If you’ve just completed a large surveying activity, forum conversation or just have a large amount of written feedback and comments to make sense of, then you might need to utilise qualitative analysis techniques.
Unpacking large amounts of qualitative data can be a daunting task but with a little preparation and some simple steps, drawing insights from you data can be made just that little bit easier.
In this article, we look at a simple process for organising and coding qualitative data.
The steps outlined below are especially useful if you have thoroughly planned your projects prior to engaging with your community.
I highly recommend our recent webinar with Dan Popping for a good overview of planning for online engagement.
4 simple steps To Do Qualitative Analysis
Step 1: Gather your feedback
The first step towards conducting qualitative analysis of your data is to gather all of the comments and feedback you want to analyse.
This data might be captured in different formats such as on paper or post-it notes or in online forums and surveys, so it’s important to get all of your content into a single place.
For this activity, you might consider a master spreadsheet as a place to collect all of your feedback or you might have other digital tools such as EngagementHQ to help you organise your content.
As part of organising your content you want to setup your analysis template. If you are using a spreadsheet you might consider using variables as seen below to help get you started.
In the first column, you can see a field for data source. This variable will allow you to filter through your responses to compare views collected via different means.
Next, you can see a stakeholder type variable, which comes in handy for drilling down into different stakeholder groups which you might need to report on.
Obviously, you are going to need to have a field in your data for the actual feedback you collected and you might label this field “feedback” or “qualitative data”.
The most important variable required for your dataset is the code field which you will use to code and organise you data in the next step.
Finally, you can also include an identifier for the question the data was collected for to further help you drill down into your insights.
In your master data template you can also include multiple columns for collecting your coding and you might also be required to add any demographic fields you have captured.
You should allow a suitable amount of time for organising your data, especially if you are collecting it and entering it from a variety of sources.
Step 2: Coding your comments
The next step in this process is about coding your comments and most importantly reading and making a decision about how each one should be organised.
There are two ways to approach this;
The first way assumes that you are looking for a pre-defined set or list of issues or themes, whilst the other method is focused on unpacking themes without having any prior expectations about what they should be.
For the first method, it’s crucial you articulate your coding legend.
A good way to do this is to create a simple table outlining what each code is and what it covers.
This can be mapped to the areas you need to report on or the key components of your project.
In this legend, you can outline your theme and description and if you want to take it a step further you might even add issues as a secondary tag within a theme.
If you decide to do the alternate method and unpack your qualitative data to try and derive themes for your code list, you are going to need to read a sample of your comments.
We recommend reading at least 25% of your comments and making a first pass judgement about where each piece of feedback might sit.
Once you have completed this you should ask a colleague to read through the same sample and check to see if they agree with your coding.
When this is complete, refer to the themes you have identified and complete a coding sheet as per above.
Regardless of which option you choose, you will be required to read through your comments and make some decisions about them.
Complete your coding by reading through each comment, using legend will be your guide.
If anything falls outside of your coding list, simply mark it with an identifier such as “unsure” and come back to it later.
It’s important to note, while there are now AI and machine learning applications which can quickly scan qualitative input and provide insights for you, there is nothing better than intimately knowing what your community thinks about your project.
Intimate knowledge of the feedback is often missed with automated tag clouds and sentiment analysis and it can encourage lazy practice and unintentionaly lead you to jump to incorrect conclusions.
This coding process can also be completed using EngagementHQ’s comment analysis tool, allowing you to digitally code comments and feedback across all nine engagement tools, including essay and single line text questions in surveys.
Step 3: Run your queries
Once you have coded all of your data, it is time to run your queries.
In essence, this means looking for insights in your data.
Your reporting requirements will determine the extent and type of the queries you run during this step.
Below are some recommended queries to run on your data:
- Which are the most used codes or themes? Represent this visually to get a sense of the most important areas.
- How did people respond via different formats? Were there any differences in views based on the submission type?
- Which issues are of most concern to different demographics segments?
- Are there any relationships between issues? ie. are people who are concerned with one issue more likely to be concerned with another?
Once you have run your queries and explored your data you should have a good foundation and enough insights to begin your reporting.
Step 4: Reporting
The final step is reporting on your findings.
This is a critical step as it’s your opportunity to tell the story of what you learnt from your consultation.
If you fail to do this step well, your community will absolutely lose faith in your process and you might even face potential community outrage.
Being transparent and timely is the best way to avoid this situation.
Use your insights to create a narrative about the issues and opportunities which your community have identified.
When framing your insights you might consider using the following as a useful way of talking about and quantifying your findings;
- “Participants frequently raised concerns around the flexibility in defining ‘sympathetic additions’ with some participants suggesting that altering heritage buildings in any way could impact the character of the neighbourhood.”
- “Many participants commented that a three storey maximum height was preferred for the eastern precinct of the Futureville Village, rather than four.”
- “Most participants suggested a three storey preferred height limit in Howitt Street, rather than four. The main concern was that this would affect the ‘village atmosphere’ of the street.”
You should also include relative charts and visuals to help your community further explore your data
Generate these in a spreadsheet or other data visualisation software application such as Microsoft Power BI, Google Studio or Qualtrics to name a few.
Once you have compiled and circulated your report, it’s good practice to again ask your community for final comments and input.
At this stage, you can test whether you have framed their concerns and issues correctly and allow yourself to make and final changes before you submit a final report and make your decisions.
As you can see, these four steps provide a simple process to follow for organising your data, determining your coding tables, running queries and reporting on your consultation.
Make these steps a part of your project planning process and ensure you always have an end to end picture of how you are going to collect and report on your data before you begin your consultation.
Image Credit: Cam Adams