general Flashcards
what is a z score?
tells us how many sd from the mean the score falls
what is type 1 error?
saying there is a difference when one doesnt exist
if CI overlap for two sets of data what does this imply?
we do not expect to see a statistically significant difference between the data
4 assumptions for parametric stat analysis
ratio or interval measurements, equality of variance, normal distribution, n>30
why is coefficient of variance useful?
it removes units and allows values to be compared fairly
what stat operation would we use to test for equality of variance?
levenes test, we want a value larger than .05 to assume equality of variance and continue with calculations
what stat operation would we use to test for normal distribution?
vodka test
if the CI for the true difference between means includes zero than _______
you cannot be confident that there is a true difference between the means
when calculating the U value for the Mann Whitney U test, there will be two values calculated, which one do you use?
the smaller U
how do we calc DF for a kruskall wallace ANOVA?
of groups - 1
for the signs ranked test what is the critical value and what must happen with the calculated value to reject the null hypothesis?
the critical value is .05, if the calculated value is higher than .05 fail to reject the null hypothesis
in which two tests do you want the calculated value to be smaller than the critical value to reject null hypothesis?
mann whitney U test, wilcoxon signed ranks test
in which nonparametric test do you rank data across the tables? —————->
friedman ANOVA, (one way repeated measures ANOVA non parametric operation)
how do we calc DF for a friedman ANOVA?
k-1, where k is the number of experimental conditions
how do we analyze nominal data?
use a 2x2 contingency table
what is ICC 1 useful for?
test retest reliability
what is ICC 2 useful for?
Inter-rater reliability, reliability of clinical tool
what is ICC 3 useful for?
intra-rater reliability, establishes reliability between trials and raters
what is a strong/good ICC output number from SPSS
above .80
what is a Kappa used for?
it accounts for chance agreements
what is a weighted kappa used for?
will address the varying levels of disagreement, more serious disagreements are weighed more heavily, less seriously are weighed less heavily
what type of data is ICC used with?
ratio/interval
what type of data is kappa, wieghted kappa, and percent agreement used with?
nominal data
how do we calc DF for pearson r?
of data pairs - 2
correlation is very _______ by sample size. A ______ sample size can result in a type ________ error.
skewed, small, two
4 pre requisites for causality
biologically plausible, logical time sequence, dose-response relationship, consistency of findings across several studies
for Spearman R the df of regression always equals ____
one
for spearman R the df of residual equals ____
N-2
y=a+bx what do the values in this formula represent?
a= constant
b=slope
x=value from a
y=value b to be calculated
when do you use spearman rho
if any of the data is ordinal
what is an assumption for an ANCOVA?
linear relationship between covariate and DV
more power means you will be ______ likely to make a type 2 error
less
factors effecting power
significance criterion, variance in the data, effect size, sample size
can we adjust the factors that affect power?
yes
as type i error ______ type 2 error _______
decreases, increases
what does effect size measure?
the degree to which the null hypothesis is false
when must you determine sample size?
a priori
If you do not know effect size what can you do to find a figure to use for estimating sample size?
guess, estimate from literature, pilot study, determine what would be a clinically meaningful difference and base numbers on that value
what is a large effect size index for ANOVA?
.50
what is a large effect size index for t tests?
.80