Chapter 10 Simple Experiments Flashcards
basic definition of an experiment
a study where at least 1 variable is manipulated and others are measured
control variables are also known as what?
extraneous variables or confounds
extraneous variable vs confound vs third variable
extraneous: variables that are not the focus of the research but may impact the variables being studied
confound: extraneous variable in EXPERIMENTAL RESEARCH that provides an alternative explanation
third variable: extraneous variable in CORRELATIONAL RESEARCH that provides an alternative explanation
how are confounds and third variables related to the variables of interest in their respective studies?
confounds vary systematically with the IV and MAY impact the DV
third variables are associated with both IV and DV (could be a spurious correlation)
what does “treatment” refer to in experiments?
the manipulated condition that researchers are most interested in; not necessarily medicine
IV levels vary. what are 3 types of groups seen in experimental studies?
treatment groups, comparison groups and control groups
comparison group
level of IV that differs from the treatment in a meaningful way
control group
neutral/no treatment condition
main type of control group
placebo group: group receives an inert treatment
true or false: every experiment needs a control condition
false
true or false: every experiment needs a comparison group
true
adding a control condition to an experiment may help tell what?
the direction of effects
confounds threaten which validity?
internal
what may be the cause of design confounds?
the IV operationalization/aspects of experimental procedure
systematic variability
the levels of a third variable vary alongside IV variation/dif levels of the IV
unsystematic variability
the levels of a variable vary independently of experimental groups
selection effects
characteristics of participants in one level of the IV are systematically different than that of participants in the other levels of the IV
2 ways to avoid selection effects
random assignment & matched groups
2 types of confounds
design confounds and selection effects
blocked random assignment
ensuring equal numbers of people are in each condition
goal of matched groups
ensuring groups are as equal as possible on important variables
3 steps to attaining matched groups
- measure an important participant variable
- put people with comparable scores in pairs
- within each pair, randomly assign 1 person to each condition
how is control of the IV obtained in experiments?
manipulation of the IV
how is control of extraneous variables and confounds obtained?
random assignment & holding variables constant on purpose
control over the IV can establish what?
temporal precedence
what 3 things are controlled in good experiments?
IV, confounds and extraneous variables
2 main points of independent-groups designs
-different groups of participants are on different levels of the IV
-participants’ characteristics are not matched
independent groups designs are AKA what?
between-groups design, betweem-subjects design
2 types of independent groups designs
posttest only
pretest/posttest
posttest-only design
participants are randomly assigned to IV groups and are tested on the DV ONLY ONCE
pretest/posttest design
participants are randomly assigned to IV groups and are tested on the DV TWICE (once before and once after IV exposure)
goals of pretest/posttest
show that RA made groups equal
track change in the DV over time
why is it sometimes problematic to use a pretest/posttest design?
the pretest may change participants behavior, at risk for bias
main point of within groups design
each participant is exposed to all levels of the IV
2 types of within groups designs
repeated-measures
concurrent-measures
repeated-measures design
participants are measured on a DV more than once, after exposure to each level of the IV
concurrent-measures design
participants are exposed to all levels of the IV at roughly the same time, and a preference is the measured DV
advantages of within-groups design
- participants are their own control group
- very powerful in the sense that the study can show a statistically significant result when an IV truly has an effect on the DV
- fewer participants required
which design typically has more power?
within groups
how can order effects threaten internal validity of w/in groups designs?
the order in which IV levels are presented creates a confound because one IV level can affect participants’ reactions to later IV levels
3 main types of order effects
practice effects, fatigue effects and carryover effects
practice effects
performance improves in later conditions only because participants were able to practice on the DV in earlier conditions (not because of the IV)
fatigue effects
tiredness and boredom can lead to worse performance in later conditions (as time goes on)
carryover effects
contamination carries over from one condition to the next/ effects of one condition carryover to other conditions
counterbalancing
different participants are exposed to different orders of the conditions
goal of counterbalancing
to control for order effects
full vs partial counterbalancing
full: all possible sequences of conditions are presented
partial: only some possible sequences are presented
when is partial counterbalancing typically used more often?
when there is a large number of conditions (by 4 conditions theres already 24 sequences, 5 conditions 120 sequences, etc)
what is a Latin square generator?
used for partial counterbalancing; it ensures each condition appears once at each serial position in the sequence and that each condition follows and precedes each other condition once
disadvantages of w/in groups designs
- order effects
- might not be practical or possible
- demand characteristics
demand characteristics
cues that lead participants to guess the experiment’s hypothesis
construct validity in experiments
how well were the DVs measured and how well were the IVs manipulated
manipulation check
an extra dependent measure to test the effectiveness of manipulation
pilot study
very simple study to test the effectiveness of manipulations of the IV
external validity in experiments
can the causal claim be GENERALIZED to other people/situations?
there is said to be a trade off b/w which 2 validities? why?
internal and external; bc experimental control may require artificial situations in the lab
internal validity in experiments
are there design confounds, selection effects or order effects?
statistical validity for experiments
how large is the effect?
Cohen’s d
the standardized difference b/w 2 means
more thorough explanation of Cohen’s d
consider the mean difference b/w groups and the spread of scores within each group
the larger d is, the larger the _______ is
effect
Cohen’s d cutoffs
0.2- small effect size
0.5- moderate
0.8- large