Content Analysis Guidelines

  • by Dr. Janet Waters 
    (revised 2017)

    Research Design:

    In a content analysis, you might be categorizing the content of a narrative or of a media, web, archival or aesthetic object. An example could be a content analysis of the frequency and quality of depictions of mental illness in popular culture, in film or television: you might wish to evaluate the accuracy of such depictions; the presence of violence in the character who is depicted as mentally ill; the SES, gender, culture & ethnic background of the character; the character's actions, & the outcome of the character's actions.

    Content analysis can also be done of art and of narratives. In past research content analysis has been done to classify and analyze the content of dreams, imagery, fiction and art. For example, researchers have analyzed the depictions of female characters in children's literature, and noted female characters were less frequent than male characters and were depicted with sex-role stereotyped characteristics (e.g. being emotional or nurturing) and in sex-role stereotyped occupations of relatively less power (e.g. as nurses or homemakers).

    Research Ethics for Content Analysis: Although content analysis is usually not done with participants, as with all research methods, make sure your research proposal has been approved by your instructor or supervisor before conducting your content analysis. Your research proposal must include what content you are collecting and analyzing, and if it does involve participant self-report (of their dreams, for example), how you plan to gather the data. If collecting participant self-reports, also include a copy of your consent form and always go over the consent form with your participant before they sign, and ensure their anonymity and confidentiality is protected in your research process (within the legal requirements).

    Data collection:

    Before conducting your content analysis, you should have prepared a checklist or coding sheet ahead of time. An easy way to do this is to first do a practice session to view the content, the art or film/tv show, or read the narrative. List categories of the content within your topic as they occur in your data. There are excellent coding systems already developed for content analysis that you can use; check out a good textbook on qualitative research in sociology or the social sciences. 

    For a general content analysis, you could use Hall & van de Castle's categories (used to analyze narrative self-reports of dreams). The general classes include characters, relationship references, objects, activities, events, places, & time. In the case of content like dreams or imagery, you might also code the content with respect to the overall quality of the experience, including valence (emotional tone), whether it is a memory/current concern/future extrapolation, whether it is realistic or fantastical, & whether it is tied to the immediate situation/reactive to the situation/irrelevant to it. For example, in terms of valence or emotional tone, artistic content or a respondent's protocol would be coded as negative in tone or valence if more than 50% of the references were negative in tone, or if the main or important references were negative when positive references occurred but were clearly secondary. If the protocol was a mixture of both qualities in the pair, both should be noted as present.

    For a specific topic, you will need to narrow down the list of all possible categories & qualities to only record the ones relevant to your topic. For example, in your content analysis of depictions of mental illness in tv & film, you might note categories & qualities of:

    a) the character - gender, age, SES, relationship with other characters (stranger, family, friend, subordinate vs superior, etc.). 
    b) the character's life circumstances - e.g. homeless or hospitalized, married or single, occupationally successful & functional, etc. 
    c) the character's actions in the story - e.g. aggressive & violent, submissive & helpless, etc. 
    d) the character's diagnosis (mental illness) - its symptomology, intensity or degree, the accuracy of the depiction of its symptoms. 
    e) the situation - danger or threat, etc. Note the outcome of the situation for the mentally ill character. 
    f) the type of story (if applicable & relevant) - e.g. comedy, romance, police drama, thriller, horror.

    Other kinds of content analysis would call for other types of categories. For example, an analysis of the characters in a Disney movie would probably include categories a to c above, & other categories as well (for example gender role stereotyping rather than character diagnosis).

    Having identified the content categories you will be coding, prepare a checklist based on your practice session. Your checklist should have room in each cell for comments. Make sure you include space for additional categories, as they occur during your observation.

    Then review the film/t.v show or art, or read the narrative, newspaper or document that you are analyzing, making note of occurrences of the categories on your checklist as they occur. Usually you will have to review the content two or three times to complete your coding. 

    Data analysis:

    The first step is to code the data. Having checked the occurrences in your checklist, categorize them further if you need to, & count each subcategory.

    Then you may be able to do descriptive statistics, presenting means, for example, in a table of the content. If you have taken Statistics, you could do a Chi2 of the frequencies of the categories or an ANOVA if you are comparing differences. (With the help of Excel these statistics are easy to do).

    Presentation of your results:

    Use the standard APA style research report format. In the Introduction, briefly review past research & theory in your topic question (e.g. summarize current research or debate on your topic - on depictions of the mentally ill, for example). Use APA referencing style to cite your sources. The Method section would vary according to the content you are analyzing. If you are analyzing participants' dreams or imagery or art, you would have a Participants section to describe their number, gender, age, etc. Otherwise, if you are simply analyzing the content of aesthetic products like film, television shows, or books, you may only have Materials & Procedure, since you don't have any participants. In the Materials section, list the content you have analysed, the fiction, film, etc. you may have used. In the Procedure section, note that your general research strategy was a content analysis, & describe your methods of data collection.

    In the Results section of the report, present your results in both a table (which would present the frequencies of each type of content) & in words. If you have done a statistical analysis, present those results as well.

    In the Discussion section, relate your results to past or current research & theory you had cited & described in the Introduction. For example, you could compare the percentage of occurrences of violence in mentally ill characters in the t.v shows you analyzed to the true percentage of violence in mentally ill individuals, according to the research literature in Psychology. You would also discuss the accuracy or inaccuracy of the depictions of the mentally ill in popular culture from your findings, & relate your findings to discussions among other sources (such as your text) of the possible effects of these inaccurate & extreme portrayals.

    © Janet Waters, 2017