Using Spreadsheets for Qualitative Coding
Qualitative Data Analysis (QDA) software packages offer powerful tools for qualitative researchers in the coding and analysis processes. However, they also come with a price tag that not all researchers can afford. Spreadsheet software packages offer a cheap or free alternative to reproduce the most important QDA package function: queries.
While coding in a QDA package essentially involves the same manual process as doing coding by hand, the main reason I recommend QDA packages to qualitative researchers is queries. Say you coded interviews with a group of unemployed individuals about their job search strategies including a subset of codes that identified different kinds of strategies. With a query, you could look at the relationship between different kinds of strategies and identify the packages of strategies that your respondents utilized. In addition, you could even look at these patterns by characteristics of your respondents such as length of time unemployed, prior work experience, race, or gender. Queries allow researchers to slice and dice their coded data to look for cross-cutting themes.
The same types of inquiries are possible in spreadsheet software packages like Excel, Google Sheets, or Numbers through two common functions. First, you can use pivot tables to produce descriptive statistics about your coding to help you identify key themes for further exploration. For instance, for the project described above, you might produce a pivot table that tells you the number of respondents who reported each job search strategy. This would help you identify the most and least frequently used strategies. Depending on how you organized your spreadsheet, which I discuss further below, you could also produce counts of the different combinations of strategies that respondents used.
From there, you could apply the second function: sorting. Simply speaking, sorting is reorganizing your data based on the values in one or more columns. A one-column sort would reorganize your spreadsheet by alphabetizing or ordering numeric values in the specified column. While you may be familiar with why this is useful for reordering a list of names in alphabetical order or zip codes in numeric order, using this function with coded data is particularly useful for sorting based on more than one column. For instance, sorting by job search strategies and length of time unemployed would allow you to explore whether respondents who have been unemployed for shorter periods of time use different job search strategies than those who have been unemployed for a longer period of time.
In order to implement these strategies, you need to think carefully about how to organize the codes in your spreadsheet. Each column will represent a code, but there are two different ways to approach this. You could include a column for each code including any subcodes and indicate the presence or absence of the code or subcode by typing 1 or X to indicate presence and 0 or leaving the cell blank to indicate an absence. For the example above, this would mean having a column for each job search strategy such as looking for job ads online, talking to friends and family about openings at their jobs, and visiting local businesses to inquire about openings. With this approach, identifying packages of strategies for job search would require sorting across multiple columns, but the results of a pivot table would be much clearer with a count for each individual strategy.
Alternatively, you could use one column to represent the code and enter all relevant subcodes that apply using a number or letter that represents each subcode. In this case, our example would have one column for job search strategies in which you would enter a number or letter that represents each subcode such as 1 for looking for job ads online, 2 for talking to friends and family about opening at their jobs, and 3 for visiting local businesses to inquire about openings. In the analysis stage, this would allow you to sort based on one column to look for packages of strategies, but would only allow you to produce counts of packages of strategies with a pivot table.
Applying pivot tables and sorting to spreadsheets is a way to reproduce the power of queries without a QDA software package. To read more about using spreadsheets for qualitative coding, take a look at the Contextual Text Coding approach, which provides a detailed and systematic approach to large-scale qualitative data analysis without QDA software.