Transcript Basic Content Analysis

Content analysis
Marika Lüders /[email protected]
Structure of lecture
1. Texts and communication with a parallell to my lecture a week ago:
interviews as texts.
2. Qualitative content analysis
3. Quantitative content analysis
2>
Department of Media and Communication
Texts and communication
Interviews as texts
”For the qualitative-minded researcher, the open-ended interview
appearantly offers the opportunity for an authentic gaze into the soul of
another, or even for a politically correct dialogue in which the researcher
and researched offer mutual understanding and support. The rhetoric of
interviewing ”in depth” repeatedly hints at such a collection of
assumptions” (Silverman: 343).
Or perhaps Silverman reads too much into the intentions of researchers
conducting qualitative interviews?
3>
Department of Media and Communication
Texts and communication
Interviews as texts
Raw material in the form of transcribed qualitative interviews evidently
implies analysing textual content.
As such, we have already introduced textual content analysis:
The aim of conducting interviews: to discover the details and nuances of
lived realities. Themes, similarities, discrepancies. Yet always requires an
awareness of interpretation as part of the research process. Hence, never
direct access to the realities of the informants.
Even when using CAQDAS for analysing interviews. Researchers read,
re-read and code interviews and hence interpret.
4>
Department of Media and Communication
Texts and communication
Shared realities?
John Durham Peters (1999) Speaking into the air
Communication is something we do to take part and participate in the
creation of a collective world, they are not means that will ever help us
cross the fundamental chasm separating us.
”The problem of communication is not language’s slipperiness, it is the
unfixable difference between the self and the other. The challenge of
communication is not to be true to our own interiority but to have mercy
on others for never seeing ourselves as we do” (Peters: 266-267).
5>
Department of Media and Communication
Texts and communication
Shared realities?
Perspectives on communication easily romanticize communication as
facilitating sharing, communality and understanding between individuals;
closing the gap between subjects and transcending differences.
Interpretation implies the very likely chance of misinterpretation.
Does not mean we have to surrender to solipsism, only that we need to be
aware that communication, whether in the form of face-to-face
interactions or written texts, are always subject to interpretation.
6>
Department of Media and Communication
Texts and communication
Interpretation of texts through
history: the power of texts
S
M
BUT: we always interpret texts in contexts
TEXT APPROACHING PASSIV RECEIVER
7>
Department of Media and Communication
R
Texts and communication
Three minutes discussion
http://www.martinlutherking.org/
Discuss the meanings of this text.
Stuart Hall
Intended meaning
Negotiated meaning
Oppositional meaning
8>
Department of Media and Communication
Qualitative content analysis
Analysing texts
Transcribed interviews are texts
Silverman: text as a heuristic device to identify data consisting of words
and images that have become recorded without the intervention of a
researcher (348).
Content analysis, usually used to describe quantitative analysis:
establishing a set of categories and count the number of instances that fall
into these categories (described in detail later in the lecture).
9>
Department of Media and Communication
Qualitative content analysis
What is it?
Aim (from Silverman: 348):
Understanding the participants’ categories - see how these are used in
activities such as telling stories.
The process through which texts depict ”reality”.
Semiotics
Ethnographically oriented narrative analysis
Discourse analysis
In any case, limit the amount of data, as qualitative analysis (whether of
interviews or texts) implies conducting a detailed anlaysis.
10 >
Department of Media and Communication
Qualitative content analysis
Semiotics
SEMIOTICS (Saussure: semiology)
Saussure on the signification process:
SIGN
SIGNIFIER
SIGNIFICATION
REFERENT
external reality
SIGNIFIED
A sign can be defined as something that stands for something else and
is the physical vehicle of meaning in a language.
Arbitrary relationship: no intrinsic connection. Cultural convention.
We are active makers of meaning.
11 >
Department of Media and Communication
Qualitative content analysis
Semiotics
2. Roland Barthes’s Myth Today. 1957
Denotative meaning: Black soldier giving a
military salute
Connotative meaning: “France has a great
empire; all her sons, without distinction of
colour, serve faithfully under the French
flag and there is no better answer to the
critics of colonialism than this black’s zeal
in serving his supposed oppressors.”
12 >
Department of Media and Communication
Qualitative content analysis
Meanings of texts? (McCullagh)
1. No stable meanings
2. The contexts of texts as linguistically constructedd
3. As products of the discourses in which they were produced
4. The power of the reader
5. Summaries are always subjective
6. Any reading will be biased
7. Problems of hermeneutical circles
13 >
Department of Media and Communication
Qualitative content analysis
1. The uniformity of meanings
Always several different discourses being used, and discourses change
over time.
... a discourse is not a disembodied collection of statements, but groupings of utterances
or sentences, statements which are enacted within a social context, which are
determined by that social context and which contribute to the way that social context
continues its existence (Sarah Mills (1997) Discourse: 10).
The contexts of texts will limit the range of possible meanings. Means
researchers/historians have to look around the specific text in question.
Trying out possible meanings in a text according to context to look for
the best fit.
14 >
Department of Media and Communication
Qualitative content analysis
2. The contexts of texts
Basic meaning of the text/the text’s conventional meaning - uncertainties
resolved with reference to the context of the text. Consistent with what is
known of the author’s beliefs and attitudes, situation and behaviour.
Locating texts within specific social sites -> disclosing the political,
economic and social pressures that condition a culture’s discourse. The
reality of social contexts of texts.
15 >
Department of Media and Communication
Qualitative content analysis
3. Auhors with intentions?
the power of discourses
Texts as biased according to the author’s intentions?
Michel Foucault: the systematicity of ideas, opinions, concepts, ways of
thinking and behaving are formed within a particular context. The effects
of those ways of thinking and behaving.
Truth: the types of discourse it harbours and causes to function as true, the techniques
and procedures valorised for obtaining truth. Does not appear in transcendental ways.
Power: not a possession or a violation of someone’s rights (thus very unlike
conceptions of power within political science). Rather power is dispersed through social
relations. A productive model of power.
Knowledge: the result of power struggles.Whose version of events is sanctioned as
”true knowledge”
16 >
Department of Media and Communication
Qualitative content analysis
3. Authors with intentions
Yes, ideas are partly developed from discourses prevalent at the time, yet
authors also have their own intentions, and discourses are always also
detested and challenged.
Figuring out the intended significance of a text: finding the best
explanation of all the relevant evidence. Checking informed imagination
against plausible alternatives.
Always uncertainty.
17 >
Department of Media and Communication
Qualitative content analysis
4. The power of the reader
Do texts have meaning of their own? Rather, meaning as interpreted by
readers? Reception studies, powerful receivers
Derrida: meaning does not reside in a text but in the writing and reading
of it.
Barthes: there are as many texts as there are receivers.
Yet, we are expected to interpret texts according to intended meanings.
Meanings can usually be fixed?
18 >
Department of Media and Communication
Qualitative content analysis
5. Subjective summary interpretations?
Aiming to provide accurate, comprehensive and informative summaries
of texts.
Accurate: supported by the details of the texts, not inconsistent with any of them.
Comprehensive: relate to all major points in the text.
Informative: picks out the distinctive characteristics.
See McCullagh’s example of interpretations of the political writing of
John Locke (page 133).
19 >
Department of Media and Communication
Qualitative content analysis
6. Historians’ bias
Preconceptions and interersts influence the way we read texts.
Searching for evidence inconsistent with a preferred hypothesis.
Compared to the continuous attempt to falsify hypotheses, and the
impossibility of verifying claims completely (Karl Popper).
Researcher correcting each other.
20 >
Department of Media and Communication
Qualitative content analysis
7. Hermeneutic circles
McCullagh’a answers:
The meaning of the components of a texts are usually quite uanmbigious.
The meaning can be decided from the ”intra-textual” context.
Sometimes the ”extra-textual” context resolves ambiguity: the context in
which the text was produced.
Independent information about the author and the circumstances to
clarify the author’s intentions.
21 >
Department of Media and Communication
Quantitative content analysis
Quantitative content analysis
Robert Philip Weber (1990): Basic Content Analysis
Classifying textual material - more manageable bits of data
Several aims, such as:
To reflect cultural patterns of groups and socities
To reveal the focus of individual, group, institutional, or societal attention
Describe trends in communication content.
Counting, sorting and presenting content.
22 >
Department of Media and Communication
Quantitative content analysis
What is it?
A method for registering and analysing texts, with the aim of reaching a
systematic, objective and quantitative description of the content.
The many words of texts are classified into fewer content categories,
presumed to carry similar meanings.
Bernard Berelson (1952):
Merely analysing manifest content (possitivism)
Hence, very different from semiotical textual analyses
23 >
Department of Media and Communication
Quantitative content analysis
Quantitative content analysis
Berelson’s claim is rather disputed.
The requirements that the coding process must
be focused on the actual text remains.
Interpretations are often part of the actual
analytical process. To develop meaningful
conclusion. What does the text mean?
The methodological ideal is moreover to
combine different methodological approaches.
24 >
Department of Media and Communication
Quantitative content analysis
Requirement: Objective analysis
The research process must be performed according to explicitely
formulated procedures.
Coding must therefore be reliable, in the sense of being consistent.
Different people should code the same text in the same way.
25 >
Department of Media and Communication
Quantitative content analysis
Requirement: Systematic analysis
The empirical material - form, content, variables, categories - must
emerge from general principles.
How it should not be done (example from Ole Holsti):
Weyl and Possony’s study of racial intelligence (1963) - carefully selected texts that
could be used to support their hypothesis that Jews are smarter than other people.
26 >
Department of Media and Communication
Quantitative content analysis
How it’s done
Reducing texts (as texts or TV-shows or whatever) into smaller parts, the
recording units, and then to categorise these units into variables and
values/categories.
Units and variables must be clearly defined (to secure a reliable and valid
categorisation).
Sampling populations, means universe must first be identified (such as
all of Norway’s regional newspapers).
Simple random sampling
Stratified sampling (ensures each subpopulation is represented)
27 >
Department of Media and Communication
Quantitative content analysis
Recording units
Word
Sentence and paragraphs: If the researcher is interested in words and
phrases that appears together. Coders can be instructed to code whether
sentences convey a positive, neutral or negative reference to US foreign
policy.
The whole text: quite common in media research (Sigurd Allern:
Newsvalues in Norwegian newspapers). Usually requires a shorter text.
28 >
Department of Media and Communication
Quantitative content analysis
A hypothetical example
A historical study of how the 2003 war in Iraq was reported in three
Norwegian newspapers, Aftenposten, VG and Klassekampen.
UNIT: article
Variables: size, priority, news angle
Values: big/small, first-page/other page, positive/negative
29 >
Department of Media and Communication
Quantitative content analysis
A hypothetical example
Variable
Priority
1. First page story
2. Story referred to on the first page
3. Article starts in news-section in the newspaper
Values
4. Article starts in opinions-section in the newspaper
5. Article starts in cultural-section in the newspaper
6. Article starts in economical-section in the newspaper
7. Other
30 >
Department of Media and Communication
Quantitative content analysis
A hypothetical example
Size text
xxx cm
Size illustrations
xxx cm
Size main text
xxx cm
31 >
Department of Media and Communication
Quantitative content analysis
A hypothetical example
Concepts used on war
1. Libaration
2. Occupation
3. War
Source for article
32 >
1.
Newsagency
2.
Freelance
3.
The newspaper’s own correspondent/journalist
Department of Media and Communication
Quantitative content analysis
Coding
Values on variables must be mutually exclusive as well as cover all
possibilities.
Simple and one-dimensional variables.
Again: remember securing reliability by clearly defining units, variables
and values.
33 >
Department of Media and Communication
Quantitative content analysis
Analysis - words as recording unit
Looking for the highest frequency words
-> Ordered word-frequency list.
Example from Democratic and Republican Party Platforms:
Carter 1980:
Development, rights, health, women, work, education
Reagen 1980:
Soviet, military, tax, defense
34 >
Department of Media and Communication
Quantitative content analysis
Analysis: Swedish Radio on 9/11
Marina Ghersetti
”A Question of Partisanship: Swedish Radio on September 11”
The aim of the study: ”to map out the information given in the coverage
and to study whether or not there were stories or pieces that were either
inaccurate or biased” (204).
Empirical material: Swedish Radio’s programs on the attacts from 11 to
13th of September
a total of 35 hours and 48 minutes of programming
divided into 1143 units of individual news stories
35 >
Department of Media and Communication
Quantitative content analysis
Analysis: Swedish Radio on 9/11
Two main format groups: Reports on events and analysis/commentary
36 >
Total
Ekot
Other
Reporting on events
75
96
55
Analysis/commentary
24
4
44
Other reporting
1
-
1
Sum
100
100
100
N
1143
577
566
Department of Media and Communication
Quantitative content analysis
Analysis: Swedish Radio on 9/11
Position taking in reporting: figuring out whether claims of biased
reporting in Swedish radio could be justified.
Impartiality: balance and presentation
Balance: does the coverage provide room form all main actors and their arguments?
Measured: how many different actors appeared in each story?
Presentation: the extent to which coverage is netural - does not take a stand either for or
against one or the other side.
Measured: The presence of biased journalistic statements, positive or negative
statements.
37 >
Department of Media and Communication
Quantitative content analysis
Analysis: Swedish Radio on 9/11
Treatment of main actors
Value judgements about the US and US actors. Percentage
38 >
Total
Ekot
Other
Suoport, positive judgements
67
77
58
Criticism, negatie judgements
33
23
42
Sum
100
100
100
N
241
93
148
Department of Media and Communication
Quantitative content analysis
Analysis: Swedish Radio on 9/11
Treatment of main actors
Most common Judgement about the US, Percent total judgement:
The US has support and help from around the world
25
The US will find and punish the terrorists
18
The US is arrogant, gets to taste its own medicine
11
The US is united, supporting the President
8
US security forces have failed
8
The US has the right to defend itself and retaliate
7
39 >
Department of Media and Communication
Quantitative content analysis
Analysis: Swedish Radio on 9/11
Treatment of main actors
Value judgement about Terrorism/Terrorists, Number and Percent
40 >
Total
Ekot
Other
Suoport, positive judgements
5
0
9
Criticism, negatie judgements
95
100
91
Sum
100
100
100
N
183
75
108
Department of Media and Communication
Quantitative content analysis
Analysis: Swedish Radio on 9/11
Treatment of main actors
Most common judgments about Terrorism/Terrorists, Percent total
judgment
The Terrorists/terrorism is ruthless, evil, cowardly
31
Terrorism must be defeated
27
Terrorism is a threath to open, democratic society
19
The terrorists shall be punished
12
41 >
Department of Media and Communication
Quantitative content analysis
Analysis: Swedish Radio on 9/11
Treatment of main actors
Hence, the number of value judgments about bin Laden was significantly
lower than the number of judgments about US
A total of 95% of judgments about terrorism were negative, compared to
33% of judgments about US.
US was favoured.
The results do thus not support accusations of SR having been biased and
inaccurate in disadvantage to the US.
42 >
Department of Media and Communication
Quantitative content analysis
Asian values through content analysis
Massey, B. L. And Chang, L.A (2002) ’Locating Asian Values in Asian
Journalism: A Content Analysis of Web Newspapers’. Journal of
Communication, 5 (4): 987-1003
’This study tested arguments in the largely anecdotal debate over the existence of Asian
values in Asian journalism. News stories uploaded to 10 Asian online newspapers were
content analyzed for the prevalence of ”harmony” and ”supportiveness,” which the
literature suggests as key Asian values. The findings show that the journalistic emphasis
on Asian values is concentrated in the Southeast Asia subregion and tracks restrictions
on press freedom.’
43 >
Department of Media and Communication
Quantitative content analysis
Asian values through content analysis
Massey and Chang develop five research questions to guide their
analysis:
RQ1: To what extent do Asian journalists practice the Asian value of harmony by
keeping conflict out of their news reports?
RQ2: To what extent does their news reporting reflect the Asian value of
supportiveness?
RQ3: Can Asian journalism be distinguished from Western news work on the basis of
Asian values?
RQ4: Does the degree of harmony and supportiveness in Asian journalism vary
between the press of East, Southeast, and South Asia?
RQ5: Does the occurrence of Asian values in Asian reporting track the degree to which
a press system’s freedoms are restricted?
44 >
Department of Media and Communication
Quantitative content analysis
Asian values through content analysis
Sampling-frame narrowed to Web news sites published by general
circulation, English-language daily newspapers:
10 newspaper web sites, representing two East Asian countries, six Southeast Asian
countries, and two South Asian countries.
Unit of analysis: the current day’s ”hard” news story: articles about
international, regional, and domestic news events, and business and
sports news.
45 >
Department of Media and Communication
Quantitative content analysis
Asian values through content analysis
Each story coded for
The home-country of the newspaper in which it appeared
That country’s Freedom House press freedom rank
Whether it was supplied by an Asian or Western source
Operationalisation of ”harmony”: reliably inferred in news by an absence
of conflict as a central storytelling device
Operationalisation of ”supportiveness”: (of the state’s nation-building
efforts): story was judged to be supportive if it emphasised political,
economic, or social stability or strength, or social cohesion or cooperation.
46 >
Department of Media and Communication
Quantitative content analysis
Asian values through content analysis
1845 news stories harvested from 10 online newspapers.
Asian values of ”harmony” and ”supportiveness”
Harmony, as absence of conflict
Supportiveness
No conflict
Conflict
n
Supportive
Critical
n
Home news at home
59,6%
40,4%
778
49,3 %
50,7 %
505
Home news abroad
64,4%
35,6%
45
56 %
44 %
25
Foreign news at home
67, 3%
32,7%
49
43,8 %
56,3 %
32
Foreign news abroad
46,7%
53,3%
137
26,1 %
73,9 %
92
Ethnocentric orientation
47 >
Department of Media and Communication
Quantitative content analysis
Asian values through content analysis
1845 news stories harvested from 10 online newspapers.
Asian values by subregion
Harmony, as absence of conflict
Supportiveness
No conflict
Conflict
n
Supportive
Critical
n
East Asia
45,5%
54,6%
152
22,5 %
77,5 %
102
Southeast Asia
66,4%
33,6%
654
56,8 %
43,2%
375
South Asia
43,1%
56,9%
204
36,5 %
63,5 %
178
Geographic division
48 >
Department of Media and Communication