HSC 403 - Quiz #3
z-test v.s t-test
Z-test: -1 equation -z-distribution -compare 1 individual/sample v.s population -population level data T-test: -2 equations -t-distribution -compare 2 samples -sample level data
critical values
how far from the mean our score needs to be for us to reject the explanation tat it happened by chance (always come from the tables)
noise
is a colloquial way of describing any sort of unexplained variability (ex: taking someones BP several times in a ro, you may get slightly diff values from reading to reading)
z-test
used to compare individual or sample to the population (ex: individual vs population or sample vs. population)
signal
scores that are affected by certain variables. these variables result in systematically higher or lower scores (comparison of a specific sample with the general population)
steps for hypothesis testing
1. calculate the stat 2. determine the critical value 3. compare the stat (from step 1) to the critical value (from step 2) 4. make a decision
t-test steps
1. state null & alternative 2. determine critical value(s) 3. calculate the t-score 4. compare the t-score to the critical value 5. make your conclusion
null hypothesis (H0)
no difference between the individual/sample and population we always test this hypothesis
one-tailed
only testing one end of the distribution we have a reason to expect either HIGH or LOW scores
two-tailed test
test for extreme scores at the both ends of the distribution the default in research
alpha
the % of scores for which we'll accept the Ha aka significance level
dependent (or paired) samples
the two samples are related to eachother (ex: pre- & post- from the same person or identical twins)
independent samples
the two samples come from unrelated groupes (ex: men v.s women, treatment v.s control groups)
alternative hypothesis (Ha)
there is a difference between the individual (or sample) and the population never test this hypothesis