Week 6 content analysis
Content analysis aims to quantify content in terms of predetermined categories and in a
systematic and replicable manner
What can you study: any question related to communication content
The nature of communication (descriptive)
How communication is related to other factors (Explanatory)
Ok: how are the presidential candidates evaluated in the us mainstream press?
Not ok: how do readers evaluate the coverage of the US elections in the US mainstream
press?
The art of measuring clear operation definition of what you want to measure with very
detailed instructions and training.
Variables must be
Clear
unambiguous
Answer categories must be
Mutually exclusive not overlap
Exhaustive
Manifest versus latent
Registration unit the specific level at which you actually measure (sentence or scene)
Reliability:
Aim to develop a code book that leads to consistent coding
One coder: intra coder reliability (within)
Multiple coders: inter coder reliability (between) faster, forces very clear
instructions leading to higher reliability and less subjective
Formal tests
Reliability vs. validity
Simple codes are often more reliable chances that coders will assign the same code
is higher for simple code consistent coding higher reliability
BUT validity may suffer from simple codes
Stratified sampling: random but you want all the layers to be present in your sample
Convenience sample: not random
ACA: automated content analysis saves time, money, no restrictions on sample size,
reliable, sees data patterns BUT complex tasks are challenging (meaning of text), data quality
issues can be hard to trace, generating the algorithm is a subjective human process, errors can
have great consequences and algorithms sometimes secret so no transparency or replicability
, Week 7 surveys
Survey: a series of formatted questions delivered to a defined sample of people with the
expectation of responses immediately or within a few days.
Capture public opinion at a point in time
With appropriate sampling, we can generalize with a known statistical confidence from a
sample to a wider population
Advantages Disadvantages
Relatively low cost Generally limited to scaled or check-list
questions
Relatively fast delivery No control over response rate
Rapid data processing Provides information more than
understanding
Can research large populations Increasing public resistance
Multiple methods: phone email, internet, f2f Difficult to explore issues in depth
Types of surveys:
cross-sectional: asks 1 group of people the same questions at 1 point in time
longitudinal: observations are taken more than once, you examine
differences/relationships over time, not able to detect causal relationships, takes into
account influence of time
o Trend: a longitudinal study that examines changes in a population across time
(not clear why) different people over time
o Cohort: a longitudinal study that examines changes in a cohort across time
(provides more detail than trend study, data from different individuals within
the same cohort, no cause and effect study) same type of people over time
o Panel studies: a longitudinal study that examines changes in individual across
time (same individuals, opportunity for more in-depth analysis, more detailed
than cohort study, but still no cause and effect) same individuals over time
o Cross-lagged panel surveys: best form of survey research design for
determining cause and effect. Measures independent and dependent variable
at several time points, you need panel data
To determine causality, you need a true experiment. However, panel studies with cross-
lagged surveys may provide you with some indication of causality.
Research validity:
- Internal validity is LOW
- External validity is HIGH large and random sample
-
Survey mode:
With Response Speed Costs Complexity
interviewer questionnaire
- face to face Good Slow (very) high (very) high
- phone Reasonable Fast High (very) high
Without
interviewer
- paper and Good Slow Low/moderate Low
pencil
- email/web Reasonable/ Very fast Very low Very high
Content analysis aims to quantify content in terms of predetermined categories and in a
systematic and replicable manner
What can you study: any question related to communication content
The nature of communication (descriptive)
How communication is related to other factors (Explanatory)
Ok: how are the presidential candidates evaluated in the us mainstream press?
Not ok: how do readers evaluate the coverage of the US elections in the US mainstream
press?
The art of measuring clear operation definition of what you want to measure with very
detailed instructions and training.
Variables must be
Clear
unambiguous
Answer categories must be
Mutually exclusive not overlap
Exhaustive
Manifest versus latent
Registration unit the specific level at which you actually measure (sentence or scene)
Reliability:
Aim to develop a code book that leads to consistent coding
One coder: intra coder reliability (within)
Multiple coders: inter coder reliability (between) faster, forces very clear
instructions leading to higher reliability and less subjective
Formal tests
Reliability vs. validity
Simple codes are often more reliable chances that coders will assign the same code
is higher for simple code consistent coding higher reliability
BUT validity may suffer from simple codes
Stratified sampling: random but you want all the layers to be present in your sample
Convenience sample: not random
ACA: automated content analysis saves time, money, no restrictions on sample size,
reliable, sees data patterns BUT complex tasks are challenging (meaning of text), data quality
issues can be hard to trace, generating the algorithm is a subjective human process, errors can
have great consequences and algorithms sometimes secret so no transparency or replicability
, Week 7 surveys
Survey: a series of formatted questions delivered to a defined sample of people with the
expectation of responses immediately or within a few days.
Capture public opinion at a point in time
With appropriate sampling, we can generalize with a known statistical confidence from a
sample to a wider population
Advantages Disadvantages
Relatively low cost Generally limited to scaled or check-list
questions
Relatively fast delivery No control over response rate
Rapid data processing Provides information more than
understanding
Can research large populations Increasing public resistance
Multiple methods: phone email, internet, f2f Difficult to explore issues in depth
Types of surveys:
cross-sectional: asks 1 group of people the same questions at 1 point in time
longitudinal: observations are taken more than once, you examine
differences/relationships over time, not able to detect causal relationships, takes into
account influence of time
o Trend: a longitudinal study that examines changes in a population across time
(not clear why) different people over time
o Cohort: a longitudinal study that examines changes in a cohort across time
(provides more detail than trend study, data from different individuals within
the same cohort, no cause and effect study) same type of people over time
o Panel studies: a longitudinal study that examines changes in individual across
time (same individuals, opportunity for more in-depth analysis, more detailed
than cohort study, but still no cause and effect) same individuals over time
o Cross-lagged panel surveys: best form of survey research design for
determining cause and effect. Measures independent and dependent variable
at several time points, you need panel data
To determine causality, you need a true experiment. However, panel studies with cross-
lagged surveys may provide you with some indication of causality.
Research validity:
- Internal validity is LOW
- External validity is HIGH large and random sample
-
Survey mode:
With Response Speed Costs Complexity
interviewer questionnaire
- face to face Good Slow (very) high (very) high
- phone Reasonable Fast High (very) high
Without
interviewer
- paper and Good Slow Low/moderate Low
pencil
- email/web Reasonable/ Very fast Very low Very high
