Good Theories

1
Good Theories Basic Methodologies

• Psych 231 Research Methods in Psychology

2
Properties of a good theory
3
Properties of a good theory
My theory by A. Elk. Brackets Miss,
brackets. This theory goes as follows and begins
now. All brontosauruses are thin at one end, much
thicker in the middle and then thin again at the
far end. That is my theory, it is mine, and
belongs to me and I own it, and what it is too.
Link to entire Monty Pythons My theory
transcript
4
Properties of a good theory

• Organizes, Explains, Accounts for the data
• If there are data relevant to your theory, that
  your theory cant account for, then your theory
  is wrong
• Either adapt the theory to account for the new
  data
• Develop a new theory that incorporates the new
  data

5
Properties of a good theory

• Organizes, Explains, Accounts for the data
• Testable/Falsifiable cant prove a theory, can
  only reject it

No amount of experimentation can ever prove me
right a single experiment can prove me wrong.
6
Omnipotent Theory
7
Omnipotent Theory

• Beware theories that are so powerful/ general/
  flexible that they can account for everything.
  These are not testable
• Karl Popper claimed that Freudian theory isnt
  falsifiable
• If display behavior that clearly has sexual or
  aggressive motivation, then it is taken as proof
  of the presence of the Id
• If such behavior isnt displayed, then you have a
  reaction formation against it. So the Id is
  there, you just cant see evidence of it.
• So, as stated, the theory is too powerful and
  cant be tested and so it isnt useful

8
Properties of a good theory

• Organizes, Explains, Accounts for the data
• Testable/Falsifiable
• Generalizable not too restrictive
• The theory should be broad enough to be of use,
  the more data that it can account for the better
• The line between generalizability and
  falsifiability is a fuzzy one.

9
Properties of a good theory

• Organizes, Explains, Accounts for the data
• Testable/Falsifiable
• Generalizable
• Parsimony (Occams razor)
• For two or more theories that can account for the
  same data, the simplest theory is the favored one

Everything should be made as simple as
possible, but not any simpler.
10
Properties of a good theory

• Organizes, Explains, Accounts for the data
• Testable/Falsifiable
• Generalizable
• Parsimony
• Makes predictions, generates new knowledge
• A good theory will account for the data, but also
  make predictions about things that the theory
  wasnt explicitly designed to account for

11
Properties of a good theory

• Organizes, Explains, Accounts for the data
• Testable/Falsifiable
• Generalizable
• Parsimony
• Makes predictions, generates new knowledge
• Precision
• Makes quantifiable predictions

12
Using theories in research

• Induction reasoning from the data to the
  general theory (data driven)
• In complete practice this approach probably needs
  a new theory (or an adapted one) for every new
  data set
• Deduction reasoning from a general theory to
  the data (theory driven)
• Here the theory (if it is a good one) is
  sometimes viewed as more critical than the data.
  
• It also will guide the choice of what experiments
  get done

13
The chicken or the egg?
Theory
Induction
Deduction
Data

• Typically good research programs use both

14
Research Approaches

• Basic (pure) research - tries to answer
  fundamental questions about the nature of
  behavior
• e.g., McBride Dosher (1999). Forgetting rates
  are comparable in conscious and automatic memory
  A process-dissociation study.
• Applied research Theory sometimes takes a
  backseat. This is research designed to solve a
  particular problem
• e.g., Jin (2001). Advertising and the news Does
  advertising campaign information in news stories
  improve the memory of subsequent advertisements?

15
Research Approaches

• Probably the best way to think of this is as a
  continuum rather as two separate categories.

• Often applied work may bring up some
  interesting basic theoretical questions, and
  basic theory often informs applied work.

16
Conducting Research An example

• Claim People perform best with 8 hours of sleep
  a night.

• How might we go about trying to test this claim?
• How should we test it (what methods)?
• What are the things (variables) of interest?
• What is the hypothesized relationship between
  these variables?

17
General research approaches

• Descriptive
• Observational
• Survey
• Case studies
• Correlational
• Experimental

18
Observational methods

• The researcher observes and systematically
  records the behavior of individuals
• Naturalistic observation
• Participant observation
• Contrived observation

19
Naturalistic Observation

• Observation and description of behaviors within a
  natural setting
• Can be difficult to do well
• Good for behaviors that dont occur (as well) in
  more controlled settings
• Often a first step in the research project

20
Participant Observation

• The researcher engages in the same behaviors as
  those being observed
• May allow observation of behaviors not normally
  accessible to outside observation
• Internal perspective from direct participation
• But could lead to loss of objectivity
• Potential for contamination by observer

21
Contrived observation

• The observer sets up the situation that is
  observed
• Observations of one or more specific variables
  made in a precisely defined setting
• Much less global than naturalistic observations
• Often takes less time
• However, since it isnt a natural setting, the
  behavior may be changed

22
Observational methods

• Advantages
• may see patterns of behaviors that are very
  complex and realized on in particular settings
• often very useful when little is known about the
  subject of study
• may learn about something that never would have
  thought of looking at in an experiment

23
Observational methods

• Disadvantages
• Causality is a problem
• Threats to internal validity because of lack of
  control
• Every confound is a threat
• Lots of alternative explanations
• Directionality of the relationship isnt known
• Sometimes the results are not reproducible

24
Survey methods

• Widely used methodology
• Can collect a lot of data
• Done correctly, can be a very difficult method
• Doesnt provide clear cause-effect patterns

25
Case Histories

• Intensive study of a single person, a very
  traditional method
• Typically an interesting (and often rare) case
• This view has a number of disadvantages
• There may be poor generalizabilty
• There are typically a number of possible
  confounds and alternative explanations

26
Correlational Methods

• Measure two (or more) variables for each
  individual to see if the variables are related
• Used for
• Predictions
• Reliability and Validity
• Evaluating theories
• Problems Cant make casual claims

27
Causal claims

• Wed like to say
• (variable X) causes (variable Y)
• To be able to do this
• The causal variable must come first
• There must be co-variation between the two
  variables
• Need to eliminate plausible alternative
  explanations

28
Causal claims

• Directionality Problem
• Airplanes and coffee spills

29
Causal claims

• Directionality Problem
• Airplanes and coffee spills

• Happy people sleep well
• Or is it that sleeping well when youre happy?
• Third variable problem
• Do Storks bring babies?
• A study reported a strong positive correlation
  between number of babies and stork sightings

30
Theory 1 Storks deliver babies
31
Theory 2 underlying third variable
32
The experimental method

• Manipulating and controlling variables in
  laboratory experiments
• Must have a comparison
• At least two groups (often more) that get
  compared
• One groups serves as a control for the other
  group
• Variables
• Independent variable - the variable that is
  manipulated
• Dependent variable - the variable that is
  measured
• Control variables - held constant for all
  participants in the experiment

33
The experimental method

• Advantages
• Precise control possible
• Precise measurement possible
• Theory testing possible
• Can make causal claims

34
The experimental method

• Disadvantages
• Artificial situations may restrict generalization
  to real world
• Complex behaviors may be difficult to measure

35
Next time

• Ethics in research
• Read chapter 3