Basic Concepts & Research Strategies Flashcards
four major questions in psychology
- what question am i trying to answer?
- what kind of data should i collect?
- whom should i measure?
- how will i collect those measurements and what will they be able to tell me?
what is the goal of this question?
- what question am i trying to answer?
identify the relevant variables and relationship between them
- come up with your IV and DV
what is the goal of this question?
- what kind of data should I collect?
assign values to variables in a way that is valid, reliable, and objective
measurement
construct
ideas we care about that cannot be directly measured
- happiness
operational variable
things we can observe that tell us about the construct
- 12 point scale on a person’s current mood
operational definition
the way a researcher defines a construct an observable variable in a particular study
- a way to describe construct based on how the researcher chose to measure the operational variable
- operationalize: process of turning construct into variable
what is the goal of this question?
- whom should i measure?
study a representative subset of the population
sampling
sampling method
how the subset is chosen from the population
- ways to get people in your study
what is the goal of this question?
- how will i collect those measurements and what will they be able to tell me?
draw statistically valid inferences about the relationships between variables
research strategy and statistical inference
quantitative research
- assigning numbers to the variables you measure
- analyze results using statistics
qualitative research
- interviewing participants, focus groups, and observing participants
- data is analyzed based on theory
- researchers create narratives based on theory
correlational research
- variables are observed, not measured
- no random assignment
quasi-experimental research
- participants are grouped naturally but not randomly assigned to conditions
experimental research
- IV is actively manipulated
- participants are randomly assigned to condition
external validity
how well do results represent people/context besides those in the original study
example:
people - if you are only sampling psyc 70 students, your study does not fully represent the college student population because participants are different than the average person
environment - say you are doing an experiment about friendliness in the workplace. you instruct the participant to say hello to as many people as possible. however, people are more likely to not be as friendly because your study is a more formal setting. whereas, their usually workplace contains people they are familiar/friends with. therefore, your study does not fully represent the realness of the situation which may lead participants to act differently
generalizability
do the results apply to other studies, other people, or real-world situations?
generalizability aims to enhance a study’s applicability and validity to multiple situations outside of the experiment
threats to external validity
- sampling bias
- novelty effect
- operation definition
- poor ecological validity
sampling bias
non-representative samples make it difficult to generalize to the population
example:
- studying the effects of burnout on young people
- sample bias: only sampling college students
- not all young people go to college
novelty effect
being in an experiment can cause people to act strangely/not how they normally act
- leads to results that does not accurately represent human experience
operational definitions
results could be too specific to choice of operational definition
- trying to study attitude about climate change
- do people believe in climate questionnaire vs asking them to volunteer and measuring them how much they care
- these scenarios will give you different results which are not generalizable to the overall context
poor ecological validity
study context is too different from real-world conditions we wish to understand
- laboratory environment may be too rigid or unnatural to be similar to real life situations
internal validity
how sure are you that X causes Y?
- checking that the causal claim is correct
- requires eliminating alternative explanations for the relationship between X and Y
extraneous variable
anything other than the manipulated variable that might be influencing the observed effect
- when people eat more ice cream, murder rates go up
- extraneous variable: temperature goes up which causes both anyway
must prevent extraneous variable from beocming a confound
why cant you make causal claims from correlations?
correlations do not indicate the directionality problem
directionality problem: don’t know whether x cause y or y causes x
it is unclear which variable affects the other
third variable problem
some unidentifiable variable is responsible for the observed relationship between two variables