Exam 3 Flashcards
Null hypothesis
assumes there is no difference between the populations from which the samples were drawn => aka no effect, means both = to one another
alternative hypothesis
says there is a difference between the populations aka the IV had an effect on the DV => reject the null if less than alpha
t-test
tests significance between the sample means
p-value
probability of obtaining the value of the statistic or a more extreme value if the null is true => also = to alpha
type 1 error
reject the null but the null is true => saying the IV had an effect but it didn’t (false positive)
–> also equal to alpha
type 2 error
dont reject the null but the null is false => saying the IV had an effect but we conclude it didnt (Miss)
–> equal to beta
power
the probability of correctly deciding the null is false => 1 - Beta
–> setting alpha relatively low sets Beta higher
T/F we can calculate Beta directly
false we cannot because its based on the alternative hypothesis and we dont know its exact probability
effect size
how much scores may differ due to the experimental condition
power analysis
given the effect size and level of significance, we can determine the sample size needed to detect the effect
multiple hypothesis testing
performing multiple tests increases the probability of committing a type 1 error
quasi experiments
lack internal validity due to lack of random assignment to conditions in the experiment
only one group posttest only design
single group of participants has a treatment and then behavior is assessed
one group pretest-posttest design
single group tested before and after on material => look to see for changes
history effects
events that occur during theparticipation that affects behavior
maturation
changes due to the passage of time that affect behavior
testing
taking a test may affect subsequent testing if you cannot separate the effects of repeated testing from the IV
instrument decay
changes in measuring instruments over time => includes observers
regression toward the mean
extreme scores are likely to be followed by more moderate scores
subject attrition (mortality)
participants selectively drop out of experiment => the only people left are the ones interested in the study and can perform the task
selection
when control and experimental groups are chosen in a way that they aren’t equivalent
nonequivalent control group design
uses an experimental group and control group but they aren’t equivalent (natural groups)
nonequivalent control group pretest-posttest design
shows if there is a difference in the groups at the beginning => we can use pretest scores and look for a change between groups
interrupted time series design
look at behavior before a treatment and then look at a behavior and measure it for some period of time after the treatment => look for changes in behavior
control series design
multiple group time series design => involves a control group
singe case experimental design (ABAB design)
researchers manipulates an independent variable and behavior is recorded during a baseline before the condition is changed => not a case study
multiple baseline design
used in stiuations where it would be difficult or unethical to remove the treatment after a period of time
multiple baseline across situations
measure baseline in several situations
multiple baseline across subjects
measure behavior of several subjects over time and introduce treatments at different times for different subjects
contamination
communication betweeen participants
experimenter expectancy/observer bias
treating participants differently
novelty effect (Hawthorne effect)
reactivity when participants behave differently because they know they are being studied
twin study
pheromone study where one twin wore pheromone but poorly controlled for confounding variables
developmental research designs
studies changes in behavior associated with age
cross sectional design
randomly select participants from different age groups and measure their behavior
ad/disvantages of cross sectional designs
Advantage: relatively fast
disadvantage: cohort effects
cohort
group of people born at the same time
cohort effects
due to unique circumstances of a particular generation rather than age itself
longitudinal methods of design
single group of participants over time being tested repeatedly
ad/disvantages of longitudinal effects
advantage: no cohort effects
disadvantage: can take a really long time, people may try to be consistent over time, reactivity
mortality/attrition effects
people dropping out of a study
multiple observation effects
improved performance can be from practice effects
sequential method
combination of cross sectional and longitudinal design
complex/factorial designs
2 or more IV levels are manipulated simultaneously within a single experiment => requires factorial ANOVA
main effect
deciding whether or not one IV has an effect on the DV
interaction
the effect of one IV on the other IV
mixed factorial design
one Iv is a repeated measures design and the other is an independent design
when you have a 2x2 factorial, how many F and P values do you get?
3 of each
external validity
the experiment is generalizable for other populations
internal validity
the experiment is logically designed so results tell us the truth
exact replication
repeat a previous experiment in an identical matter with only a different sample
conceptual replication
replication of the conceptual relationship between variables
- Different procedures and types of participants
meta analysis
set of statistical procedures that allow you to compare results across different studies