Foundations Flashcards
Experimental method
Involves manipulation of one or more independent variables to observe the effect on a dependent variable. Allows researchers to establish cause-and-effect relationships. Control groups and random sampling needed.
ex. a study testing the effectiveness of a new medication
Correlational method
Examines the relationship between two or more variables without experimentation/manipulation. Correlation coefficients (r) indicate the strength and direction of the relationship (positive, negative, or none). Can be affected by third variables. Correlation may not mean causation.
Cross-sectional studies
Examines different groups of participants at a single point in time. Useful for comparing different age groups, demographics, or conditions. Provides a snapshot of data but does not track changes over time. Can identify correlations but not causation.
Meta-analysis
Combines results from multiple studies (maybe hundreds) to identify overall trends and effects. Increases statistical power and provides a more comprehensive understanding of a research question.
Quasi-experimental design
Involves comparison between groups without random assignment. Useful when random assignment is not feasible or ethical (e.g., studying existing groups). Can suggest causal relationships but is more vulnerable to confounding variables. Often used in field settings where controlled experiments are impractical. Results should be interpreted with caution due to potential biases.
Ethical Considerations
“CARD/UD”
Consent
Anonymity
Right to withdraw (leave)
Debrief
Undue Harm
Deception
Research Hypothesis
This is the one you believe. It clearly predicts that there is a relationship between the dependent and independent variables. For example, one could predict that the way you frame questions about a student’s future to them, might impact how they behave
Reactivity
simply being observed can change your behavior. This is one of the problems with lie detector tests, some people will show spikes in anxiety just by being ASKED certain questions, even if they aren’t guilty.
Interference effects
one condition in the study impacts your ability to do the others
Confirmation bias
finding what you expect to find because you accidentally only look for info that confirms your hypothesis. If a person believes in astrology, they may only recall, or look for, instances in which their horoscope was correct. They’ll ignore and dismiss the instances in which it did not apply.
Hindsight bias
the “I knew it all along” phenomenon. After we seen the outcome of something like a sporting even or an election, we overestimate our ability to have actually predicted this thing.
P-hacking
Shooting an arrow and painting a bull’s-eye around it afterward. Altering your hypothesis as you go along, or picking only the best stats and data that make your experimental results confirm what you want.
Bidirectional ambiguity
chicken or the egg problem. Does exercise make people happier or do happier people just exercise more?
Internal validity
AKA construct validity. This is when your experiment is measuring what it claims to be measuring, with no extraneous variables or confounds. Conclusions and implications can be made using your data.
External validity
your experimental results can be generalized to other types of similar situations.
Ecological validity
your experimental results are also shown to be true outside of a lab/experimental situation. e.g. removed from demand characteristics and fMRI machines.
Mean
A measure of central tendency. The “average” amount of a distribution of numbers. Do not use in skewed data sets where there are extreme outliers. Outliers make the “average” a misleading number.
Mode
A measure of central tendency. This is the number that occurs the most amount of times in a distribution of numbers. Use in a bimodal distribution. So a list of numbers where there are two different clusters of central tendencies.
Regression toward the mean
When an outlier happens like you sleeping 14 hours one day rather than your usual 6 hours, eventually you will revert back to your “normal” or average. So that 14 hours was a fluke, and you will return to your normal mean. The same is true with most outliers in any situation.
Standard deviation
This is the average amount that data is different from the mean of the entire data set. For example, you make a 60 and a 100 on tests. The mean is 80 for those tests. If you made a 90 and a 70 though, the mean would also be 80. These two data sets have the same mean but the standard deviation is different.
RRR of animal testing
Reduce, refine, replace. We should reduce the number of animals tested, refine our experiments to lessen the discomfort, and replace animals with other methods of testing if we can.
Reductionism
A common criticism of bio psych. You can’t understand the video game Elden Ring simply by looking at the physical parts of an Xbox. Similarly, many argue that a person is much more complicated than the parts of their brain. Should we, or can we, really reduce human behavior to brain shapes and the chemicals that flow within them?
