What do experiments do that other types of studies may not be able to do?
Identify causal relationships
What are we doing in an experiment
A true experiment involves manipulation of one or more factors (IV) and the measurement(observation) of the effects of this manipulation on behavior.(DV)
Independent group designs have how many I.V.'s
Independent group designs have ONE I.V.
Independent groups design have how many levels (or conditions)? What are they?
Independent group designs have 1 I.V. That one I.V. must have at least 2 levels: (also called conditions)
What is the essential ingredient in an experiment?
Control is the essential ingredient of experiements.
How do we gain control in an experiment?
Experimental control is gained through:
holding conditions constant
What are the 3 conditions needed for internal validity?
An experiment has internal validity when it fulfills the three conditions required for a causal inference:
time-order events (relationship)
elimination of plausible alternative causes.
When confounding occurs is there internal validity?
When confounding occurs, a plausible alternative explanation for the observed covariation exists, and therefore, the experiment lacks internal validity.
Plausible alternative explanations are ruled out by:
holding conditions constand and balancing.
The factors a research controls or manipulates are called the:
In an independent groups design each group of subjects participates in how many conditions of the IV?
In an independent groups design, each group of subjects participates in only one condition of the independent variable.
What is another name for the independent group designs?
Between Subjects design
(because we are observing the differences between subjects)
Balancing is averaging subject characteristics (individual differnces) across the conditons of the IV.
Balancing means using random assignment to make sure groups receiving different levels of the IV are similar.
Researchers can use a pretest to make sure not all of one kind are in just one of the groups.
When random assignment is used to form independent groups for the levels of IV, the experiment is called:
random groups design
3 types of independent groups designs:
1) random group design
2) matched group design
3) natural group design
In a(n) ________, each group of subjects participates in adifferent condition of the IV.
independent groups design
The most effective independent groups design is one that uses:
When random assignment to conditions is used, the independent groups design is called a:
random groups design.
What is block randomization?
The most common technique for carrying out random assignment in random groups design; each block includes a random order of conditions, and there are as many blocksas there are subjects in each condition of the experiment. It balances subject characteristics and creates groups of equal size.Each block would consist of random arrangement of the five conditions.
What are the advantages of block randomization?
There are several
produces groups of equal sizes
controls for time-related variables
balances other time-related variables
balances any characteristices of participants
Define threats to internal validity:
an uncontrolled variable(s).
potential alternative explanations for a study's findings
What are 4 threats to internal validity?
"intact groups" - groups formed prior to study.
The systematic differences between two intact groups are almost guaranteed to threaten internal validity of the experiment.
Potential variables that are of no interest to experimenter but still may confound the experiment.
When a subject fails to complete the experiment
2 types of subject loss
1) mechanical subject loss
2) selective subject loss
mechanical subject loss
when subject fails to complete the experiment because of an equipment failure (experimenter reads wrong set of instructions, computer crashes, someone interrupts and experimental session.
Selective Subject Loss
when subjects are lost differentially across the conditions of the experiment
when some characteristic of the subject is responsbile for the loss
when this subjects characteristic is related to the dependent variable used to assess the outcome of the study.
Is internal validity typically threatened when subjects must be excluded from an experiment due to mechanical loss?
Is internal validity typically threatened when subjects must be excluded from an experiment due to selective subject loss?
Selective subject loss destroys comparable groups that are essential to the logic of the random groups design and can thus render the experiment uninterpretable.
influence the observer has over the behavior under observation.
Experimenters; expectations that may lead them to treat subjects differently in different groups or to record data in a biased manner.
cues and other information used by participants to guide their behavior in a psych study, often leading the participants to do what they believe the observer expects them to do.
How do you combat the reactivity/experimenter biases threat to internal validity?
double blind procedure
The extent to which findings from an experiment can be generalized to individuals, settings and conditions beyond the scope of the experiment.
Ways to increase external validity
replicating a study
partically replicating your study or
conceptually replicating your study
What kind of study can you do to have both internal and external validity?
What is a repeated measures design?
a research design in which each subject participates in all of the conditions. (measurement is repeated on the same subject in different conditions)
repeated meausres design: how many IV's and how many conditions?
one IV and at least 2 conditons, only difference from independent groups is that subjects participate in all conditions.
Another name for repeated measures design
studying changes within the same subjects across conditions
4 reasons we use a repeated measures design
1) only a few paricipants are available
2) more convenient and more efficiently
3) increase the senstivity of the experiment
4) some studies require it (ie. study changes in participants' behavior over time; or when its required that participants compare 2 or more stimuli relative to one another).
What is a more sensitive experiment?
The ability to detect the effect of the IV, even if it is small.
An experiment is more sensitvie when there is less variability in participants' responses within a condition of an experiment, less error variation.
The increased sensitivity of repated measures designs is especially attractive to researchers who study independent variables that have small (hard-to-see) effects on behavior.
can be due to variations in the procedure each time the experiment is conducted OR
can be due to individual differences
There is usually more variation between people than there is within people (thus error variation will generally be less in repeated measures design.)
The main advantage of repeated measures design:
there can be no confounds due to individual differences in a repeated measures design.
The bigegest threat to internal validity in a repeated measures design:
define practice effects
The changes participants undergo due to repeated testing:
1) can get better at a task due to better understanding or more practice at a task
2) may get worse due to boredom or fatigue
2 types of repeated measures design (and how they mainly differ)
they differe in the specific ways in which they control for practice effects.
practice effects are balanced for each participant by administering the conditions to each participant several times, using a different order each time.
each condition is administered to each participant only once. The ordering of the conditions is varied across participants instead of within participant (like in the complete design)
2 ways to randomize conditions
abba counterbalancing randomization
all of the conditiions of the experiment (a block) are randomly ordered each time they are presented
-to assist in randomly assigning participants to treatment groups
-to assist in randomly ordering conditions within subjects in a a repeated measures design
abba counterbalancing randomization
a random sequence of all conditions is presented followed by the opposite of the sequence.
if you can't use all possible orders in your experiment we can balance our design by using:
2 types of selcted orders
selected random order
What are complex designs
When 2 or more independent variables are studied simultaneously in one experiment.
2 or moer IV's
complex designs are also known as:
What is it called when you have one IV measuer within groups and another IV measured using between groups format?
What is the simplest type of complex design one may have?
2 X 2 (2 IVs, each having 2 levels)
HOw do you calculate how many conditions will be present?
The number of conditions can be determined by multiplying the levels of the experiment (2 x 2 = 4 conditions)
In a 2 X 4 X 7 X 9, how many iv's do we have and with how many levels and how many conditions would this result in?
2, 4, 7, & 9 levels
What is the main benefit of a complex design?
Identifying interactions between variables
What is the main effect?
The effect of each variable by itself
What is an interaction effect?
When the effect of one IV differs, depending on the level of the other IV
What format do we often use when reporting main effects and interactions?
Using descriptive statistics through
-tables (subtraction method)
-bar graphs (often used for main effects)
-line graphs (often used for interactions)
What does it mean to have a 3 way interaction?
When the interaction of 2 variables depends on the third variable in the study
What do complex designs have to do with validity?
When there is no interaction, the effects of each IV can be generalized across the levels of our other IV, so external validity of our IVs increases.
Having interaction creates boundaries by specifying conditions in which the effect of our IV occurs.
When performance reaches a maximum
when performance reaches a minimum
are ceiling and floor effects good or bad?
these effects create problems for interpretation
How are complex designs & natural groups design related? Are there any potential problems?
We use complex designs to make causal inferences about natural group differ.
Problems in determining what characteristic of our groups caused our outcome
What are the steps for making causal inferences from natural groups?
1) develop a theory (why should differences occur between the groups?)
2) Identify a relevant variable to manipulate (select an IV we can manipulate)
3) test for an interaction (test for interactions between your IV and individual difference variables to see if these variables are effecting your outcome.)
Matched Group Design
researcher makes the groups equivalent by by matching subjects
(by using a pretest task, aka matching task)
Natural group designs
individual variables are selected (age, race, gender, socioeconomic, etc.) rather than manipulated to form natural group designs.
represents a type of correlational research in which researchers look for covariations between natural groups variables and dependent variables.
Causal inferences cannot be made regarding the effects of natural groups variables because plausible aternative explanations for group differences exist.