Sociological Imagination and Investigation

1
Sociological Imagination and Investigation

• Mixing methods Sociological imagination or
  epistemological heresy?

2
Mixing our methods An epistemological faux pas?

• Epistemological and methodological debates are
  routinely set up around the discussion of
  supposedly mutually exclusive (quantitative and
  qualitative) approaches.
• Quantitative methods are generally seen as
  ideally suited to
• Questions to which the answers are quantifiable
  e.g. How often do you make use of service x
  every day, once a month, once a year, never ?
• Questions where only absolute guarantees of
  anonymity would persuade respondents to answer
  honestly e.g. questions about income,
  illegal/deviant activities, voting etc .
• When you want to be able to say something about
  the whole population.
• Qualitative methods are seen as preferable when
• You are interpreting views, opinions, ideas
• Context is important
• Research is with vulnerable or hard to reach
  groups
• Research is into complex or dynamic social
  processes
• But, different methods can be used either in
  parallel or sequentially and either discretely or
  in an integrated fashion to important effect.

3
Combining methods

• There are many ways to classify the ways in which
  how different methods can be combined (see Punch
  2005).
• The three broad approaches of combining methods
  that I want to draw your attention to are
• Triangulation
• Parallel use of methods
• Sequential use of methods

4
Triangulation of data

• Triangulation is a process of checking the
  inferences drawn from one set of data sources by
  collecting data from others.
• Within an ethnographic approach you might want
  to check your own observations of a
  situation/event against how it is described to
  you in interview.
• Alternatively you might use a diary entry or a
  set of family photos to confirm or contest
  something said in conversation.
• This process of triangulation in ethnographic
  work is a way of improving validity countering
  the view that the researcher gets a very
  partial understanding of the social world. But
  this comes about because it allows you to first
  capture differences and contradictions and work
  through them to try and understand the whole.

5
Parallel use of methods

• Triangulation is an example of the parallel use
  of methods as a check on inference from that
  data.
• However, methods can be used discretely in
  parallel in order to provide a different kind of
  data, answering a different question, but helpful
  to the overall research
• It is common for qualitative researchers to
  employ an element of small-scale survey for the
  purposes of generating descriptive but not
  inferential statistics, for example.
• The dangers in this kind of parallel work is that
  the data from different sources may simply
  present different types of information that
  cannot be easily integrated and thus produce no
  valuable analytic yield.

6
Mixing methods sequential approaches

• An alternative approach is to use different
  methods sequentially.
• The most common example of this is the
  recognition by survey researchers that the
  formulation of questions for questionnaires is
  benefited by using qualitative research methods
  either interview or focus groups to understand
  how respondents are likely to interpret
  questions. The aim of such mixing of methods is
  to reduce misreporting and cross-cultural
  confusion and thus the qualitative element is
  designed to improve the main technique which is
  quantitative.
• Although less common, qualitative specialists may
  also use quantitative work to identify
  interesting cases for study, or else to
  statistically test hypotheses thrown up by
  ethnographic work.

7
Drug use An empirical example

• 1. Choosing drugs
• Which young people begin experimenting with
  drugs and in what contexts? Who chooses not to
  use drugs? Are gender, class, ethnicity or age
  significant factors?
• 2. The role and function of drugs in youth
  groupings
• To what extent do drugs talk, symbols, style
  and image shape everyday communication among
  youth? How does this vary for young people who
  use drugs regularly, who have experimented with
  drugs and who have never used?
• 3. Drugs and youth cultural formations
• What is the relationship between youth cultural
  style and drug use? Are certain youth style
  groups associated with certain drug preferences
  (in terms of use/non-use and particular drug
  empathies)?
• 4. Narratives of drug use
• What, in young peoples own narratives,
  constitutes experimentation with drugs,
  habitual, regular or dependent drug use?
  How important are peer influence, drug
  dealers and personal choice in young peoples
  own narratives of drug decisions?
• 5. The normalization of drugs
• Does the accommodation of drug use in the peer
  group, personal drug experimentation and/or
  regular drug use indicate an acceptance of drug
  use as normal?
• 6. Responses and inputs to drug prevention
  initiatives
• To what extent do those responsible for leading
  drugs prevention work in Russian cities listen to
  young people's experience? To whom do young
  people talk, and listen, about drug use?

8
Mixing methods Logic and practice

• We took three regions as case studies and chose
  three different towns/cities in each region as
  fieldwork sites. We employed the following
  methods
• a representative survey based on self-complete
  questionnaire (n2914) of 14-19 year olds in all
  9 locations
• semi-structured interviews with a sub-sample of
  survey respondents (n95) in all 9 locations
• ethnographic studies in 3 of the 9 field sites
  with a total of 19 groups of young people
• expert interviews with key personnel engaged at
  the regional and city/town level with devising
  and implementing health education responses to
  drug use by young people in 3 of the 9 field
  sites.

9
The quantitative element

• Method
• First check no such data exist already e.g. in
  official statistics.
• Be careful not to confuse incidence with
  prevalence data.
• Check for already existing survey data (e.g.
  household survey data).
• Choose a sampling method
• A mixed method approach may require compromises!

• Question
• How many young people encounter drug offers?
• How many young people experiment with/use drugs?
• Has drug use been desubculturalised?
• These are questions about a population in
  general. We therefore need to ask every single
  member of the population about their
  behaviour/attitudes or survey a representative
  sample.

10
The interview element

• Semi-structured interview (loose interview guide
  with prompts).
• Importance of teasing out what respondents mean
  by normal.
• Allows space for respondents to talk about drug
  use experiences in their own words.
• Allows unexpected outcomes e.g. dont recognise
  beer as alcohol.

• Students at the mining college in Vorkuta after
  finishing the questionnaire.
• Although 3 respondents went on to give
  interviews, there was no cross referencing
  between questionnaire and interview responses for
  ethical reasons.

11
The ethnographic element

• Ethnographic research helped us
• understand the context (group setting) of drugs
  decisions.
• check narratives of drug use against behaviour.
• see drug use trajectories (careers).

This is a photograph of two key informants
seeing the ethnographer off at the railway
station when she left Vorkuta.
12
Expert interview element

• Expert or elite interviews are used where the
  target respondents are people with specific
  insider, or specialist knowledge.
• In our case these were individuals from NGOs,
  drugs clinics, local government and education
  departments involved in drugs education.
• Access to these respondents is obtained
  differently and agreements about confidentiality
  of information provided are often more formal.

13
Conclusion

• Epistemological divides in the social sciences
  are weakening, but there remains some
  considerable scepticism about mixing methods.
• We have suggested that there are no absolute
  divisions between deductive and inductive
  approaches or between statistical/survey
  quantitative and interview/observation based
  qualitative methodological techniques.
• These approaches and methods should be viewed as
  complementary rather than mutually exclusive and
  we should focus on selecting the most appropriate
  methods for a particular question and considering
  how different methods might be integrated (or
  triangulated) in the course of investigation.
• We should also pay attention to the different
  ways in which methods can be use discretely or
  in an integrated fashion in order to maximise the
  benefits of a mixed method approach.