Benefit of adding dependent variables to an experiment
A single IV can be the cause for multiple effects
Converging varints = incease strength that there is an affect present
what can you determine from a one-tailed t-test?
where there is a size difference between the groups
what can you determine from a two-tailed t-test?
whether the two groups are equal or not
student t-test assumptions
Sientific Integrity
values and beliefs on how scientist treat each other, data, and the public
Education Act (1989)
NZ unis must act as the critic and conscience of society:
Freedom of:
* testing and questioning wihtin the law
* engage in research
* regulate content being taught
* how best to teach
* staffing
Research Ethics and Te Tiriti
Partnership
Protection
Participantaion
Practice
Protection of Human Participants
Rerspect for persons - autonomy
Beneficence - risks + harms = benefit
Justice - burdens and benefit is equal and fair
Animal Etics Committee
Reduce - only use as many animals as nessasry
Refine - reduce pain and harm
Replace - dont use animals if possible
Type 1 error
False positive/alarm = rejecting H0 when its true and theres no effect
Type 2 error
False negative/ miss = failing to reject H0 when its false and there is an effect
test statistic
measures the variability tht is caused by manipulation (the effect), compared to the vaiability that occurs natuarely in the population
Independent t-test
between-subject
1 categorical IV, 2Levels
Continuous DV
t = independent samples
a group differnence / Standard Error of the Mean
when SE produces big t
when t is big or bigger when H0 is true
t is walys a bit bigger than 0 because of SE
sampling error
the variability of the observed diffence compared with the true diffrence
how the DV differ between samples due to extranious variables
sampling distribution
how often i should expect to observe diff t-stats values by randomly sampling from a single pop.
impacted by df = participants
tells us how often to exect a particular t-value is H0 is true
a big t stat =
unlikely to be from the same pop
two-tailed t-test
when looking for a diffrence that is non-directional
one-tailed t-test
when looking for a difference that is directional
two-tailed t-test and H0
we reject H0 if the t is far enough from 0 either way +/-
p-value will be double = more conservative
one-tailed t-test and H0
we rejcet H0 if the t is a big enough posive/ negtaive number
Assumptions of the Independent Student’s t-test
idependence
homogeneity of varince
normalicey
homogeneity is violated - Levene’s test is significant
Welch’s t must be significant to reject H0
normality is violated - shapiro-wilk’s test is significant
Mann-Whitney U-test
more powerful test (conservative) so the chnace of Type 2 is reduce