Experimental Design (Sept 11/13) Flashcards
What is the difference between experiment design and non experimental design
Experimental design allows us the establish a causal relationship between 2 variables
Non experimental doesn’t
What is the purpose of ED and NED
ED - to establish a cause and effect relationship
NED- to observe and describe phenomena
How much control do we have over variables in ED vs NED
ED- high control over independent through manipulation
NED - no manipulation of variables, variables are observed as they naturally occur.
How does random assignment work in ED vs NED
ED - participants are randomly assigned to groups (control and experimental)
NED - no random assignment because groups are pre-existing or naturally occurring which would make it impossible or unethical.
When should you chose experimental design
- have a hypothesis about causal relationships between 2 variables
- you are able to manipulate the indepent variable
- you can randomly assign participants to different groups in the indepent variable
What is the difference between random assignment and random sampling
Random sampling is where we randomly pick participants from the population whereas random assignment is randomly choosing which participants are going to be apart of which group in the experiment.
When is a non-experimental design good?
- you are interested in a single variable
- you are interested in a non-causal relationship between 2 variables (correlation)
- you have an exploratory research topic
NED: what is cross sectional
You compare 2 or more pre-existing groups (you don’t create the groups
NED: what is correlational
You see if 2 variables are correlated
NED: what is an observational study
You observe the behaviour in a natural or lab setting without manipulation
True-experimental design vs Quasi- experimental design
True - random assignment is possible
Quasi- random assignment is unethical or impossible
How does intervention and random assignment work in QED vs NED
QED - an intervention is made but no random sampling
NED - no intervention is made, no random sampling or group manipulation
How does causality work in QED vs NED
QED - attempts to establish causality but less certainty due to lack of randomization
NED - cannot determine causality only correlation
Is there a need for comparison groups in QED vs NED
QED - often include comparison or control groups but not randomly assigned
NED - may not have comparison or control groups
Dependent vs independent variables
Depend - the variable you measure
Independent- the variable you manipulate.
What are levels in independent variable
Levels are groups within the indepent variable
Ex. Sex - male and female
What is an extraneous variable
Variables that are not the indepent variable but may have an effect on the dependent variable
(Example slide 15 experimental design)
What are confounding variables
An extraneous variable that potentially ended up affecting the dependent variable (example slide 16)
What is a lurking variable
A lurking variable is a variable that affects both independent and dependent variables
Pretext- Protest design
DV measured before intervention - intervention is made - DV measured again
If DV isn’t measured pre-intervention it’s not a pretext-protest or true experimental design
What is factorial design
When there are more than one indepent variable in a single experiemnt without the need to run separate experiments it is factorial.
What is a single-factor design:
when you have a single factor/independent variable in the experiment.
What is a univariate design
When you have a single dependent variable
What is a single-factor univariate design?
has one independent variable and one dependent variable
