paired t-test + effect size Flashcards
what are the 3 different measures of effect?
statistical significance, 95% confidence interval, effect size
explain statistical significance
p is less than 0.05, there is a difference that is probably not due to chance
how would a statistical difference of p=0.011 be written up in research?
there is a 1.1% chance that we would have found a mean difference as big as the one in our sample, if there was no difference between these groups in the population
_________ is a standardised measure of how important the difference/effect is that you found in the sample
effect size (how important is the effect)
what is the cohen’s d used for?
estimate of magnitude of effect (effect size)
what 2 statistical tests is cohen’s d used for?
independent and paired t-test
how is the p value stated in SPSS output?
sig(2-tailed)
why is effect size a useful measure?
avoids problems that the p value has with big sample size, as when N is big you’re likely to find a statistically significant effect regardless of whether this effect is important or not.
can interpret the magnitude of effect independent of the scale and values used, so can compare studies that use different measures
interpretation of cohen's d value: less than 0.2 = 0.2-0.5 = 0.5-0.8 = more than 0.8 =
trivial effect,
small effect,
medium effect,
large effect
what does a negative cohen’s d value show?
result showed the opposite effect of what was expected
what assumptions need to be met with the paired t-test?
data needs to be approx normally distributed, no clear outliers.
don’t need to worry about variances because its within participants
what is the number of data that you can freely vary and still get the same descriptive stats in your sample?
degrees of freedom (df)
what is the equation of df?
df = n-1
total sample size - 1 person
