Quantitative and Qualitative Methods for Research

1
Quantitative and Qualitative Methods for Research

• Lecture 1
• Introduction and Psychology as a Science
• Alan Garnham

2
Course Documents

• At the moment there are problems with the
  official teaching pages for this course
• For my materials, for the moment please see
• www.biols.susx.ac.uk/Home/Alan_Garnham/Teaching/QQ
  M
• Dan Wright, who is taking the next five sessions,
  will probably also use his personal pages to post
  course materials

3
Courses over the Year

• Autumn Term
• Quantitative and Qualitative Methods for Research
  (this course)
• SPSS 1
• Spring Term
• Statistical Methods
• SPSS 2
• Summer Term
• Advance Statistical Methods (optional except for
  MRes)

4
Aims for the Year

• To be able to examine psychology research and
  methods critically.
• To have in depth understanding of the basics of
  statistical analysis.
• To being able to analyse (fairly complex) data
  using SPSS.
• To be acquainted with some advanced techniques
  and be able to apply these to data sets.
• Not to be afraid to learn new statistics.
• To understand some of the skills necessary for
  being a researcher.

5
Additional Aim

• To meet each other and learn about what different
  people are doing in other areas of psychology.

6
Background

• The course content is heavily influenced by
  what the UKs Economic and Social Research
  Council (ESRC) suggests for postgraduate research
  training.

7
Practicalities This Term

• Friday Mornings lectures (Alan Garnham, Dan
  Wright, Richard de Visser, John Drury, Pete
  Clifton).
• This is the Quantitative and Qualitative Methods
  for Research course
• Friday/Tuesday Afternoon Computing Workshops.
  (David Reby with demonstrators Sam Knowles,
  Gustav Kuhn, Dawn Watling)
• This is the SPSS 1 course (starting in week 2)

8
Assignment for Quantitative and Qualitative
Methods

• Assessed assignment
• Write a grant proposal (see the course handbook
  for more details).

9

• Statistics are about accurately describing
  the appropriate patterns in data to others using
  numeric, graphic, and verbal forms of
  communication.

10
Outline for this Term (QQM)

• Philosophy/Conceptual Issues
• Psychology as a Science, hypotheses, sampling,
  causation, measurement.
• Revisiting statistical concepts
• Chi-square, t-tests, power, ANOVA, correlation,
  regression, making good graphs.
• Qualitative Methods
• Phenomenological Methods (Richard de Visser)
• Discourse Analysis (John Drury)
• Ethics (Pete Clifton)

11
Psychology as a Science

• Public perception Are you reading my mind?
• Different types of Psychologist (Not all
  Scientists)
• Some are the off the wall therapists
• Are the professionals better?
• Harvard Medical School psychiatry professor (John
  Mack) estimated that 3.7 million Americans have
  been abducted by aliens
• BPS survey (Counselling and Clinical sections)
  1/3 believe memories cannot be created
• 97 believe the essential accuracy of Satanic
  Ritual Abuse memories
• Pendergrast (1996 Victims of Memory) what
  the therapists would have said about past lives
  or alien abduction memories (p. 586). Morton
  Oh, thank God we didnt ask them about that!

12
How Science Progresses

• Steady accumulation of knowledge
• Popper - falsifiability
• can imagine it being false, can imagine testing
  this
• Used in Daubert (1993)
• Kuhn Paradigms
• Einsteins gravitational theory completely
  different from Newtons (eg., mass completely
  different)
• Lakatos - core, generally accepted
• Competing theories (lots say this)
• At a micro-level (in Psychology) null
  hypothesis significance testing (NHST)

13
Cronbachs (1957) Two Disciplines of Psychology

• (APA Presidential Address)
• Experimental
• Correlational (non-experimental)
• NB - both are empirical - not all Psychology is
  empirically oriented

14
Group Work Experimental vs Non-Experimental

• Which is better for psychology?
• and no theyre both good, complementary
  answers
• Write down your answer and give two reasons.
• Get into groups of 3-5. Choose people who gave
  the same answer (ie., experimental or
  non-experimental) as you.
• Discuss (and introduce yourselves)
• What did you use in your project? Why?

15
Laws in Nature hypotheses / models in science)

• X -gt Y
• Want Chocolate -gt Have Chocolate
• iff Y if and only if
  X
• X Y different descriptions of same
  thing
• for every action there is a reaction an equal
  and opposite reaction is action equals reaction
  
• (Feynman et al., 1963, p. 10-2)

16
Types of Hypothesis

• Strict if X occurs Y occurs (X
  is sufficient for Y)
• If you want
  chocolate, you have it.
• if Y has occurred X must
  have occurred
• (X is necessary for Y)
• Only have
  chocolate if you want it
• Associative Y is more likely to occur if X
  than if not X
• Purely statistical, data
  driven
• Ceteris Paribus X -gt Y all other things being
  equal
• Want -gt Have unless Tea Bar closed
• The all other things being equal, or CP, is
  important.

17
All Other Things being Equal

• Suggests it is clear what the other things are
  that have to be equal.
• Leads to the notion of Experimental Control
  (other things are deliberately made equal, so
  that the X -gt Y relation can be studied)

18
CP (Ceteris Paribus) Laws

• Fodor (1991) argues that psychological laws must
  be CP (compare, for example, Boyles Law in
  Physics).
• BUTis proposing CP Laws cop out?
• X -gt Y all other things being equal
• Is this just saying
• X -gt Y unless it doesnt ?
• The conditions must fit in with the theories
• There are physical realizations of X (like the
  brain states for want chocolate), and certain
  completer events (like have money), that
  together lead to Y.
• Should psychologists be concerned with the
  conditions necessary for the effects?

19
CP Hypotheses

• the goal of psychological theory construction
  is not to predict most (or even all) of the
  variance it is to explicate the underlying
  mechanisms upon whose operation the variance
  depends (Fodor Pylyshyn, 1981, p. 154).
• NB - associations predict variance
• So what does this approach buy us?

20
Experimental Research and CP Laws

• Experimental research is associated with the
  search for/ assessment of CP Laws, but in a small
  number of situations.
• There exists some situation (a completer) such
  that X -gt Y
• Falsifying this falsifies only the strict
  (sufficient) X -gt Y
• Reliably predicting failures means the law needs
  to be adjusted
• Example (Script theory at least in 1977)
• People know they eat food and then pay for it in
  restaurants.
• Exceptions are fast food restaurants.
• Schank Abelson (1977) create a track for fast
  food places
• Can get circular ... fast food equate with pay
  first
• For science, exceptions need other predictive
  differences

21
Non-Experimental X Y associated

• Quasi-experimental
• females v males
• Cant conclude, for example, that gender effects
  something
• Can conclude that woman score higher on some task
• Naturalistic
• Often test relationships in multiple situations
• Can sample situations
• Ecological validity
• More difficult statistically
• Models describe patterns.
• They simplify our observations of nature. These
  may help us to understand how these patterns
  occurred.

22
Two Types of Scientific Goals

• Laws that describe the causal mechanisms
• Experimental Studies
• Models that describe the patterns in nature
• Correlational studies
• All scientists should be interested in both.
• There are proper procedures for both
• (which you are here to learn)
• Next week (Dan Wright on)
• Statistics are about measuring the effects of
  causes, rather than the causes of effects.