Stats semester 2 Flashcards
What is estimated standard error?
s / root N
how do you work out t?
t = (m-u) / s/rootN
v = N - 1
what is effect size (d)?
measure of difference in means taking into account the sd
d = (u1 - u2) / meanSD
meanSD = (theta 1+2)/2
how do you do a related paired t test?
find chance (post - pre) and do a normal t test.
what is the null hyporthesis for finding the difference between 2 populations?
mean A = mean B
what is variance?
sd ^2
what is t for an independent t test?
t = (mA – mB) / √( [SA2/nA] + [SB2/nB])
what is v for an independent t test?
(nA-1)+(nB-1)
nA+ nB-2
how can you measure correlation?
pearsons r
spearmans rho
describe pearsons r
allows us to conduct a parametric test to decide
whether there’s a real relationship between
variables (or if its quite likely that observed data
could have arisen just by chance)
describe covariance and sample covariance.
Covariance describes how much two variables co-vary
(the amount of variance they share).
sample covariance (C) is given by C = TC/(N-1)
Positive covariance - indicates that higher than average values of one variable
tend to be paired with higher than average values of the other variable.
– Negative covariance - indicates that higher than average values of one variable
tend to be paired with lower than average values of the other variable.
– Zero covariance - if the two random variables are independent, the covariance
will be zero. However, a covariance of zero does not necessarily mean that the
variables are independent. A nonlinear relationship can exist that still would result
in a covariance value of zero.
how do you work out r in pearsons r?
r = covariance/total individual variance.
between -1 and 1
r = c/ SxSy r = c/ sdx sdy
Table is set up for you to simply look in row N
V = N-2
???
what is the difference between pearsons r and a p value?
Pearson’s r tells you the strength/direction of the
relationship between variables
– The p-value tells you the probability that this correlation
coefficient could arise by chance assuming the null
hypothesis is true
A large Pearson’s r DOES NOT tell you that there is necessarily a
correlation between your variables
when do you use the 1 variable chi squared test?
We use this when we have one categorical variable
with at least 2 categories.
Used when we want to assess whether observed
frequencies in each category different from what is
expected
what does the chi squared test tell us?
the chi-square statistic tells us about
whether observed frequencies depart from what is
expected (which will be known).
equation for chi squared?
sum (E-O)^2 / E
(E-O) = sum of difference
1 variable - V = C-1
2x2 V = (C-1)x(R-1)
2x2 cell chi squared test?
expected =
multiply totals of column and row / grand total
then use sum (E-O)^2 / E
what are the non parametric tests?
spearmans rho (r) mann whitney and wilcoxon (t)
Non-parametric test are less powerful than their parametric
friends… but they are assumption free
how do you rank numbers?
remember equal values become the mean of the rank
how do you do spearmans rho?
Convert scores to ranks
• Calculate difference in
ranks
• Square the difference
equation?
when do you use wilcoxon and mann whitney?
Wilcoxon signed rank test (Wilcoxon) is an
alternative for the paired/related t-test
The Mann-Whitney U test (Mann Whitney) is an
alternative for the independent t-test
how do you do a mann whitney test?
rank regardless of condition, then find mean of each condition.
mann whitney tells us if difference in mean rank is significant.
how do you do a wilcoxon test?
calculate difference (after - before). rank the non zero difference scores, ignore signs but take ties into account.
how do you work out shared variance from r?
r^2
