PSY 3044 - LECTURES > Evaluating Claims / Scientific Method

I. EVALUATING SCIENTIFIC THEORIES (and other CLAIMS)

A. Things to look for when evaluating a claim...

1) FALSIFIABILITY (testability) - a strong theory should theoretically be able to be proven wrong; it should make unambiguous predictions that can be tested against the facts
Beware of: Undeclared claims (too broad, vague) and Multiple outs ("excuses, excuses")

2) SIMPLICITY- the Law of Parsimony advises us on what type of explanations to adopt

All things being equal, the simpler of two accounts is preferred

3) GENERALITY - "Laws" of nature, explaining phenomenon

4) FRUITFULNESS (generativity)

5) AGREEMENT with AVAILABLE DATA (Comprehensiveness)

6) REPLICABILITY - safeguards against fraud, error, or coincidence ("co-incident, not causal)

B. Some other general guidelines...

1. What is the claim and who is making it?

2. Is the claim based on scientific observations?

3. What do statistics reveal?

4. Are there plausible alternative explanations for the claim?

"Absence of Evidence does not equal Evidence of Absence"

..but "Absence of disconfirming evidence is not the same as the presence of confirming evidence.."

Just because there is no evidence against a claim, that is not as good, valid as having evidence for your cliam, theory, etc.

II. The SCIENTIFIC METHOD

A. The scientific method sets apart formal vs informal theories about behavior

1) Observation (systematic to avoid error and bias)

2) Forming Hypotheses by inductive reasoning

a) determining cause and effect relationships

b) "Correlation does not equal Causation"

c) In terms of experiments, we can think of Cause as the independent variable (IV) and Effect as the dependent variable (DV)

d) Confounding variables: a variable other than the stated IV but than can affect the DV (and hence the interpretations of an experiment); confounding variables are usually said to systematically vary with the IV, if not they are called extraneous variables

e) Placebo effect (expectation effect) - single and double blind studies

f) Syntax: the rules and definitions that state how the IV and DV are to be measured and how they are related

3) Testing Hypotheses / Types of Data Collection/Studies

Case study - a single individual is studied in great detail

Naturalistic observation - The study of naturally-occurring behavior in its typical setting

Correlational research - Scatterplots and coefficients tell us the degree of relationship between variables

Survey research - A representative sample is polled on particular issues

The experimental method - The effect of IVs on DVs is measured

Measurements are compared between a control group and an experimental group

B. Statistics and psychologists

Descriptive or Summary (e.g. average / mean, mode) and inferential statistics (e.g., t-test, ANOVA) allow psychologists to reach conclusions from research

Research ethics - APA (and other Federal and State agency) guidelines for human and animal participants are closely followed