Use LEFT and RIGHT arrow keys to navigate between flashcards;
Use UP and DOWN arrow keys to flip the card;
H to show hint;
A reads text to speech;
72 Cards in this Set
- Front
- Back
descriptive statistics
|
describe what's going on in the data you have (average)
|
|
inferential statistics
|
generalizations, inferences from samples to populations
|
|
parameters
|
something that you measure from the population
|
|
sample statistics
|
measured from a subgroup of the population
|
|
variable
|
trait within a study that changes
|
|
constant
|
doesn't change
|
|
nominal
|
categories are names, no order (cats, dogs, eye color)
|
|
ordinal
|
order (big, bigger, biggest) intervals between adjacent levels are not necessarily the same
|
|
interval
|
order; intervals between adjacent values are equal
|
|
ratio
|
has an order, intervals between adjacent values are equal, aboslute zero point (ex: weight, time)
|
|
contiuous
|
no gaps between adjacent values, infinite number of values (weight)
|
|
discrete
|
gaps between values, finite number of values (number of kids)
|
|
independent variables
|
variable you are observing/manipulating
|
|
dependent variables
|
variable using to assess whether independent variable has an affect
|
|
correlational
|
observing the relationship between 2 groups
|
|
measures of central tendency
|
mean median mode
|
|
linear transformation
|
changing a set of data by a constant (addition subtraction multiplication division) will not change the shape of the distribution
|
|
central limit theorem
|
the mean of the sampling distribution of the man is the same as the population mean
standard deviation of sampling dist of means is called standard error the shape of sampling dist of means if population is normal: sampling distribution will be normal; if pop dist isn't normal: sampling dist will approach normal when N gets big enough |
|
independent random sampling
|
each selection must be independent from all the others (selection of one doesn't influence selection of other)
|
|
if p<.05
|
reject the null (there is a difference)
|
|
if p>.05
|
fail to reject then null (no difference)
|
|
type I error
|
rejecting the null hypothesis when it is in fact true
|
|
type II error
|
failing to reject null hypothesis when it is in fact false
|
|
alpha
|
level of significance, lower prob of making a type I error
|
|
beta
|
prob of making a type II error
|
|
power
|
finding difference when there really is a difference
|
|
as N increases, t calc
|
increases
|
|
as N increases, standard error
|
decreases
|
|
as N increases, chance for significance
|
increases
|
|
confidence interval
|
measure of precision and error, estimate for the population mean, tests the null hypothesis
|
|
homogeneity of variance
|
variances are not twice as big as the other
|
|
fail to reject levene's test
|
do have HOV
|
|
reject levene's test
|
yes HOV
|
|
when to use two sample t test
|
quasi experiment, true experiment, only when dependent is interval or ratio scale
|
|
as alpha gets smaller
|
power gets smaller
|
|
as the alternative distribution moves to the right,
|
beta gets smaller, power gets bigger
|
|
delta
|
expected t value, center of alternative distibution
|
|
if n increases, delta
|
gets bigger, power gets bigger, beta gets smaller
|
|
.2
|
small effect size
|
|
.5
|
medium effect size
|
|
.8
|
large effect size
|
|
large sample size, less variability, confidence interval gets
|
narrower
|
|
power should be
|
.8
|
|
you can increase power by
|
increasing effect size
increasing sample size |
|
standard error between difference of means
|
difference between sample mean difference and population mean difference
|
|
large sample size? (>100)
|
use z test
|
|
are sample sizes equal?
|
pooled variance test for equal sample sizes
|
|
can the population variances be assumed equal?
|
yes= pooled variances test
no= separate variances test |
|
how do we test for significance using confidence interval?
|
if the null hypothesis value is in the interval
|
|
perfect correlation
|
an exact match one match with another (ex: final grades that are all exactly 10 points lower than the midterm grades)
1 = positive -1= negative |
|
pearsons r
|
used to represent correlation
|
|
positive correlation
|
variables moving in same direction
|
|
negative correlation
|
variables moving in different directions
|
|
bivariate outlier
|
a point that falls outside of the correlation that throws the correlation
|
|
as n increases, correlation
|
gets closer to rho
|
|
uses of pearsons r
|
reliability (.7 or higher)
validity (am i measuring what i think i'm measuring) |
|
reliability
|
test-retest
split-half inter-rater |
|
larger differences between M1 and M2
|
larger poewr
|
|
effect size
|
measure of separation/overlap between 2 population distributions
|
|
r^2
|
coefficient of determination
|
|
k^2
|
coefficient of nondetermination
|
|
when to use regression
|
prediction
statistical control regression with manipulated variables |
|
best fit regression line
|
smallest sum of squared residuals
|
|
repeated-measures
|
same people being measured more than once
|
|
degree to which matched t is bigger than the independent groups t for the same data depends on
|
how highly correlated the two samples are
|
|
as r increases, t
|
increases
|
|
increasing r increases
|
variability of difference scores
|
|
as correlation gets stronger
|
less error
|
|
role of correlation in matched t test
|
correlation changes denominator (r is subtracted)
|
|
matched pairs t test
|
match two people on similar traits
|
|
order effects
|
fatigue or practice
|
|
carryover
|
being exposed to the condition more than once
|