Qualitative Coding: How to Analyze Open-Ended Responses

by Kelly J. Larson | January 2016

Introduction
This document explains how to code open-ended responses (qualitative data) from surveys or other sources. To allow for meaningful and comparative analysis, each response should be reviewed and given a code or multiple codes based on similarities among the responses.

When referring to qualitative data, a code represents a word or short phrase that summarizes or attributes meaning to open-ended responses.[1]

The coding for each open-ended question should begin with “1” and continue for as many codes as needed to make sure the responses are appropriately considered when analyzing the data. For example, responses to the question “What do you like best about working at Questar?” might be organized with the following codes (or coding scheme):

  1. Coworkers
  2. My job
  3. Work environment
  4. Compensation
  5. Benefits
  6. Other

Once the responses are coded, the frequency of each code can be tallied and placed in a table, thereby allowing qualitative data to be measured and analyzed similarly to quantitative data.

Coding Process
The following steps outline the process of coding open-ended responses.

  • Step 1: Extract the open-ended responses from the raw data.
  • Step 2: Organize the responses in Excel. Each question and its responses should be in its own spreadsheet.
  • Step 3: Review each question’s responses and look for recurring comments and themes. Read each response carefully at least twice.
  • Step 4: Create a coding scheme for each question by listing the recurring responses and then numbering them, beginning with “1” up to a maximum number of reasonable codes.
  • Step 5: Apply the coding scheme to each response. Some responses may have multiple codes depending on how extensive and detailed the response is. Only code the most frequently mentioned responses. Code the least frequently mentioned responses as “other.”
  • Step 6: Have at least two other people read the coded responses to check for accuracy, consistency, and agreement among the reviewers.
  • Step 7: Create a filter in Excel to review the responses by code. This is helpful for reviewing the codes.
  • Step 8: Create code frequency pivot tables in Excel.

When applying the coding scheme, duplicate the responses that require more than one code. For example, a response to the question “What do you like best about working at Questar?” might be “my awesome coworkers and having a job I’m excited and passionate about.” If two of the codes from the coding scheme are 1. Coworkers and 2. My job, “my awesome coworkers” would be associated with one code and “having a job I’m excited and passionate about” would be associated with the other. Figure 1 shows what the response would look like in Excel.

figure 1

Once each response is given a code or multiple codes depending on the depth of the response, a filter can be applied to the Excel spreadsheet to review the responses by code, and pivot tables can be created to display the code frequencies and allow for comparative analysis of the open-ended responses. The following steps outline the process of creating a pivot table in Excel.

  • Step 1: Each open-ended question and its responses should be in its own Excel spreadsheet. The heading for the column containing the responses should be “Response” and the heading for the column containing the codes should be “Code,” as shown in Figure 2.
  • Step 2: Click on the “Insert” tab.
  • Step 3: Click on “PivotTable” on the far left.
  • Step 4: A “Create PivotTable” box will appear. Make sure all the responses and codes are included in the Table/Range (including the headings!). Place the pivot table in the existing worksheet. Click on a cell to determine the pivot table’s location within the spreadsheet. Click OK.
  • Step 5: A “PivotTable Fields” box will appear. Drag the “Code” field to the bottom left box labeled “Rows.”
  • Step 6: Drag the “Code” field to the bottom right box labeled “Values” to create the code frequencies. It might automatically appear as “Sum of Code,” but you must change that by clicking on it and then choosing “Value Field Settings.” Change “Sum” to “Count.” Click OK.
  • Step 7: Drag the “Code” field to the “Values” box again, except this time you will produce the code percentages. It might automatically appear as “Sum of Code,” but you must change that by clicking on it and then choosing “Value Field Settings.” Change “Sum” to “Count.” Click OK. Next, right click on any cell in the corresponding pivot table column labeled “Count of Code2.” Place your cursor above “Show Values As” and click on “% of Grand Total.” You can change the number of decimal places by right clicking on the column and choosing “Format Cells.” Click “Percentage” and change the decimal places to either one or two. Click OK.
  • Step 8: Change the table’s labeling from “Row Labels” to “Code,” from “Count of Code” to “Freq.,” and from “Count of Code2” to “%.” Format as needed.

Figure 2 presents a code frequency table in Word that was created using pivot tables in Excel.

figure 2

The next step after creating the code frequency tables is to summarize the results in a summary analysis report. Create charts when possible to display the results, include the code frequency tables in an appendix, and insert some responses in the body of the report as examples when presenting your findings.

[1] Saldaña, J. (2013). An introduction to codes and coding. In The coding manual for qualitative researchers (p. 3). Los Angeles: Sage Publications.

Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s