In a word, Discourse
Analysis means Doing Analysis. (Antaki et al., 2003)
I think that the overall arching connection between these
readings was that qualitative research is equivalent in its own right to
quantitative research, but that a problem that comes up in this line of work is
that the researcher doesn’t pay it the respect that it deserves. The research isn’t always done accurately,
the analysis is sloppy, or time hasn’t been devoted to developing the craft as
they should have. Researchers need to
spend time learning about what an interview is and isn’t. They need to recognize how to really do the
analysis, so that it is conducted and verified by the data itself, such as
being followed up in the next turn. This
isn’t easy work and if done correctly, provides everything that quantitative
researchers are looking for, i.e., validity, without all of those messy
numbers.
In terms of the Golato article, I found the concept of DCTs
interesting, but I was left with that feeling of ‘Duh!’ Humans are infallible. They are different. No interaction can ever truly be replicated,
because nothing can ever been the same in a subsequent interaction. For example, I am sitting on my couch typing
this paper and my husband has been researching vacation destination for us for
winter break. He was excited about taking
me on a beach vacation, because he knows how much I love the beach. When he got my deflated reaction about the
possible cold weather in late December, he went back to looking. From this point forward, every vacation
option he encounters that involves a North American beach will be viewed
differently than before. Our interaction
changed his perception. If he or someone
else tried to replicate our interaction, they would approach it with a
different stance.
Golato talked about the function of a compliment in a
classroom and non-institutional setting and it made me think about a southern
term that we have blogged about before, ‘Bless your heart.’ Anywhere in America that saying would appear
to be a sympathetic statement, but in the south, it probably means the speaker
thinks that the intended is an idiot.
Context matters. Background
knowledge matters. This also made me think
of Peter Johnston’s Choice Words or Opening Minds. The simplest of phrases can be interpreted in
a number of ways based on the context in which they are said or based on the
tone of voice or the mood that the participants are in.
Potter, J. &
Hepburn, A. (2011). Eight challenges for interview
researchers. For J.F. Gubrium and J.A.
Holstein (Eds) (forthcoming). Handbook of Interview Research (2nd
Ed.). London: Sage.
The irony is that qualitative interviews are massively overused, but
their potential has been massively restricted.
Our aim in this chapter is to make the case that
interviewing has been too easy, too obvious, too little studied and too open to
providing a convenient launch pad for poor research…faces up to a series of 8
challenges.
Set 1 (reporting to the interview study):
1)
improving the transparency of the interview
set-up (how participants are recruited);
2)
more fully displaying the active role of the
interviewer (interview is interactional, but not followed through in research
practice);
3)
using representational forms that show the
interactional production of interviews;
4)
tying analytic observations to specific
interview elements.
Set 2 (analysis of the interview):
5)
how interviews are flooded with social science
categories, assumptions and research agenda;
6)
the varying footing of interviewer and
interviewee;
7)
the orientations to stake and interest on the
part of the interviewer and interviewee;
8)
the way cognitive, individualist assumptions
about human actors are presupposed.
Golato, A.
(2003). Studying compliment responses: A
comparison of DCTs and recordings of naturally occurring talk. Applied Linguistics, 24(1), 90-121.
Golato argues that ‘recording naturally occurring
talk-in-interaction enables the researcher to study how language is organized
and realized in natural settings (pg. 90.’
Comparative studies (e.g. Yuan 2001) of role plays and
naturally occurring conversations have indeed shown that what is said and, more
importantly, how it is said differ drastically in role plays and in actual
conversations (pg. 94).
Study of classroom compliments versus non-classroom
compliments…’One cannot assume that compliment serve identical functions and
have the same design in ordinary and institutional talk (pg. 96).’
Goodman, S.
(2008). The generalizability of
discursive research. Qualitative research
in psychology, 5, 265-275.
‘I show how such findings (discourse analytic findings) can
be considered generalizable to the extent that they can show how a particular
discursive strategy will often bring about the same interactional results (pg.
265).’
-quantitative vs. qualitative comparison and definition of
generalizability: ‘The process of making statements about the general
population on the basis of relevant research (e.g., experiments or surveys).’
(pg. 265)
Qualitative researchers tend to accept that their findings
cannot be generalized in this way: instead generalizability is sacrificed in
favor of a more detailed understanding of the issue being researched (pg. 266).
Guba (1981):
qualitative researchers could replace this with what he called
transferability where findings within one context can be applied to another if
there is sufficient knowledge of the contexts in question (pg. 266).
To Edwards and Potter, social action refers to the
interactional accomplishment that a piece of discourse brings about. This may be, for example, the action of
remembering, blaming, performing prejudice, or any other outcome of social
interaction (pg. 267).
To conclude, where it has been widely accepted that
discursive psychological findings are not generalizable, I have shown that a
discursive strategy can be generalizable to the extent that the ‘action’ that
it accomplishes can be generalized across contexts (pg. 273).
Antaki, C., Billig,
M., Edwards, D. & Potter, J. (2003).
“Discourse analysis means doing analysis: A critique of six analytic
shortcomings.” Discourse Analysis Online,
1. Available from:
<http://www.shu.ac.uk/daol/articles/v1/n1/a1/antaki2002002-paper.html>.
‘Discourse analysis still can be misunderstood by those who
have been schooled in quantitative analysis.
It might appear to quantitative researchers that ‘anything goes’ in
qualitative work in general, and discourse analysis in particular.’
-Problems: researchers self-education, work produced
embodies basic problems
“Writers are not doing analysis if they summarize, it they
take sides, it they parade quotes, or if they simply spot in their data
features of talk or text that are already well-known. Nor are they doing analysis if their
discovery of discourses, or mental constructs, is circular, or if they
unconsciously treat their findings as surveys.”
‘It is safe to say that analysis means a close engagement
with one’s text or transcripts, and the illumination of their meaning and
significance through insightful and technically sophisticated work. In a word, Discourse Analysis means Doing
Analysis.
Golato Notes…
-purport actual language use: the forms and formats of a
compliment response, the comparison of compliment responses in different
languages (pg. 91)
-‘My results suggest caution in using DCTs if one’s goal is
to describe actual language use (pg. 91).
-administrative advantages of DCTs…researchers can better
control variables, quickly gather large amounts of data without transcription,
better ability to compare native and non-native speakers (pg. 92)
Goodman Notes…
Definitions:
-Validity: showing what it is claiming to show
-Construct Validity: show that the effect demonstrated can
be generalized from the measures used in the study to the fuller construct
-External Validity: being able to generalize the research
findings to the population in general
-Ecological Validity: the extent to which the research
findings can be generalized to other settings
-Population Validity: the extent to which the research
findings from the sample studied to the wider population
-Reliability: the extent to which a given finding will be
consistently reproduced where it is deemed that similar results will be
consistently found from the same research study (pg. 265-266)
-best known example of a generalizable conversational
strategy is that of the three-part list, where lists are consistently seen to
include three items to show that the list is complete, and that what is been
described is normative (pg. 268)
The analyst should be able to state that the following is
true in order to make a claim of generalizability:
1)
A discursive strategy can be shown to achieve a
certain rhetorical accomplishment.
2)
This strategy can be identified as being used in
a range of conversational settings in an attempt to bring about this rhetorical
accomplishment.
3)
If this strategy often brings about the same accomplishment
this strategy can be described as a successful strategy.
4)
It can be shown that successful strategies will
be used by a range of speakers in a range of contexts to bring about the same
rhetorical end. To this extent, it is a
generalizable example of an action performed by a rhetorical strategy.
5)
We may eventually begin to see opposition to
successful and generally used strategies (pg. 272).
Antaki et al…
Qualitative/Quantitative: Both want to do something with the
data. Neither is content merely to lay
the data out flat.
Transcription prepares the data for analysis. However, it is not the analysis itself.
Hutchby & Wooffit Revisited…
The most central of these assumptions is that ordinary talk
is a highly organized, socially ordered phenomenon (pg. 11).
At the most basic level, conversation analysis is the study
of talk. It is the systematic analysis
of the talk produced in everyday situations of human interaction,
talk-in-interaction (pg. 11).
One aim of CA therefore is to reveal this sequential
order...describable ways in which turns are linked together into definite
sequences (pg.42).
The next-turn is the place where speakers display their
understanding of the prior turn’s possible completion. The next speaker has
performed on the typed of utterance the prior speaker has produced (pg.42).
‘Paired Action Sequences that conventionally come in
pairs: questions-answers,
greetings-return greetings, invitations-acceptance/declinations (pg.42)
The format for agreements is labeled the ‘preferred’ action
turn shape and the disagreement format is called the ‘dispreferred’ action turn
shape (pg. 46).
‘dispreference markers’: turns that is some way depart from
what seems to be expected incorporate a variety of these, ‘well’ or ‘um’ (pg.
47)
Preferred actions are characteristically performed
straightforward and without delay, while dispreferred actions are delayed,
qualified and accounted for. The concept
refers to these structural features of turn-design and not to individual
motivations or psychological dispositions (pg. 47).
Organization of Turn-Taking:
1)
turn-taking has occurs
2)
one speaker tends to talk at a time
3)
turns are taken with as little gap or overlap
between them as possible (pg. 49.
Overlapping talk may be considered evidence of an incoming
speaker’s failure to take notice of whether the current speaker is or is not
finished (pg. 54).
Four varieties of repair sequences:
1)
Self-initiated self-repair: repair is both
initiated and carried out by the speaker of the trouble source.
2)
Other-initiated self-repair: Repair is carried
out by speaker of the trouble source but initiated by the recipient.
3)
Self-initiated other-repair: The speaker of a
trouble source may try and get the recipient to repair the trouble—for instance
if a name is proving troublesome to remember.
4)
Other-initiated other-repair: The recipient of a
trouble-source turn both initiates and carries out the repair. This is closest to what is conventionally
understood as ‘correction’ (pg. 60).
The first two places in which repair can occur are within,
or immediately after, the turn construction unit containing the trouble source
(pg. 62).
The second place in which repair can be done occurs
immediately at the next transition relevance place after the trouble source
(pg. 63).
CA also focuses on:
Gaps and pauses
Breathiness
Talk in Institutional Settings
Turn-type pre-allocation means that participants are
normatively constrained in the types of turns they may make according to their
particular institutional roles. Typically, the format involves chains of
question-answer sequences, in which the institutional figures ask the questions
and the witness, pupil, or interviewee is expected to provide the answers. This format is oriented to by participants,
but at the same time normative rules operate which mean that participants can
be sanctioned if they refuse to stay within the boundaries of the
question-answer framework (pg. 141).
A three-part list is simply a way of packaging a point or
position in an argument using a list of three separate items (pg. 183).
A contrastive device is means of packaging a point, where
one argument or approach is contrasted with another in such a way that the
speaker’s favoured position is seen to be superior (pg. 183).
While CA aims to describe the ways that participants display
that they are aware of specific contextual factors (by observably modifying the
ways that they talk, for instance), CDA maintains that there are other factors,
external to the situation the speakers are in, and of which the speakers may
not be aware, that impact on the production of their talk (pg. 210).
