Methods of Research in Psychology

1
Methods of Research in Psychology
2
Some really basic stuff.

• Basic assumptions
• External reality
• Time has direction
• Causation
• The universe operates according to rules
• How do you prove or disprove these?

3
Assumptions, Laws

• Law of gravity, of thermodynamics, of supply and
  demand
• Laws usually describe an event or process
• Laws have been confirmed so many times that we
  always accept them

4
Theories

• Summarize and organize experience (from
  observations and experiments)
• Generate testable predictions (hypotheses)
• Theories
• Are internally consistent
• Are compatible with available evidence
• Are tested over a wide range of evidence

5
Theories

• Can evolve to incorporate new evidence
• Are neither proven nor disproven we can discard
  them as being no longer useful

6
Hypotheses, Observations, Experiments

• Hypotheses are predictions, not descriptions of
  reality
• Observations and experiments are subject to many
  problems perception, design, interpretation
• In science, observations and experiments must be
  replicated

7
Scientific Attitude

• Humility
• Skepticism
• Creativity
• Curiosity
• Critical thinking Q1
• Examine assumptions

8
Scientific Method - Q2

• What are the elements of the scientific method?

9
Q3 - Methods of research

• Longitudinal
• Study the same cohort of HS students from
  freshman year to senior year
• How has this group changed in conformity over
  four years?
• Cross-sectional
• Study entire HS population at one time
• How do freshmen and seniors differ in conformity?

10
Descriptive methods of research

• Case Study Q4
• Why use a case study?
• Strengths?
• Weaknesses?

11
Q5 - Descriptive methods

• Naturalistic observations
• Strengths?
• Weaknesses?
• Clinical observations
• Strengths?
• Weaknesses?

12
Descriptive methods

• Benefits
• Often easy to do
• Provides information, descriptions of behavior
• May suggest ideas for research
• Dangers
• Overconfidence
• Hindsight bias
• No predictions, no cause effect
• Observer effects - behavior different when
  watched?
• Researcher effects - wording, etc
• False consensus
• Social desirability bias

13
Q6 Research design

• See research design handout

14
Q7 - Sampling

• Finding a random sample
• Finding a representative sample
• Sampling problems
• Limited population
• Self-identified subjects

15
Q8 - Confidence levels

• Research usually done to .95
• Determined in part by your sample size
• For more info, take a stats class

16
Q9 - Correlations

• Identify relationship
• Variables must be
• Height in inches OK?
• Male or female OK?
• Relationship
• Predictions?
• Cause and effect?

17
Correlation coefficients

• Q10 -
• -1.0 to 0.0
• 0.0 to 0.5
• 0.5 to 1.0

18
Correlation example

• Students who get more sleep every night do better
  in school

19
Operational definitions Q11

• get more sleep - measure this
• do better in school - measure this
• Can you use whether or not the student graduates
  as a correlation variable?

20
Operational definitions

• Define the variables so that the research can be
  repeated (replicated)
• Op defs must be measurable
• Op defs must be repeatable

21

• Collect data, probably with a survey
• Why would observations be difficult?

22
Q12

• Plot your data on a scatterplot graph
• The graph on the next page has a correlation of
  .83

23
(No Transcript)
24
Best fit - trend - line

• Best fit - trendlines
• http//argyll.epsb.ca/jreed/math9/strand4/scatterP
  lot.htm

25

• If correlations have predictive value, what can
  you predict from this graph?

26

• Lets try another
• Number of letters in last name vs height in
  inches. Hypothesis more letters taller
• What are op defs?

27

• Dangers of correlational studies?
• False consensus Q13
• Illusory correlation
• Survey issues
• Sample issues
• Wording effects Q14
• Researcher effects

28
Experiments

• Determine cause and effect
• Hypothesis - what is a hypothesis again?
• Subjects should be assigned randomly from a
  random and representative sample Q15
• Independent and dependent variables Q16
• Operational definitions
• Control and experimental groups Q17

29
Lets do an experiment

• Hypothesis women who drink a lot of alcohol in
  pregnancy will have children with lower
  intelligence.
• Where do we start?

30

• Identify the independent and dependent variables
• Define the variables operationally

31

• Do we need control and experimental groups?
• How do we choose our subjects?
• What sample do we choose our subjects from?

32

• Design the experimental procedure
• Q18 Double-blind / single blind

33

• What problems are we likely to run into?

34
Dangers!

• Procedural problems
• Overgeneralization
• Placebo effect
• Experimenter effects
• Ethical problems

35
Coming up

• Ethics
• Statistics

36
General Principles

• Do good
• Be responsible
• Have integrity
• Be just
• Be respectful

37
APA Research Ethics

• Informed consent Q19
• Inducements / force
• Pain Q20
• Deception
• Debriefing Q21
• Animal care
• Fabrication
• Plagiarism
• IRBs Q22, Q23

38
Statistics

• Measures of central tendency
• Measures of dispersion
• Bell curves and other curves
• Graphs
• Significance

39
Q24 - Central tendency

• Mean
• Mode
• Median

40
Dispersion

• Range
• Variation average of squared distances from the
  mean
• Q25 - Standard deviation - square root of
  variation

41
Range

• The difference between the largest and smallest
  data point
• How useful is this?

42
Variation

• The average of the squared differences of each
  data point and the mean
• Why square them?
• Is this useful?

43
Standard deviation

• The square root of the variation
• N? or N-1? Some thoughts for stats students
• Is this useful?

44
Normal curve

• Q26 - Or normal distribution, bell curve
• a bell curve
• http//acsweb.fmarion.edu/Pryor/bellcurve.htm
• Do all normal curves look the same?

45
Skewed distribution

• Q27 - Distribution other than normal
• Skewed distributions
• http//davidmlane.com/hyperstat/A11284.html
• Do all skewed distribution curves look the same?
• Positive/negative skew the tails direction

46

• Whats the mean and standard deviation on a
  normal curve?
• Is there a relationship between std dev and the
  normal curve?
• Do all normal curves have the same mean and std
  dev?
• Do all skewed curves have the same mean and std
  dev?

47
Graphs - displaying data

• Scatterplots and bar graphs more common in our
  work
• Many instructive, but silly graphs (PG-13 rated)
• http//graphjam com

48
Q28 - Within one SD

• With any standard distribution, the distribution
  will always be
• 68 within 1 SD of mean
• 13.5 between 1 and 2
• 2 between 2 and 3
• 99.7 within 3 SD of mean
• 68 95 99 rule

49
Chebychev

• For a non-standard distribution
• 75 with in 2 SD
• 89 with in 3 SD
• What is a z-score?

50
Q29 - Inferential statistics

• From our statistics, we can make predictions
  about the population as a whole.
• How accurate will these predictions be?
• We would apply laws of probability to our data to
  determine the accuracy of our predictions -
  essentially we will determine the likelihood that
  our results are real or the result of chance.
• For more info, take a stats class

51
Q 33 - Statistical significance

• Sample averages are reliable
• Average differences are relatively large
• Low probability that findings are purely the
  result of chance
• Then my findings are statistically significant