test 3 Flashcards
descriptive stats are…
statistics used to organize/summarize info
inferential stats
tests that allow researcher to make inferences about some unknown aspect of population
statistical inference
provides tools to assess reliability of inferences made about population
central tendency is a…
descriptive stat
central tendency measures…
the center score in group of data (mean, median, mode)
mean
sum of all values in group divided by number of values group
median
midpoint in set of scores / use when there are extreme scores
mode
value that occurs most frequently
use mode when…
data is categorical
use mean when…
data has no extremes or categorical data
variability is…
how different each score is from the mean score (range, standard deviation, variance)
variability is measure of how much each score differs from the..
mean
range is…
difference between largest and smallest scores
standard deviation is…
average distance from mean
variance is…
standard deviation squared
what type of skewness are there
positive and negative
continuous iv + continuous dv (analysis method)
correlation
descriptive stats used for measuring…
frequencies
continuous iv + categorical dv (analysis method)
logistic regression
categorical iv + continuous dv (analysis method)
t-test/anova
chi square test compares…
observed frequencies with frequencies we expect under the null hypothesis
categorical iv + categorical dv (analysis method)
chi-square
(observed-expected)2 / expected
chi square formula
how can you tell if the chi square stat is significant?
greater difference between observed and expected frequencies lead to greater chi square
how to calc degrees of freedom?
(rows-1) x (column-1)
if chi square is > than critical value…
relationship is significant
phi and cramer’s v tell us…
the degree of difference between our variables
phi is used for
2x2
Cramer’s v is used for larger than
2x2
weak relationship if phi and cramer’s v is
less than .10
moderate relationship if phi and cramer’s v is
between .10 and .25
strong relationship if phi and cramer’s v is
greater than .25
most elementary method for comparing two+ groups mean scores
t-test
independent t-test
comparing means of 2 nominal (categorical) groups
paired t-test
comparing means of 2 variables, comparing mean of variable at 2 points in time (one group)
to conduct a t-test, you need
means, # of units in each group, standard deviation, degrees of freedom
the further apart the means are…
the more confident we are the group means are different
knowing what the ___ is very important and .05 value indicates significance
p-value; .05
when looking at significance on charts, look at…
sig. (2-tailed)
three ways to check for significance
- compare t-value to critical value (t-value should be greater than crit val)
- look at significance value (less than .05)
- look at confidence interval (sig if doesn’t cross 0)
what can anova test be used for?
can be used to simultaneously investigate multiple levels of independent variables (ex: political orientation)
__- way anova refers to
of iv’s
__- way anova refers to
of iv’s
if the F statistic is large, the difference is…
greater and more significant
can look at the ___ subsets (groups that are same as each other) to make things easier
homogenous
correlation determines the degree to which…
two ordinal, interval, or ratio variables are associated
correlation coefficient may take value between
-1.00 and 1.00
direction of a correlation coefficient indicated by…
positive and negative
magnitude of a correlation coefficient indicated by
1 (perfect correlation) and 0 (no relationship)
r is the correlation formula indicates…
average product of all pairs of z-scores
large sample size, r __ and small sample size, r __
goes down / goes up
regression is..
stat technique that relates a dv to one+ iv
when you want to summarize strength + degree of relationship btwn two+ numeric variables
use correlation
when looking to predict, optimize, explain # response btwn numeric variables (how x affects y)
use regression
this regression uses one iv to explain outcome of dv
simple linear regression
this regression uses two+ iv to predict outcome
multiple linear
logistic regression used for
categorical outcome variables (iv can be cont, categ, mix of both)
error term is…
residual variable that accounts for lack of relationship btwn iv and dv
