TEXT ANALYSIS:
PRACTICAL
CONSIDERATIONS
Miriam Matejova, PhD
November 14, 2023
Agenda
1) Critical discourse analysis
2) Content analysis
Recall: discourse
◦ Michael Foucault – discourse denotes the way in which a
particular set of linguistic categories relating to an object and the
ways of depicting it frame the way we comprehend that object
◦ The version of an object comes to constitute it
◦ A discourse is constitutive of the social world that is a focus of interest or
concern
Language shapes the way we think
Critical discourse analysis (CDA)
◦ Emphasizes the role of language as a power resource
◦ Discourses should be examined in relation to social
structures, including the power relationships that are
responsible for them
◦ Aims:
◦ Understand how meaning is created in context
◦ How language is used to exercise power (reveal power
relationships)
◦ challenge the use of power
Critical discourse analysis: how to?
◦ Seeks to reveal the meaning of a phenomenon by
exploring how:
◦ the discourse came to have a particular meaning today;
◦ the discourse draws on and influences other discourses;
◦ the discourse is constructed through texts;
◦ the discourse gives meaning to social life and makes certain
activities possible, desirable or inevitable;
◦ particular actors draw in the discourse to legitimate their positions
and actions
CDA: steps
◦ Analysis of a discursive event in three
steps:
1) examination of actual content,
structure and meaning of the text
under scrutiny,
2) examination of the form of discursive
interaction used to communicate
meaning and beliefs,
3) consideration of the social context
in which the discursive event is
taking place
Qualitative content analysis
◦ Two general strategies: analytic induction
and grounded theory
◦ Analytic induction: a rough research
question, a hypothetical explanation,
collection of data
◦ Grounded theory: development of theory
out of data
◦ iterative approach – data collection
and analysis proceed in tandem,
repeatedly referring back to each other
Qualitative data analysis: basic operations
◦ Coding as a starting point
◦ Of what general category is this
item of data an instance?
◦ What does this item of data
represent?
◦ What is this item of data about?
◦ What question about a topic
does this item of data suggest?
◦ Results in fragmentation of data;
data taken out of context
Quantitative content analysis
◦ Approach to the analysis of documents and texts that
seeks to quantify content in terms of predetermined
categories and in a systematic and replicable manner
◦ Objective and systematic
◦ Rules clearly specified in advance for the documents/text to
categorize
◦ Transparency in procedures
◦ Application of rules done in a consistent manner to minimize bias
Content analysis: examples of research
◦ Visual images of women’s and men’s
magazines (Malkin et al. 1999);
◦ Politicians’ speeches (Proksch and
Slapin 2010)
◦ Obituaries (Fowler and Bielsa 2007)
◦ Alcohol brand references in the lyrics
of popular songs (Siegel et al. 2013)
◦ Social media posts and comments
(Beullens and Schepers 2013)
◦ The content of websites and blogs
(Davis et al. 2015)
Step 1: choose your research questions
◦ Clearly formulate your research questions
◦ Questions of: who (gets reported), what (gets reported), where
(does the issue get reported), location (of coverage within the
items analyzed), how much (gets reported)
◦ Notice: What is not important to reporters? How much does the
coverage change over time?
Step 2: select your sample
◦ What type of media to sample? How many pieces?
◦ What time period(s) to focus on?
◦ But could have an ongoing general phenomenon – random
sampling
Step 3: decide on what to count
◦ Consider different units of analysis
◦ Significant actors: Who has produced the item? Who is the main
focus of the item? Who provides alternative voices?
◦ Words: counting of frequency with which certain words appear,
the use of some words rather than others
◦ Can use computer-assisted content analysis (e.g., dictionaries)
◦ Subjects and themes: search for both manifest and deeper,
hidden content
◦ Dispositions: are journalists favourable or hostile towards a topic?
Reporting slant, tone
Step 4: coding
◦ Two main elements: design a coding schedule and design a
coding manual
◦ Coding schedule: a form onto which all the data relating to an
item being coded will be entered
◦ Coding manual: a statement of instructions to coders
◦ Must do it even if only coding yourself
Step 5: intercoder reliability
◦ “The extent to which independent coders evaluate a characteristic of a
message or artifact and reach the same conclusion” (Lombard et al. 2002)
◦ Critical component of content analysis – without it the data and
interpretations of the data cannot be considered valid
◦ Need at least two coders
◦ Many different techniques but only a few regularly used
◦ Percent agreement - percentage of all coding decisions made by coders on which the
coders agree
◦ Krippendorff’s alpha - designed to be used for variables at different levels of
measurement; accounts for chance agreements
Content analysis: advantages
◦ Very transparent research method; objective method of analysis
◦ Can allow tracking changes over time
◦ Unobtrusive method – does not entail participants in a study
having to take the researcher into account
◦ Flexible – can be applied to wide variety of unstructured
information
Content analysis: disadvantages
◦ Can only be as good as the documents on which the researcher
works (authenticity, credibility, representability, meaning)
◦ Impossible to devise coding manuals that won’t require some
interpretation on the part of coders
◦ Sometimes difficult to accurately capture latent content
◦ Can’t answer the “why” questions