slide1 n.
Skip this Video
Download Presentation
My background …

Loading in 2 Seconds...

play fullscreen
1 / 24

My background … - PowerPoint PPT Presentation

  • Uploaded on

DISMANTLING THE QUANTITATIVE – QUALITATIVE DIVIDE Comments On Hypothesis Testing, Induction, Statistics, Fiction And Epistemological Anarchy. Presentation at the 3rd International Conference on Interdisciplinary Social Sciences, Prato, Italy, July 2008

I am the owner, or an agent authorized to act on behalf of the owner, of the copyrighted work described.
Download Presentation

PowerPoint Slideshow about 'My background …' - leigh

Download Now An Image/Link below is provided (as is) to download presentation

Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author.While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server.

- - - - - - - - - - - - - - - - - - - - - - - - - - E N D - - - - - - - - - - - - - - - - - - - - - - - - - -
Presentation Transcript

DISMANTLING THE QUANTITATIVE – QUALITATIVE DIVIDEComments On Hypothesis Testing, Induction, Statistics, Fiction And Epistemological Anarchy

Presentation at the 3rd International Conference on Interdisciplinary Social Sciences, Prato, Italy, July 2008

Michael Wood ( and Christine Welch

Portsmouth University Business School, UK

There is a revised draft paper at

and this presentation is at


We suspect this presentation may be a bit of a mess because it’s largely about things which don’t make too much sense. So we’ll try and impose a clear framework …

what conceptual framework is best for understanding research
What conceptual framework is best for understanding research?
  • Perspective 1


  • Perspective 2


  • Perspective 3


  • Perspective 4

Which do you think will be the winner?

perspective 1 two kinds of research

Generalisation through statistics

Research progresses through hypotheses and deductions

Observer independent

Large random samples

Etc, etc

Social Constructionism

Generalisation through theoretical abstraction

Research progresses through rich data and induction

Observer part of study

Small purposive samples

Etc, etc

Perspective 1: Two kinds of research

Based loosely on Easterby-Smith et al (2002: 30). They list 8 dimensions. Other authors may use different labels for the two types: commonest probably Quantitative vs Qualitative.

the two kinds of research
The two kinds of research?
  • Quantitative / positivist / deductive
    • Hard and spiky
  • Qualitative / phenomenological / social constructivist / interpretivist / etc
    • Soft and cuddly

Often suggested that researchers need to choose one or the other.

which side are you on
Which side are you on?
  • To hard and spiky people, soft and cuddly research is lacking in rigour
  • To soft and cuddly people, hard and spiky research is superficial and lacking in richness and relevance

… but is this a genuine dichotomy?

perspective 2 more than two types of research e g
Perspective 2: More than two types of research: e.g. ….

Easterby-Smith et al (2002) list 8 bipolar dimensions: this leads to the possibility of 28 or 64 types of research.

what does qualitative mean
What does qualitative mean?
  • General term from right hand side of Perspective 1 ………?
  • Opposite of statistical ……… ?
  • Focussed on qualities not quantities … ?
  • Detailed information / in depth analysis (eg not superficial questionnaire) …… OK
But …
  • Why not use different approaches in one project?
    • E.g. Glebbeek and Bax (2004) found statistical evidence for the hypothesised inverted U shaped relationship between staff turnover and firm performance, but why not back this up with case studies to look at possible reasons for this effect?
    • Britten et al (2000) identified categories of misunderstanding between doctors and patients, but why not do a statistical survey to see how common each category is?
  • So …
perspective 3 multimethodology or mixed methods
Perspective 3: Multimethodology or Mixed methods

Use different approaches in one project …

But …
  • There are problems with many of the bipolar dimensions used to pigeonhole research
    • Concepts used may be vague, ambiguous or prone to misinterpretation
    • May not be bipolar: reasonable approaches may be omitted
two bipolar dimensions
Two bipolar dimensions?





“ …” indicates vague terms we don’t like

or do these fit on the same dimension
Or … do these fit on the same dimension?

Deterministic laws

Empirical possibilities






Fictional possibilities

Is this one dimension?

extended statistical non statistical dimension
Extended statistical – non-statistical dimension
  • Deterministic laws – what always happens
  • Statistics – what sometimes happens
  • Illustrative inference – what has happened at least once (some “qualitative research”)
  • Fiction – what is possible / imaginable
and what about this
And what about this … ?

Deductive (e.g. applying a model)




Using a framework or paradigm

to define questions

Much research is neither hypothetico-deductive nor inductive. There is no obvious linear dimension here, which is why we’ve made the layout of this slide a bit of a mess.

So …

fourth perspective
Fourth perspective
  • Avoid unhelpful concepts. If in doubt, shut up, or use more straightforward language!
    • Induction?
    • Qualitative?
    • etc
  • No useful general categorisation schemes for research (=only useful grand narrative)
    • If you stick to such a scheme you risk ignoring useful possibilities
    • Epistemological anarchy. (= Postmodernism?)
      • Feyerabend: “anything goes”
  • But some concepts are worth careful thought …
statistical methods
Statistical methods
  • Important approach
    • Over-estimated by proponents, underestimated by opponents
  • Focus on null hypothesis tests is usually stupid
    • Better to measure size of effect
    • E.g. Glebbeek and Bax (2002)
    • Statistics = formal methods for doing induction
  • Sampling / context needs care
    • E.g. Glebbeek and Bax (2002)
  • “Qualitative” data often analysed statistically
    • Should be done properly
following a framework or paradigm
Following a framework or paradigm
  • Neither hypothesis testing nor (pure) induction
  • Kuhn’s normal science
  • Obviously a good idea but not in the standard menu of approaches
  • More helpful concept than induction because focuses attention on the framework and presuppositions
  • Guiding vs formal
  • Formal hypotheses are tested. They may be
    • Null
      • Sadly, statistical null Hs tend to dominate idea of hypotheses.
    • Non-null. Popper’s bold conjectures.
      • Require imagination. Not boring!
    • Rigour is in the testing process
  • Guiding hypotheses are explored
  • Not restricted to statistical approaches. In fact hypotheses are usually best avoided with statistics
fictional data
Fictional “data”
  • Made up data may be more convenient
    • E.g. confidentiality problems
    • Thought experiments
  • Fictions, fables, utopias, dystopias to explore …
  • Widely used in mathematical modelling
    • Interesting possibilities fabricated to play out what-ifs
  • Sometimes what is possible may be more interesting than what has actually happened
    • E.g. if we are interested in improving things
    • Dogmatic empiricism may be unreasonable? If we want to change things why focus exclusively on facts?
does all this matter we think so because
Does all this matter?We think so because …
  • Avoids impoverishing research by adhering to very restricted perspectives
  • Suggests new possibilities
  • Avoids wasting time talking rubbish
  • Fitting methods to the enquiry is important (e.g. check the two CRITIC acronyms in
  • Fitting them to your favourite paradigm is not!
  • All comments and suggestions welcome. These slides and a revised draft paper are at