Summary for Research Workshop: Content Analysis

CONTENT ANALYSIS NOTES

LECTURE 1 - SLIDES
-> Content analysis: quantitative analysis of characteristics of media content
(media messages)- analyzes in a systematic, objective, replicable manner

FOUNDING FATHERS=
● Bernard Berelson(1912-1979) -> in communication research
● Ole Holsti(1933-) -> for the social sciences and humanities
● Klaus Krippendorff(1932-) -> into to methodology

CONTENT ANALYSIS
● Cause of media effects
● Outcome of media production and environment factors

- ADVANTAGES
● Accessible
● Nonobstrusive
● Non Reactive (-> control)
● Longitudinal in no time
● Endless relevant possibilities

-> more descriptive and explanatory research questions
- still dominant in scientific content analyses
- dominant in practice-oriented content analyses

DESCRIPTIVE vs EXPLANATORY
D-> less theoretical, less knowledge gain, less scientific
E-> more theoretical, more knowledge gain, more scientific

SAMPLING
- all units in the population are explained -> we call this a CENSUS

Non-Probability -> Consecutive Period
all editions of periodically appearing publications within chosen period are
analyzed without interruption

,-> SUITABLE for analyses of continuing news story in a specific period (e.g.
election campaign)
-> UN-SUITABLE for longer period -> you ignore potentially large variation in
content over time as a result of current events

Probability Sample -> often chosen if publications appear daily, such as news
media
- Simple Random Sample = not always effective or efficient
- Systematic Sample = suitable for smaller groups and longer periods -
control over dispersion of selected units across sampling frame
- Stratified Sample = suitable for diverse population
- Constructed period = suitable for daily analyses over a longer time period
= effective and sufficient
● accounts for selection bias due to cyclical variation in media content
- cyclical variation = systematic variation due to rhythm and logic of
medium or source
- Multistage sample = suitable when sample should reflect all dimensions
of media landscape and content


Content analysis is central in communication
-> content is the mediator between production and effects
-> they are connected with experiments but have low ecological validity


LECTURE SLIDES WEEK 2

MANIFEST CONTENT
- directly observable (concrete)
- clear meaning
- unidimensional
- single-term measure

LATENT CONTENT
- not directly observable (anstract)
- ambiguous meaning
- multidimensional

, - measured through multiple items

The MORE LATENT
● THE MORE its meaning is judged (constructed) instead of observed by
the coder
● THE MORE its meaning becomes dependent on pre-existing mental
schemes (frames or reference) of the coder
● THE MORE the meaning making shifts from the content to coder
● THE MORE subjective the substantive interpretation

VALIDITY OF MEASUREMENT
● a valid measure of a characteristic is a measure that covers in a correct
and compete way the content and meaning of that concept

- Operationalizing latent media content (constructs):
Translation into a set of manifest characteristics of content (‘indicators’)

- Always start from literature (theoretical insights,, empirical evidence)

WAYS OF EVALUATING VALIDITY OF MEASUREMENT
● face validity: most basic form - seems common sense
● construct validity: relation between concept and observable measures
● association between measurement instrument and :
- other version of same type of measurement instrument (concurrent
validity)
- plausible outcome variable (predictive validity)
● comparison outcomes with those of different type of instrument
(convergent validity)
● cross check among producers or users of content

SOCIAL VALIDITY
- the social significance and meaning of the analyzed content beyond the
academic community (e.g. operationalizing violence)