Lecture 14: The Experimental Research Design l Flashcards
what is the most powerful research method and why?
the experimental method because it contains high constraints on variables
goals of the experimental strategy
- to arrive at a causal explanation
- to provide a comparison of situations (conditions) in which the proposed cause is present or absent
the experimental research design
Manipulation of one variable (IV) to demonstrate its effect on another variable (DV) while holding other potential influences constant
how does the experimental strategy establish causality?
strict control; creating an unnatural situation where variables are isolated from the influence of other variables
what designs can establish causality?
experimental and quasi-experimental
is it possible to prove causality?
no, it is only possible to show its likelihood (alpha value associated with statistical test)
2 steps to establishing a cause-and-effect relationship
- Establish that the effect happens after the cause occurs
- Establish that one specific variable (IV) is responsible for changes in another variable (DV)
4 factors that one must establish to establish that two variables are causally related
- time order
- co-variation
- rationale/explanation
- non-spuriousness
time order
the cause x must have occurred before the effect y
example of time order
individuals arrested for domestic assaults tend to commit fewer subsequent assaults than similar individuals who are accused in the same circumstances but are not arrested
co-variation
changes in the value of x must be accompanied by changes in the value of y
example of co-variation
the more years of schooling, the higher the projected income
rationale/explanation
there must be a logical and compelling explanation for why these two variables are related
example of rationale/explanation
the relationship between childhood poverty and petty theft
non-spuriousness
it must be established that x and only x have caused the observed changes in y; alternative explanations must be ruled out
example of non-spuriousness
grade schools with larger libraries cause or produce students who are better readers (possible third variable = parents’ education)
two types of experimental designs
Between-groups design
Within-groups design
between-groups design
- Two or more samples (groups) are formed at random from a pool of subjects
- Each group is composed of different participants
- Each group is assigned to a different condition (value of the IV) and the values of each group are compared
experimental group
the participants in your experiment are exposed to an experimental manipulation
control group
a group in your experiment that is not exposed to the manipulation and that is used for comparison purposes
placebo control group
the group receives a placebo treatment with no real treatment
placebo effect
just believing that the treatment will have an effect can cause a response
effect of treatment/ placebo comparisons
they indicate the effect on DV beyond placebo effects
waitlist group
participants who sign up to receive the same treatment
strength of waitlist groups
it controls for motivation across groups
limitation of waitlist groups
comes at a cost: the experimenter must offer the treatment eventually
within-groups design
- Only one sample (group) is formed and each person participates in all conditions (levels of the IV)
- Values are compared across conditions within each participant (values from the same participant in different conditions)
independent vairable
- The variable we think is the cause
- The manipulated variable in experimental studies
- Is manipulated by creating a set of treatment conditions
dependent variable
- The variable we think is the effect
- Results from the manipulation of the IV (cause)
- The outcome (what is measured in each of the conditions)
extraneous variables
- Any other variable in the study
- They do not matter if they do not affect the outcome and do not correlate with the manipulated variable
- Ex. age, gender, personality, etc.
4 basic elements of the experimental research strategy
- manipulation
- measurement
- comparsion
- control
manipulation
The researcher manipulates one variable by changing its value to create conditions
measurement
a second variable is measured for each condition, resulting in a set of scores in each treatment condition
comparison
the scores in each condition are compared with the scores in other conditions
control
all other variables are controlled to be sure they do not influence the variables being examined
what elements of the experimental research strategy are unique to this strategy?
manipulation & control
steps of manipulation
- First, decide on the levels of the IV you would like to examine
- Manipulate one variable and observe the second variable: is it affected by this manipulation?
purpose of manipulation
to determine the direction of the effect
what does a finding of causality assume?
that the researcher has controlled all other vairables
manipulation of outside variables
Manipulation can also be used to control the influence of outside variables
active manipulation
Active manipulation of the IV allows us to ensure that the IV is not changing along with another variable that can account for the relationship you are interested in
when can you be confident that changes in the IV are not caused by an outside variable?
if the third variables are controlled
control
- Must make sure that there are no other factors contributing to the changes observed in the DV beyond that of the IV
- Must rule out all other possible explanations for changes in the DV (eliminating confounding variables that can influence the IV & DV)
confounding variables
- A special class of extraneous variables that changes systematically with an IV and can influence the DV
- Therefore, they can mask the true effect of the IV
- They matter because they vary systematically along with the IV and can affect the outcome
- When the IV and confounding variables change together systematically, we cannot conclude cause and effect relationships between the IV & DV
- Provide alternative explanations of the results
extraneous vs. confounding variables memory example
- You are conducting a memory experiment and there is noise due to construction around the lab
- Noise would be an extraneous variable if it affects all conditions
- If only one condition is affected and there is reason to believe noise will impact memory, it is a confounding variable because the effect of the noise on memory becomes confused with the effect of the IV on memory
confounding variables and internal validity
- Confounding variables are an important threat to internal validity: we don’t know if it’s the IV or the confounding variable that is causing the change in the DV
- The presence of confounding variables offers an alternative explanation for your results
- It introduces ambiguity
Schmidt memory study and confounding variables
Schmidt 1994 examined the effects of humour on memory but the beneficial effects of humour on memory were due to surprise
what makes extraneous variables become confounding variables?
- The variable must affect the DV
- The variable must vary systematically with the IV
how do we deal with confounding variables?
Examine extraneous variables for possible influences on the DV
three categories of extraneous variables
Environmental variables
Partcipants variables
Time-related variables
importance of control
We need to introduce experimental control so that extraneous variables do not become confounding variables
5 ways to control for possible confounding variables
Remove them
Hold them constant
Use a placebo control
Match them across conditions
Randomize them
