1/37
Looks like no tags are added yet.
one-tailed hypothesis
(directional) predict the relationship between variables
two-tailed hypothesis
(non-directional) predicts that variables have a relationship, but not how
inferential statistics
estimation and significance testing
estimation
confidence that sample represents population
significance testing
process of setting an error threshold and using it to see if results were statistically significant
estimates need to
pass normality assumption and have a random sample
normality assumption
the variable of interest is normally distributed in the population
normal distribution
theoretical distribution representing the location of deviations about the mean and the probability of deviations occuring
normal distribution needs to
68/95/99.7
68% are within 1 standard deviation, 95% are within 2, and 99.7% are within 3
kurtosis
abnormal distribution of pointedness/flatness; leptokurtic is pointed and platykurtic is flat
skewness
positive skew is high at the beginning then tapers off, and negative skew is opposite
central limit theorem
assume a normal distribution in a population; random sample is best; bigger the sample means more normal distribution
standard error of the mean
can’t get actual mean, so we assume there’s some error
confidence level
% confident we are in the result, set by the researcher; typically 95%
confidence interval
range of scores we’re % confident in
hypothesis testing
examines how likely differences between groups and relationships between variables occur by chance
statistical significance
the sample findings are present in the population
null hypothesis
statistical relationships do not exist; statistics say if we can or cannot reject the null
critical value
what needs to be met for significance to be shown
procedure for hypothesis testing
set significance level, find critical value, compute test, compare test statistic to critical value, null decision, and hypothesis interpretation
type 1 error
reject the null when the null was true
type 2 error
fail to reject the null when the null was false
power
probability of rejecting a null that is false
test of difference
tests h/rq of difference; nominal IV, scores are independent from one another, and relatively similar frequencies within IV groups
chi square
used when the DV is nominal or doesn’t exist
chi square goodness of fit
determines how well the frequency distribution for a sample fits the distribution of one variable
chi square independence
tests the differences in the categories of two variables to one another
t tests
do the means of two groups differ from one another
independent samples t test
each person is unique
paired samples t test
each person is paired in some way
anova
used when there are at least 3 groups
test of association
relationship between variables means there’s shared variance
measures of association
when one variable changes, the other does too
regression analysis
correlation techniques can be used to predict an outcome when only one variable is known
pearson’s r
independent scores, linear relationships, and correlational; direction and magnitude of relationship
restricted range and outliers for pearson’s r
makes it hard to detect relationships and can produce false correlations
effect size
how much variance is explained by statistical relationship, by percentage