Transcript introduction - UC Berkeley School of Information

INFO 272. Qualitative Research Methods
‘Corpus Construction’
‘Corpus Construction’

Defining the sites and subjects of in situ
work
 Making decisions about your field site(s) –
how a social phenomenon of interest is
mapped out onto spatial terrain
 Selecting people to follow, observe and/or
interview
 Selecting media / artifacts from the setting
for further analysis
Competence and Innovation

Competence (Bauer and Gaskell)
 Systematic
 Issues of public accountability

Innovation (Becker)
 Challenge conventional thinking
Doing Innovative Research

Starting Where You Are (Lofland
and Lofland)
 Commitment and Curiosity
 Access and ‘getting in’

Willingness to go where
others won’t
 The inconvenient and
uncomfortable
 The illegitimate
Approaches
Total enumeration (i.e. census)
 Statistical random sample
 Snowball sample (iteration again)
 Convenience sample (bad)

Random vs. Systematic

Random Statistical
Sampling
 Distribution of already
known attributes
 Sample has a
distribution of criterion
= population as a
whole
 Popular
misconception – the
greater the # in the
sample, the more
accurate

‘Corpus
Construction’
Typifies unknown
attributes
 Systematic selection
to some alternative
rationale (not a
convenience sample)

Unknowable Populations
Many populations of ‘individuals’ are
knowable, however…
What about ‘actions?’
 What about ‘situations?’
 Open systems (i.e. language) = infinite
populations

Mapping the Unknowable
Social strata, functions and categories (known)
Representations
(unknown)
Varieties of:
Belief
Attitudes
Opinions
Stereotypes
Ideologies
Worldviews
Habits
Practices
[Bauer and Gaskell]
Mapping the Unknowable
Iteration until Saturation
 Don’t collect too much data [logistical
limits]

Problems of Social Strata in CrossCultural Research
Demographic Form
Extending Selection Strategies: Sampling
for ‘Innovation’
Identify the case that is likely to upset your
thinking and look for it – (the counterexample) e.g. morphine, opium, heroin
addicts
 If someone says it has already been studied,
its probably time to study it again.
 Studying the non-serious and the ‘boring’

Loose Ends: Selecting Field Sites
Some work is clearly ‘sited’
 Some is not (amorphous social settings) –
and therefore locating such work will be
more involved
 Sites may be ‘open’ or ‘closed’

Loose Ends: Collecting text,
images, data
Text produced in the process of research
vs. texts produced for other purposes
 Bauer and Gaskell’s simplified treatment of
newspapers, etc. – newspapers as…
 vs. Becker’s concern with the ‘sociology of
record keeping’
 in media studies, the ‘active audience’

In Conclusion - Representativeness?
The problem of unknowable populations
 Rather than ‘representativeness’ we are
seeking ‘range’ and variation in the social
phenomenon under study
 To what effect? Challenging notions of
what is ‘natural’ or ‘universal’ about a
phenomenon

To Review
Population and the problem of unknowable
populations
 Selection for range/diversity of the social
phenomenon rather than
representativeness
 Selection for innovation
 Stopping criterion

For Thursday
Read Lofland and Lofland section on
logging data
 Read UC guidelines for protection of
human subjects
