Intro (ch 1)
(Read each chapter after the lecture)
Illusory correlation: perceiving a relationship between two noteworthy (often uncommon) events
Consequences of “bad” reasoning: does it matter?
- Understanding planetary motion → technological advances
- bloodletting → Ouch
- Avoiding vaccines → increasing disease
- Antibiotics cure coughs → resistant bacteria
- Therapy for depression?
Knowledge-based on evidence
- Knowledge builds on itself.
- Natural biases in thought
- Learning not to be fooled
- Evaluating claims in the popular press
- “What do you mean?”
- “How do you know?”
- Extraordinary claims require extraordinary proof.
An experiment usually includes a comparison between groups that are treated differently
- Manipulate the independent variable, measure the dependent variables
- Statistics:
- Tools to decide if the groups differ significantly in terms of dependent variables
- Research methods:
- Tools for constructing groups so that the independent varilabe is the only explanation for our results
- Research methods → making good comparisons
Why the world would be a better place if the world were just like me
- Psychology = study of behavior
- To study behavior we want to be able to:
- Describe behaviors and characteristics
- Behaviors: how high I can jump, how well I can solve math problems, how fast I can run
- Characteristics: height, heart rate, favorite ice cream flavor
- There is variability in the real world
- Variability: a general “thing” we can measure for each person ( behavior or characteristic)
- Examples: height, introversion, IQ
- To describe everyone in the real world we use descriptive statistics
- Predict behavior
- Determine the cause of the behavior
Is my new teaching method more effective than my usual teaching method? Example
- Usual method: stand here and lecture
- New method: hand puppets
- Independent variable:
- Dependent varilabe:
The results of our experiment in the real world:
- Avg test score of puppy group: 85
- The average test score of the normal group: 65
To infer or draw conclusions about everyone in the real world we observe small groups of people and we use inferential statistics
- P-value:
- The probability of obtaining the result that you obtained if...
- There is no real difference between the groups you are comparing (for t-test, ANOVAm chi-squared)
- There is no real relationship between the variables you measured
- Alpha:
- The probability that is “rare enough” for you to conclude that there IS a relationship between variables or a difference between groups (often alpha = .05)
- So if p < alpha, you reject the null hypothesis
- When performing inferential statics on a computer, almost all statistical software packages report:
- The test static (r,t, F)
- The p-value
Ways to understand or gain knowledge about the world:
- Intuition
- Authority
- Scientific approach
Characteristics of scientific approach:
- Empiricism
- Use of objective observations to answer questions about behavior
- Test a specific idea
- Openly exchange ideas
- Peer review
Techniques for testing your ideas:
- Objective observations/description
- Control over the situation
- Match the quality of the evidence to the nature of the claim
Ways that we fool ourselves into believing what is not true:
- We don't understand randomness
- Gamblers fallacy:
- After a series of losses, the probability of a win is higher
- “Clusters” of a single type of outcome don't seem random
- We give too much weight to confirmatory information
- Friday the 13th
- Count how many good/bad things happen on that day
- But also count on the other dates
- We tend to believe what we want to believe
- Ex. survey of 1 million high school seniors
- 70% thought they were above average in leadership ability
- 2% thought they were below average
Ideas
- Idea:
- An assertion about the world that can be tested through objective measurement
- Ideas can be labeled as true or false depending on whether or not the evidence supports them
- Scientists test an idea by constructing arguments for or against the idea
- Arguments:
- The evidence and logic used to support the truth or falseness of an idea
- “The coin toss came up heads”
- “The earth is flat”
- When given an idea and arguments for or against the truthfulness of that idea, a person has to decide whether to believe that the idea is true or false
- Scentiitsit holds beliefs that are consistent with the degree of truthfulness of ideas
The life cycle of knowledge (Lecture 3)
- Get an idea
- Perform a research project
- Present research at a conference ( lecture, poster) <6 months
- Prepare and sub, a manuscript for publication < 6 months
- Peer review 6 months - 1 year
- Publication 1- 2.5 years
- Republication 1-5 years
- References in other journal articles 1-2 years
- References in books 3-5 years
Publication process
- Prepare and submit a manuscript
- Editor evaluate suitability
- They could reject the manuscript
- If they reject it… submit a manuscript to a different journal
- Revise methodology and run the study again
- OR distribution to peers for review
- Editor evaluates the reviews, makes a decision
- Could reject it
- OR author revises the manuscript
- Resubmit manuscript to an editor
- Send the manuscript back to peers for review (in step 2)
- Editor evaluates the reviews, makes a decision (again)
- Manuscript accepted
- Author reviews / edits proofs
- The paper appears in the journal
Goals of the scientific study of behavior
- Description of behavior
- “Pace of life”: LLevine and Norenzayan
- Ex. fast walkers have a quicker pace in life
- Prediction of behavior
- Determining causes of behavior ( conditions of determining causes of behavior)
- Temporal precedence
- Cause has to come before the effect
- Covariation of cause and effect
- Elimination of plausible alternative explanation
- Ex. effects of seating arrangement on class participation
- One lab is seated in a circle while the other lab is seated in rows
- Explanation of behavior
Ways to achieve goals
- Correlation research → description and prediction
- Mearuee two variables, determine if the relationship exits
- Experimental research → determining and explaining
- Control or manipulate one variable and look for changes in the dependent variable
Basic vs applied research
- Benjamin Franklin and electricity → he started by performing basic research
- He theorized about nature and lightning
- Observed that lightning is attracted to sharp points
- Basic research can lead to applied research:
- development of the lightning rod
- Psychical and psychological benefits
Lecture 4
A goal for this course: encourage critical thinking
Sources of research ideas
- Theory:
- A systematic body of ideas that organizes what is known about a topic from past observations and makes predictions about future observations
- Examples: gravitational theory, cognitive dissonance theory
- Research provides evidence for or against a particular theory
- Theories will amply support and become accepted fact
- Disapproved theories are modified or dropped
- Past research
- Observations
Theories
- To be used by us, theories must be falsifiable (there is some conceivable outcome that contradicts the theory's predictions)
- The most useful theories are those that can be wrong in most ways
Hypothesis
- Hypothesis:
- An assertion about what is true in a particular situation
- Ex. a person will enjoy work less if they are overpaid
- Hypotheses must be testable to be useful in research
- The hypothesis is NOT testable when:
- Concepts are poorly defined
- Ex. person committed a crime because they are “mentally disturbed”
- Hypothesis is circular
- This means the hypothesis does not say anything new
- Ex. Bobby is distractable and has trouble reading in school because he has an attention deficit disorder
- Inlvoveds non-scientific ( unobservable or unmeasurable) ideas or forces
- Testable hypotheses help us avoid after-the-fact explanations and confirmation biases (focusing on evidence that confirms our expectations
- Ex. rorschach ink blot test, barnum effect, jurassic park, cold reading
- The most useful hypotheses are those that are testable
Lecture 5
Examine the evidence
- “What do you mean
- How do you know?
Consider the plausible alternative explanations
- “What would happen if the conditions were any different?”
Ethical research
- Mlgrams obedience experiment
- There is an experimenter, a fake test subject, and a subject
- The subject gives “shocks” to fake test subject
- The subject doesn't know that the shocks are fake
- Fake test subject pretends to feel the shocks
- What they found:
- Subjects kept giving shocks because the person in authority kept telling them to
- Belmont report:
- Ethical principles and guidelines for the protection of human subjects of research
- 3 basic ethical principles of research:
- Beneficence
- Respect for persons (autonomy)
- Justice
Beneficence: maximize benefits and minimize risks
- Possible benefits:
- Education, treatment for the condition
- Satisfaction associated with contributing to research that may produce beneficial applications
- Possible risks:
- Physical harm
- Loss of privacy and confidentiality
- Naturalistic observation and privacy issues
- Psychological stress
- Debriefing
- Where the researcher reveals everything about the experiment
Autonomy (informed consent)
- An informed subject is capable of making decisions as to whether or not to participate
- Need an informed consent form
- Reseaosnable alternatives to particpation must be avaible
- No excessive inducements
- No technical jargon
Justice - selection of participants
- Make sure not to select just one group
APA ethics code: five general principles
- Principle A:
- Beenficne and nonmaleficence
- Principal B:
- Fidelity and Responsibility
- Establish relationships of trust
- Principal C:
- Principal D:
- Principle E:
- Respect for people's rights and dignity
Lecture 6
Institutional Review Board (IRB)
- At least 5 people one from outside the institution
- Submit an application to IRB for review and approval
IRB classifications of risk
- Minimal risk
- Greater than minimal risk
- Physical or psychological harm
- Privacy/confidentiality
One way to avoid placing a subject under stress:
- Ask how they think they would behave in a situation (rather than putting them in the situation)
- Problem:
- Subjects do not always know how they would behave in a situation
Withholding information and deception
- Sometimes fully informed subjects may alter their behavior
- Deception is used only when there are no viable alternatives
- It cannot be used if physical pain or emotional distress is likely
- If used, the subject must be informed as soon as possible
Problems with excessive use of deception
- Loses effectiveness as subjects begin to expect it
- Undermines public trust in experts
Debriefing
- The researcher reveals the full nature of the experiment
- The researcher assures that the subject has not experienced distress
- The researcher gets information about the subjects' perspectives on participation
- Good for both the subject and the researcher
Fabricating data
- Plagiarism
- Authorship
- Duplicate data
- Ethical obligations of reviewers
Different Variables
- Situational variables
- Describe the characteristics of an enviorment
- Ex. size of classroom
- Response variables
- Describe the responses of individuals
- Ex. recording where people choose to sit in a classroom
- Participant or subject variable
- Describe the characteristics of different individuals
- Ex. recording your age
Operating definitions
- In behavioral research, we must have testable hypotheses
- For our hypothesis to be testable, the variables in our hypothesis must be measurable
- Operational definition:
- Definition of a variable in terms of the operations or techniques that a researcher uses to measure or manipulate it
Operational definition examples
- Example hypothesis: boys show more affection toward their father than their mother
- Possible operation definitions: affection =
- Rating on a scale from 1-7 where 1=does not like at all and 7= very much
- Number of hugs shared with a person in a 10-min period where more hugs = more affection
- The method of measurement is clear and defines
- Another example
- Popularity of a course =
- Ratings on a course evaluation
Lecture 7
distinguishing the difference between correlational research and experimental research
we can perform correlational research
- Measure two variables
- Determine whether the variables are related to each other
- To do this…
- Compute a correlation (ex. Between the amount of sleep and GPA)
Experimental research
- To control and explain behavior
- Create two or more levels of the independent varilabe
- The researcher manipulates and controls the level of IV presented to each subject
Diifeencees between correlational and experimental research
- In experimental, the researcher randomly assigns groups
- In correlational, the researcher puts participants in groups based on characteristics
Lecture 8
Some types of validity
- Internal validity:
- The extent to which we can conclude here is a casual relationship between variables
- The biggest threat to internal validity:
- Plausible alternative explanations
- External validity:
- The extent to which our results can be generalized
- Construct validity:
- The extent to which our operational definition of a variable measures that variable
- Statistical validity:
- The extent to which our statistical conclusions are accurate
Nonexperimental vs experimental methods
- Nonexperimental:
- Observe or measure variables of interest
- Behavior is observed as it naturally occurs
- Experimental:
- Involves direct manipulation and control of variables
- Manipulates one viable and measures some aspect of behavior
- Nonexpermeint methods:
- Measure two variables to determine if they are related
- Also called the “correlational method”
- Ex. kids who watch a lot of aggressive TV shows exhibit aggressive behavior
- Cant determine the direction of cause and effect
- Third variable problem (confounding variable):
- A third variable may act as a cause of the two that were measured
- Useful if we are interested in prediction
- predictor: criterion (or outcome)
- Experimental research:
- Manipulation and control
- Manipulate one variable (IV) and measure the effects on another (DV)
- The logic of experimental design:
- Change only one thing (IV) and look for the effects of that change (DV)
Important questions to ask when evaluating claims
- Examine the evidence
- “What do you mean”?
- “How do you know?”
- Consider plausible alternative explanations
- “What would happen if the conditions were different?”
- The big question you should ask:
- “Compared to what”?
- Example: Examine the effectiveness of a particular instructional technique
- Prediction: using hand puppets will enhance learning
- Teach students to use hand puppets
- 75% of students get an A on the following exam
- Conlcue that hand puppets are an effective aid for teaching
- But… “compared to what”?
Experimental designs
- Two basic experimental signs / two ways to make comparisons
- Independent groups design (between-subjects design):
- Compare two different groups, treated differently
- Repeated measures designs (within-subjects designs):
- Compare one group to itself, before and after treatment (or under different conditions)
Independent groups design (between-subjects design)
- Ideally, start with two identical objects and treat them different
- Different groups experience different levels of the IV so the IV is manipulated “between groups”
- Easy to do in the physical sciences
- Hard to do in psychology
- How can you create equivalent “things” in a psychology experiment?
- An early approach to obtaining equivalence: matching
- Impossible to match subjects on all variables that might affect the outcome of the experiment
- Big discovery: random assignment
- Every participant has an equal chance of being assigned to any given experimental condition
- Random assignment equalizes groups
- Allows us to take advantage of what we know about the statistics of sampling
Types of Independent Group Designs
- Posttest-only design:
- Use random assignment to achieve equivalent groups before treatment
- Measure the behavior of both groups after the treatment
- Differences in behavior must be due to treatment
- Pretest-posttest design:
- Administer tests before treatment to assess the equivalence of the experiment and control groups
- Pretest does not guarantee equivalence
- Pretets might tip off participants to the purpose of the study
- Pretest useful if groups are less than 20
- Useful in evaluating reasons for mortality (dropping out of experiment early)
- Solomon four-group design
- Evaluated the effects of taking a pretest
- Ex. effects of caffeine on the ability to solve sudokus
- Group 1: pretest → treatment → postest
- Group 2: pretest postest
Lecture 9
Consequences of bad comparison groups
- Confounds:
- Factors that vary systematically with the IV (and thus provide alternative explanations for results)
- Not simply anything in the design that varies:
- Example:
- Petunia wants to determine which two types of plant food are most effective. She puts bob big blooms on half of the geraniums and puts them on her deck. She puts fems flowers on the other half and places them in the bay window of her kitchen. She then waits two weeks and measures the size of each flower
- IV: type of plant food
- DV: size of flowers
- Confound: location of flowers
- FFF flowers
- On kitchen window
- Fix: put some of each flower in both locations
- Example:
- Dr Fairchild would like to study whether people with visible tattoos are treated differently than those with tattoos. Each assistant is told to approach a shopper at the mall and ask for help carrying a box into a store. He found that people agree to help non-tattooed people more than tattooed assistants.
- IV: visible tattoos vs nonviable tattoos
- DV: How many people help
- confound: tattooed and non-tattooed were recruited from different populations
- Tattooed people:
- Non tattoed people:
- Fix: have one person wear a temporary tattoo or no tattoo
Repeated measures design (within-subjects design)
- Advantages:
- Fewer participants needed
- More sensitive to the effect of IV
- Disadvantages:
- Order effects
- Practice effects: performance gets better every time
- Fatigue effects: performance gets worse over time
- Another form of order effect:
- Contrast effects (interference effects):
- behavior in one condition affects subsequent behavior
- Soulton to order effects: counterbalancing
- Complete counterbalancing:
- All possible orders are presented ( to all subjects or different subjects)
- Two conditions: A, B OR B, A
- Incomplete counterbalancing:
- Not all orders are presented
- ABBA counterblancing:
- Run conditions once in one order and then again in reverse order:
- Counterbalancing does not eliminate order effects, it just spreads them across conditions
- Cant use repeated measures if treatment changes the subjects in some way
- Matched pairs design
- It is useful if only a few subjects are available
- Measure DV or something closely related
- Rank order results, and assign pairs of subjects to different groups