Constant comparative method–I totally get it, now.

But, I’m not an expert or anything. It’s a detail-oriented, time-consuming process.

So many details...

I recently did a content analysis of 2010 election speeches using the constant comparative method (CCM) (Glaser, 1965). Prior to this analysis, I read extensively about the CCM, but I still wasn’t exactly sure how to employ it. I’m a kinesthetic learner, so it’s often best for me to just jump right in and try something as soon as I have a general idea of how to do it.

I did not use any fancy software. I coded my documents by hand and used MS Word to tally the results as I went. After that, I used MS Excel to sort and clean up the data. This is a time-consuming method of analysis (how many times am I going to say that?). It took well over 10 hours to code two-20 minute speeches. I rechecked one percent of the data, just for verification.

I proceeded in this fashion:

  • Every word was considered, but not every word was coded. Conjunctions, filler words such as “however,” weak verbs, and other words determined not to add meaning were discarded.
  • All pronouns were coded with their actual name. For example, if the word “they” was used, the word was coded with the actual name—fundamentalists, terrorists, illegal aliens, etc.
  • Similar words were grouped and counted together, but the differences were noted (vote, cast your ballot, check that box, etc.).
  • If the meaning of words was metaphoric, the entire metaphor fragment was noted in the Metaphor category.
  • Words were counted only once. If the words made up a metaphor, they were not included in the general word count.
  • After counting the words, the data was checked for errors by recount. One percent of the data was recounted.

About pronouns–the use of pronouns may have importance beyond the word count (Hahn, 2003; Krippendorf, 2004; Stemler, 2001), so you’ll need to decide as you go (emergent) if you’ll note their proximity in your research. In my research, I found the pronouns got in the way a bit–because, who is “he?” In my analysis, I wanted to know exactly who was the subject of the speech. But, because pronouns act in proximity to other words, an analysis of pronouns themselves can be revealing also. As I mentioned above, in my research, whenever I came upon a pronoun, I replaced it with a proper noun (terrorists, electorate, etc).

My results were very revealing. I think the emergent nature of the CCM is key to a content analysis of this type. If you have the freedom to create categories as you go, you can find things you may not have been looking for–there are little surprises you find along the way.

If you have any questions about it, feel free to post in comments. I’ll try to answer them.


Glaser, Barney G. “The Constant Comparative Method of Qualitative Analysis.” Social Problems 12, no. 4 (1965): 436-445.

Hahn, D. (2003). Political Communication: Rhetoric, Government, and Citizens (Second ed.). State College, PA, USA: Strata.

Krippendorff, K. (2004). Content Analysis: An Introduction to Its Methodology. Newbury Park, CA: Sage.

Stemler, S. (2001). An overview of content analysis. Practical Assessment, Research & Evaluation, 7(17). 1.

This entry was posted in politics, school and tagged , , , , , , , , , , . Bookmark the permalink.

8 Responses to Constant comparative method–I totally get it, now.

  1. KB says:

    On metaphors (posting here because 140 characters are not enough) :

    I use metaphors for an expert audience and only for describing how material (such as metals Aluminum and Copper) is flowing during different conditions. For example: relating the flow of aluminum to the way water flows when someone jumps into a pool; “With these weld parameters the aluminum flows like water as something is dropped in.”

  2. martin chapman says:

    Hi, im doing my 10,000 dissertation at the moment and much like you am using a constant comparison method of data analysis. The design is qualitative using phenomenology to display the narrative of the participants, which were interviewed in groups of two (focus group).

    At the moment when im coding, everything seems to be jumping out at me and I cant seem to narrow down the themes that area clearly there, each seems to blend into the other.

    My subject area is looking at neurological conditions and how people with these problems are dealt with when admitted to the NHS, both by doctors and administration staff. What we intend to develop is a card/passport with the individuals needs and preferences on, in order to inform the NHS of personal and client centred needs.

    Can you please give me some help when considering how to approach this, clearly I have gone to my supervisor, however, she is of little help.

    • Hi Martin,

      Thanks so much for your comment! I do have some things to share, but I’ll need to reply tomorrow. Hope that’s not too late for you! My deadline is tonight, midnight and I’m still furiously in the invention stage of some of my writing.

  3. Hi Martin,

    I’m not familiar with phemenology. I used content analysis.

    Here’s how I did it: I went through the document word by word and just counted the words as I went. At the end, I may have had 7 counts of direct_certification, 5 counts of directly_certify, and 2 of direct_certification_process. I would look at it as 14 counts of direct certification. I sorted the data according to which words were used more often.

    At the end, I looked over the data because the constant comparative method is emergent. That means, you need to look over the whole body of data, and then apply categories. Then, you do it again to further break down the data. You also want to take extensive notes and write down exactly what you’re doing. These notes will help with the data verification process.

    Word counting is probably not a part of phemenology, but you might try it because words make meaning, and sorting by meaning is the goal of the constant comparative method. Word count helped me understand what parts of the data were actually important. High counts told me that was a main theme. So, even if you’re not using content analysis, you might try doing just a count of similar words–similar nouns and verbs. I know word count might not be important to your analysis, but it might help categories “emerge.”

    The emergent part of the constant comparative method is one thing that makes it powerful. So, that’s an idea I have for you. Let me know how it goes.

  4. stratumcorneum says:

    Hi.. Thanks for your writing,
    Got some question to ask you, could you please tell me what is the different between constant comparative and content analysis. I know that the CCM is being used to analyse data for grounded theory research, how about content analysis? Thanks for your reply

    • Qualitative content analysis is a powerful tool of scholarship used to divide a text into discernable, countable categories, so conclusions can be drawn; the Constant Comparative Method allows for emergent categories, allowing the information to shape discoveries about a text.

      So, content analysis is a way to gather usable data about a text. And, the constant comparative method is a way to analyze and understand what the data means.

      I hope this helps.

      • stratumcorneum says:

        Thanks for your answer, It helps me to understand the two methods better. My advisor wants me to use the CCM to analyze my qualitative study, my study is just a descriptive qualitative study not as fancy as GT. Im developing a tool to measure care dependency in the elderly population, and Im using mixed method to develop the tool.
        Thanks again for your quick reply

Leave a Reply

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

You are commenting using your 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 )

Google+ photo

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

Connecting to %s