluations – is there a trend or correlation between how well your employees are performing and other external factors?
- Product research – Does a certain gender feel a problem more acutely or does a certain income bracket feel less inclined to purchase this kind of product?
- Usage data – Are there any correlations between the features used and retention or frequency of usage and retention?
Practical example of cross tabulation
Let’s go through an example of cross tabulation in a set of data gleaned from a product survey with 20 respondents. It’s not statistically relevant but it’s enough for our purposes.
First, we want to discover who our most active users are. Not just in absolute numbers but also proportionally. What proportion of user groups uses the application daily, weekly, biweekly, etc.?
To do this, we’d compare data from business users and individual users against usage frequency. That would give us a basic cross tabulation table like the one below.
From there, we’d want to figure out the percentage of users in each group that used the application at different frequencies.
From the results, it can be seen that half of the business users logged in and did something with the application every day. They’re clearly the most engaged group. We can do something with customer satisfaction or likelihood to recommend.
Business users, again, have the most satisfaction with the application. This could be due to using the application more often and getting value from it.
The point is that it’s a powerful tool to drill down into your data and understand which segments are the most meaningful. You may be able to prove your own assumptions or be pleasantly surprised by what you find.
Benefits of cross tabulation
As you may have been able to tell from the few examples I used, cross tabulation is, for lack of a better word, epic.
This becomes more apparent when you’re working with larger data sets and more segments in each grouping of data. If the example above had 1,000 respondents then you could rely much more heavily on the data and make decisions.
Some of the key benefits of cross tabulation include:
Oftentimes, you may have an idea that something is true – an assumption – but cannot pin it down to specific causative factors. For example, your business may be growing at 20% MoM but you’re unable to tell what’s fueling that growth.
With cross tabulation, you may be able to find the source of your customers AND the types of customers that are signing up and driving the most growth for your business. From there, you can make decisions on what to do next or what to stop doing. The clarity you gain becomes more and more apparent as the data you use increases in complexity.
Allows you to get deeper insights
Raw data is often difficult to deal with. If you’re only looking at it in aggregate then the most important information may be missed. Cross tabulation gives you the ability to understand the nuances behind the raw data.
Instead of understanding that a lot of people like your products and services, you may learn exactly who likes them. Instead of having an understanding that males prefer your broadcasts, you may learn that middle-aged men are your core demographic.
More manageable data
Larger data sets are difficult to use – to say the least. With cross tabulation, you can focus on certain aspects of the information you have and work through it in phases. This will reduce errors while pulling out useful information.
Chi square analysis
Oftentimes, cross tabulation is used together with chi-square analysis to determine the statistical significance of the outcomes produced by a cross tabulation table. It’s a test that helps figure out whether the two groups or variables used in the cross tabulation table are independent or have a relationship.
If the variables are found to be independent then the data can be considered non statistically significant and given a null hypothesis. In other words, the conclusions drawn are not sound because the results gotten may be due to error.
Note: Null hypothesis means that the significant outcomes observed could just as well have been gotten by chance.
If the variables are found to be related then the data can be considered statistically significant then the data is considered sound. In other words, you can use the information gleaned to make decisions and be confident of the outcome.
If the probability of chance is at 0.05 (5%) or below, then it’s considered statistically significant. If the probability of chance is greater than 0.05 then it’s considered insignificant and should be treated as a null hypothesis.
As I’ve illustrated throughout this post, cross tabulation is a useful tool for drilling down into your data. It allows you to come out with deeper insights you can be confident in and quickly take action on.
Start with a few pieces of data or small data sets and try your hand at it. Over time, you’ll get a better understanding of what is possible and what isn’t. That’s when things will truly change for the better.
Let me know what you think of cross tabulation in the comments and don’t forget to share.