Chapters 8,9,10 Flashcards
what if we want to study a subject variable?
this would be a correlational study, when randomization between groups is not possible but we still want to focus on the effect of a manipulated IV we use: preexperimental designs, quasi-experimental designs
one shot study
one group is tested only once
why would we use a one shot study?
initial explanation of a new phenomenon or intervention, gives researcher a basic understanding of the potential effects before conducting more complex experiments
nonequivalent group design
static group design, no random assignment
why use static group design?
main reason is to evaluate the effect of a treatment
pretest-posttest design
one group of people is tested twice
quasi experiment
adding more observations and/or more comparison groups, lets us make more conclusions with greater confidence
pretest-posttest nonequivalent control group design
using an experimental group with a comparable but not equivalent control group, both groups are tested before and after the introduction of the IV
time-series design
used when there is no appropriate nonequivalent group
multiple time-series design
combination of the time-series design and the pretest-posttest design with nonequivalent groups, multiple observations of the experimental group and nonequivalent group, makes the best out of a less than ideal situation, groups are not equivalent but they are made as comparable as can reasonably be
factorial designs
a design that assess the effect of two or more independent variables (also called factors) on a dependent variable
asking questions
will the type of feedback impact the participants self-confidence regarding their ability to do something
hypotheses
with this in mind, you really have 2 alternative hypotheses
IV and levels
can be multiple independent variables and they can have different levels
main effects
the effect that each factor has on the dependent variable
think of it this way:
it’s almost like conducting two separate studies at once
what about the prediction?
the researcher predicts a significant main effect
why is that okay?
researchers need to make predictions about every possible effect in a study, but it is important to be aware of potential effects
interaction effect
the effect that two or more factors have in combination of the DV, is each level of the IV affected the same way by the other IV or are they affected differently?
marginal means and main effects
the mean score for each level of an IV collapsed across the other IV
some things to consider
demand characteristics, will the participant be able to guess the study
interpreting main effects
most common to examine factorial design using: ANOVA, analysis of variance, statistical procedure that compares means between groups, tells us if main effect and/or interactions are significant
watch out for interactions
with main effects they ignore all other factors, if there is no significant effect you can generalize the main effect to each level of that IV, if there is a significant interaction effect you must be careful with how you interpret the main effects
mixed designs
factorial design that also has within group factors
