chapter 1 & 3
explain the theory-data cycle
scientists collect data to test, change or update their theories
what is bivariate correlation
correlation between two variables
matched groups
make sure groups are balanced with respect to a specific characteristic
placebo effects
feature improvements without receiving active treatment as a result of their beliefs and expectations
criterion and predictor variables
criterion - the one the researchers want to understand and predict (DV)
predictor - the variables that may or may not be predictive (IV)
what makes a good theory?
data supporting, falsifiable and exhibits parsimony
interrogation of association claims
effect size
is the correlation statistically significant
are there outliers
is the relationship curvilinear
is there a restriction of range?
between-subject design
each participant is exposed to one condition of the independent variable
instrumentation effects
results from a change in the way the dependent variable is measured from one time point to the next
cross-sectional - show how the 2 variables are related at each time point
autocorrelational - examine how each variable at one time point is associated
cross-lag correlations - usually the most important ones for making the case for causation
interrogating frequency claims
construct - how well a variable is operationalized
external - how generalizable are the results?
statistical validity - are conclusions are reasonable and accurate in terms of statistical analysis
what is a longitudinal design
within subjects designs
each participant is exposed to each condition of the independent variable
regression to the mean
a groups behaviour/status is extreme at the first time point, it is statistically likely to get less extreme over time
practice effects
participants get better at the task - or bored with the task - from session to session
3 criteria for causation
covariance - are the variables related
temporal precedence - does one variable happen before the other
internal validity - are there any other alternative explanations for the observed relationships
a technique that assess the association while taking possible third variables into account
posttest-only design and pretest-postest design
post-test only design - the dependent variable is measured once, after the manipulation has occurred
pretest/post-test design - the dependent variable is measured before and after exposure to the independent variable
double-blind placebo effect
participants and experimenters don't know if they are getting actual medication or not
counterbalancing (partial and full)
full - where every possible order is represented
partial - latin squares, each condition appears in each position at least once
frequency
causal
association
pattern and parsimony
involves finding support from a variety of correlational studies that approach the question in different ways but all point towards the same causal relationship
concurrent measures and repeated measures
concurrent - participants are exposed to the different conditions at the same time
repeated - participants are exposed to the different conditions in sequence with the dependent variable measured multiple times
attrition threat
attrition - relate to participants dropping out of a study and how that effects the comparison between time points
demand characteristics
seeing both conditions may cue the subject in on the experiments hypothesis