Module 9: Independent groups Flashcards
The experimetal design
A reseatch methodlolofy that allolws researchers to calim casuality (x causes Y)
Formal experiments allow us to
establish causality
Experiments are often refered as the gold standard in resarch methodology because
they allow us to pinpoint a casuation
Observation methods+survey methods+corelataion design+ experiments difference
Observational methods may allow researchers to observe a children’s mathmatical behaviour without intervening, survey method allow researchers to gather opinions or attitudes about mathmatics, corelational designs can show the relation between the frequency of number word talk and mathmatic performance. Experiments can asses casuaity through the administration of treatment(s) or stimuli and careful observation of subjects in a controlled environment to determine effects.
Independent group designs (2)
Each participant has an equal chance of being assigned to one of the groups in the experiment (Treatment vs control). Paricipants are non-overlaping between groups
Random groups design
The simplest independent group design which includes 2 independent groups of participants allows researchers to establish causality
Random assignment
Participant in the sample has anequal chance in being assigned to one of the 2 groups
Operational definition
A percise definition of a construct such as mathmatical understanding which allow researchers to measure the construct
Control condition group
Do not receive treatment/intervention. The placebo condition group
Corelation methods are…..and helpful for circumstacers where….
non experiemental and do not lead to firm conclusions about causality and helpful for circumstances where we cant manipulate the IV
random assignment ensures that
the groups dont differ from one another based on factors that could impact your experimental outcomes
What happens with underpowered studies?
It is likely that a researcher may accept a null hypothesis (no group diff) because the sample is too small to have statistical diff
internal validity can be threatened when
there are confounding variables
Reseacehers can claim their research has internal validity when
they can say the variable that the manipulated was what caused the outcomes of their study
double blind study
Both the researcher and participant dont know which group they are assigned to
Matched group design
Researchers may chose to match their experimental groups on some additional measure (eg age) to ensure that tehir groups do not differ in other important ways other then exposure to the IV
Positive correlation
As A increase, B increases
Negative correlation
As A increase, B decreases
Zero corelation
There is no relationship between A and B *No pattern
What is r (4)?
r=1?+what it is+-+>0.5
When r=1 it means best fit to the line. It basically means how clustered the dots are on the line. If the number is - it means neg corelation, if its >0.5=strong
Effect size
How big of an effect between 2 groups
Power
The probabilitty of detecting an effect if it actually exist
More participants=
in terms of power/effect
More power to find an effect if it exist
Underpowered studies may not
replicate
Confounds decrease
internal validity
3 example of possible confounds:
- Diff in the environment (ex: room temp)
- Diff in the individuals (differential motivation for mathematics)
- Diff in the researchers (the way they act)
Single blind
Participants dont know
