Lectures 7-9 Flashcards
what determines the probability that a sample of ratio-interval scale data was taken from a population with a pre-determined mean
One-sample t-test
what is an a priori constant
pre-determined mean (c)
(one-sample t-test, 2-tailed) H0
μ = c (calculated mean = pre-determined mean)
(one-sample t-test, 2-tailed) HA
μ ≠ c (calculated mean ≠ pre-determined mean)
(one-sample t-test) test statistic
t
(one-sample t-test) degrees of freedom
n - 1 (n = sample size)
what are 2-tailed tests
no direction of the difference (not less than or greater than) || can also denote if there is change or no change
what are 1-tailed tests
testing in a particular direction (more/less, better/worse, higher/lower)
What is HA in ratio-interval scale data
whatever the question asks, H0 will be opposite to the question asked
What happens when we have a -t in 2-tailed?
ignore (-) due to symmetrical t-distribution
What happens when we have a -t in 1-tailed?
check to see if HA is satisfied by inputting mean of the sample in order to ignore (-) || if HA is not satisfied, immediately accept H0
confidence intervals (CI) and limits
it can determined with 95% confidence that the mean of a population lies between two values (the lower and upper limit)
(confidence intervals) what does a smaller interval indicate?
a smaller standard error
what would happen if a 99% confidence interval was employed
the critical value = 0.01 = more confident the larger/wider the interval gets
what does the t-statistic calculate?
calculates the probability where the sample came from the population where H0 is true || t(0.05)(1 or 2 tailed)(DF=n-1)
What to consider when dealing with 2 samples of ratio-interval scale data
see if 2 samples have equal variances (variance-ratio test), compare the means of the samples, and the relationship between those samples (does S1 values increase while S2 values decrease or increases?)
What to consider when dealing with central tendency tests
are samples paired or independent? are the assumptions met?
(central tendency tests) independent samples
values in one sample are not in any way related to values in the other sample
(central tendency tests) paired samples
each value in one sample is associated with a particular value in the other sample (i.e.: via biological link)
why use paired samples?
more direct comparison, reduces the extra extraneous variation (controlled) || cancels out other variables that may affect data
(central tendency tests) two assumptions
RSNDP and homoscedasticity
RSNDP
Randomly Sampled from a Normally Distributed Population || tests if data is normally distributed
what test can test for normality
D’Agostino Pearson k^2
what terms describe for equal or unequal variances
Homoscedastic/Heteroscedastic