The Role of Coding in Survey Data Analysis

Now, you need to analyse the results. The closed questions (“What is your age?”, “How would you rate our service out of 10?” etc…) are easy to add up. The open-ended questions (“Why are you unhappy with our service?”) are harder to analyse because people answer these questions in their own words. And therein lies the rub: these questions are useful because they get to the “why” behind the closed questions, but their qualitative nature makes them hard to analyse quantitatively.

That’s where coding comes in.

What is Coding in Market Research? 

Coding is a data annotation process that systematically assigns themes to each of your survey verbatim comments. These themes are drawn from a defined list, called a Codeframe.

For example, you might have a setup like this:

By tagging each survey comment with the relevant themes it contains, you transform the data into a numeric form that can be easily counted up. In the example above, once the comments are coded it is easy to see that although this product is cheap people have issues with the taste.

Traditionally, this process of coding was a purely manual process (using pen and paper, if you go back far enough!). The manual process had its advantages as expert coders could spend time applying their expertise coding the data diligently, accurately and precisely.

The obvious downside is that the manual process is very slow – sometimes taking weeks to deliver, causing drag on the whole research process.

The Benefits of Automated Coding Tools

In the new AI age we live in, surely all of this manual effort is a thing of the past?  

Yes and no. AI tools can vastly speed up the process, but we have to be careful not to throw the baby out with the bathwater. There is no substitute for human judgement, knowledge and expertise. So, what is needed is something that combines the best of both worlds. You can read more about this in our other blog here.  

codeit achieves this balance via a three step “Extract | Refine | Apply” process.

Here’s how codeit's “Extract | Refine | Apply” process works:

Step 1:  Extract

codeit uses AI to extract the main themes and sentiments from a set of data.  It then autocodes your verbatim data for you – building a Codeframe and applying codes in a fraction of the time of the manual approach. 

Step 2:  Refine

However, it’s important we also allow for that all important human input in the process. 

So, codeit enables you to refine and curate the autogenerated output using an industry-leading customer user interface.  We call this “human-led AI”. 

Step 3:  Apply

Finally, codeit contains a unique a machine learning layer that automatically learns from the coded results you produce. When you add more data, this learning will be automatically applied to autocode your data – further improving efficiency and accuracy every time you have more verbatims to analyse. 

Many researchers—both in market research and corporate settings—are turning to automation to handle their verbatim data. The result? Faster, smarter insights without the painful grunt work. 

So, are you ready to make the most of your verbatim data? Sign up for your 30-day free trial today.