Lecture 15: the experimental research strategy

define “experimental design”

planning a set of procedures to investigate a relationship between the IV and the DV

what do you need to design a controlled experiment? (3)

• testable hypothesis

• at least one IV that can be precisely manipulated

• at least one DV that can be precisely measured

what do you need to decide when designing an experiment? (4)

• how the IV(s) will be manipulated

• how to control for potential confounding variables

• how many participants/samples will be included

• how many participants will be assigned to conditions

what do you need to consider when building an experimental design?

internal and external validity 

why is the experimental strategy the most powerful research method?

it has high constraints on variables: you manipulate the IV to demonstrate its effect on teh DV while holding other potential influences constant

what's the goal of an experimental design?

to arrive at a causal explanation about impacts of IV on DV

how can you demonstrate a cause-and-effect relationship with the experimental design? (2)

• establish that the effect happens after the cause 

• establish that the IV is responsible for the change in the DV and not an extraneous variable 

what are the situations/conditions you compare?

present VS absent: condition with treatment VS without treatment

how can you try to establish causlity?

strict control: create an unnatural situation where variables are isolated from the influence of other variables

true or false: only the experimental design can establish causality 

false: quasi-experimental too

it’s impossible to prove causality, but what can we do?

show its likelihood relative to chance (alpha: alpha% of chance that it happened by chance)

*you don’t say that there is a cause-and-effect relationship, you say that the likelihood that this relationship happened by chance is alpha%

what do you need to establish to show if two variables are causally related? (4)

• time-order: cause (X, IV) happened before effect (Y, DV)

• covariation: change in the IV/X must be accompanied by changes in the DV/Y

• rationale: logical explanation for why the two variables are related 

• non-spuriousness: only the IV/X can cause changes in the DV/Y

define “time order”

the IV/X (cause) must have occured before the DV/Y (effect)d

define “covariation” (AKA statistical association)

changes in the IV/X must be accompanied in changes in the DV/Y

define “rationale” (AKA explanation)

there must be a logical and compelling explanation for why these two variables are related 

define “non-spuriousness”

only the IV/X can cause the changes in the DV/Y

*spurious: false, inaccurate

what are the types of experimental strategy? (2)

• between groups: 2+ samples are formed and are assigned to different conditions/IV

• within-groups: one sample and goes through all conditions/levels of IV, across time

what’s the difference between an IV, an DV and an extraneous variable?

• IV: cause, manipulated

• DV: effect, outcome, measured

• extraneous: any other variable in the study

between-subjects or within-subjects: 

• IV = lessons

• Treament: music lesson

• Control: no lessons

between: there is a control condition, only in between

between-subjects or within-subjects: 

• IV = time of testing

• DV = change in intelligence test scores from before-lessons to after-lessons

within: we look at the difference across time 

define “within-groups designs”

• one sample and all participants go through all conditions (levels of IV)

• we compare the results across conditions within the same participant

*usually across time

image

define “between-groups design”

• 2+ samples composed of different participants that will go through different conditions

• we compare the results between the groups

image

what are the groups that we only and always find in between-group designs? (2)

• experimental group: group that will be exposed to an experimental manipulation

• control group: group that isn’t exposed to the manipulation, will be used for comparison purposes

define “experimental group”

group that will be exposed to experimental manipulation

define “control group”

group that isn’t exposed to the manipulation and that is used for comparison purposes 

what are the possible control groups found in between-subjects designs? (2)

• placebo control group: participants receive a fake treatment 

• waitlisted group: participants will receive the same treatment, but later

define “placebo control group”

participants receive a fake (placebo) treatment 

define “placebo effect”

believing that the treatment will have an effect, even though you don’t necessarily get the treatment

what’s the goal of the placebo control group?

• we can compare the treatment group and the placebo group

• if there is a difference, then we can be sure that the treatments are the cause of the changes

what’s the goal of the waitlisted group?

we can control for the motivation (how do people with treatment perform VS how people in wait list perform)

true or false: if we put someone in the waitlisted group, we can change our mind and not give them the treatment

false: you need to give them the treatment

what are the basic elements of the experimental research strategy? (4)

• manipulation: researcher manipulates one variable by changing its values to create the condition

• measurement: second variable is measured for each condition, creating 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 that they do not influence the variable being examined

among the basic elements of experimental strategy (manipulation, measurement, comparison, control), what are the elements unique to experimental research? (2)

• manipulation: researcher manipulates one variable by changing its values to create the condition

• control: all other variables are controlled to be sure that they don’t influence the variables examined 

what is the first manipulation you should do in an experimental research strategy?

determine the amount of levels of the IV you want to have

what’s are the purposes of manipulation?

• determine the direction of the effect: you increase/decrease the IV to see what happens to the DV 

• control the influence of outside variable 

how can manipulating the IV help control outside variables?

it can ensure that the IV isn’t changing with another variable that can explain the change in the DV

how can you be sure that the changes in the IV aren’t caused by outside variables?

by controlling the third variable: if you can control the third variable, then you’re sure than the IV causes the change in the DV

what’s the goal of control? 

• make sure that there are no other factors contributing to the changes in the DV, other than the IV 

• we need to rule out alternative explanations for changes in DV 

what’s the difference between extraneous and confounding variables? (definition, impact)

• extraneous: affect the DV or the IV’,not both (extra variables that aren’t studied)

  • don’t matter if they don’t affect the DV/outcome or aren’t correlated with the IV

• confounding: affect both the IV and the DV

  • problematic: changes along the IV, we don’t know which one affects the DV

how can an extraneous variable become a confounding variable? (2)

• variables must affect the DV (don’t care if don’t affect the DV)

• must vary systematically with the IV (don’t care if they move randomly with IV, not a threat)

true or false: a variable that affects the DV with no relation to the IV is a threat

false: it isn’t

you are conducting a word memory experiment with 2 conditions (background music present/absent) and there is noise due to construction around the lab. what would be the confounding variable and what would be the extraneous varibles?

• confounding: noise if only one condition is affected, would impact the memory

• extraneous: noise if all the conditions are affected