CH. 1: Research Methods in Psychology Flashcards
Hindsight Bias
Overestimate our ability to predict an outcome - belief that we predicted the current outcome prior to it
Confirmation Bias - Observer Bias
See what we believe
Judgmental Overconfidence
Overconfident about judgements
Seeing Order in Randomness
As in gambling superstitions - seeing a causal pattern where there is none.
Cultural Biases
e.g. Racism, Ageism, sex and gender
Observational Bias
See what we expect to see
Rosenthal Effect
High expectations lead to improved performance in a given area and low expectations lead to worse
Major Sources of Bias in Research - Threats to Validity
- Observer Bias
- Demand Characteristics
- Non-Random Sampling
Can be helped by: Double Blind experimental design
Falsifiability
Scientific propositions, hypotheses, are falsifiable (can be proven wrong)
- We do not say that a study—especially a single study—proves a hypothesis. Instead, we say the results of the study offer evidence in support of the hypothesis. We Fail to disprove the hypothesis when the results support the hypothesis
Science in psychology is inductive (going from the specific instance to a general implication) and based in the idea of probability (which involves statistical methods of estimating likelihood of an outcome being generalizable). Deductive reasoning is going from the general to the specific instance
Only deductive reasoning can logically prove something: going from a general proposition to a specific instance. All dogs are white, Lincoln is a dog, therefore Lincoln is white
Inductive reasoning: Going from a specific instance to the general proposition: Lincoln is white, Lincoln is a dog, does that establish that all dogs are white? It supports the idea of all dogs are white but does not prove it.
Psychological science is inductive.
Hypothesis Significance Testing
The results of a study are stated in P values, probability values: the probability that the positive finding is in error. That caffeine enhances memory is wrong for example. Usually stated as p =.05 or p = .01
Research study shows there is a relationship between the studied variables and there isn’t
False positive
Research shows there is no relationship between the variables and there is a relationship
False negative
Operational definition
Defining a measured variable in terms of how it was assessed
Replication
Can the study be repeated and get the same results
- The failure of replications, publication bias
- Absent replication we would be smart to not assume the results of the original work are accurate and not due to some extraneous variable in the procedures of the experiment
- May be occurring at higher rates in biomedical research than recognized, potentially undermining credibility and self-correction,
Validity
Does the research reflect the construct under study?
- Internal - The clarity of the measures and the variables
- External - Findings apply to settings and samples other than those in the particular experiment
- Ecological - Does it apply to real-world settings
Reliability
Consistency of measurement
Sampling
A sample of a population
- Random Sampling
Normal curve - population mean
The Experiment
- Causation unlike correlational research
- Manipulation: Creation of a pattern of variation in a variable in order to determine its causal power
- Independent variable: Variable that is manipulated in an experiment
- Dependent variable: Variable that is measured in a study
EXPERIMENTAL AND CONTROL GROUPS AND CONDITIONS
Control group - Experimental group
Random assignment
Experimental condition
Control condition
Sample
Population
Double Blind
Regression Toward the Mean
Daniel Kahneman - “ Thinking Fast and Slow”
Daniel Kahneman and Amos Tversky example Israeli Air Force Pilots
Can make differences seem related to independent variables - risk of Type I error (false positive)
Events repeated tend to regress to their mean
Mathematically, size of possible regression depends on natural variability of the phenomenon
Ex: Coin toss
Problem of small samples
Fooled by randomness
The controlled experiment is to overcome regression effects - control group
BE A CRITICAL CONSUMER OF PSYCHOLOGY
Reliability—the same values are obtained when the measurements are repeated
Validity—the method measures what it is supposed to measure
A study can be reliable without being valid but the reverse is not true
Bias
- Response bias
- Sampling bias
Replication
Statistics - Measures of Central Tendency
The Mean - the average
The mean of 4, 1 and 7 is 4
Median - divides top and bottom half
The median of 4, 1, and 7 is 4
The Mode - most frequent value
The mode of 4, 1, 7, and 7 is 7
Variability
Range - lowest and highest values
Normal distribution
Statistical Significance