Exam Flashcards
compare means of two independent variables with normal distribution
Students t-test
compare means of two dependant variables with normal distribution
Paired Student t-test
compare means of three or more independent variables with normal distribution
ANOVA
compare means of three or more dependent variables with normal distribution
Blocked ANOVA
compare means of two independent variables without normal distribution
Mann and Whitney U test
compare means of three or more independent variables without normal distribution
Kruskal-Wallis H test
compare means of two dependant variables without normal distribution
Wilcoxon test
compare means of three or more dependent variables without normal distribution
Friedman test
mutual dependence of two ordinal variables
spearman correlation
mutual dependence of two-scale, ratio or interval, variables
pearson correlation
influence of one variable over the other
linear regression
impact on many variables on one
multivariate regression
grouping of objects according to the valuables of the variables
cluster analysis
grouping of variables according to their relations
factor analysis
group separation and forecasting of inclusion into them in nominal variables
logistic regression
group separation and forecasting of inclusion into them in scale variables
discriminant analysis and logistic regression
pearson 0,50
strong relationship
spearman -0.22
weak relationship
Pearson spearman below 0.30
weak realtionship
pearson spearman 0.30 - 0.49
moderate
pearson spearman above 0.49
strong
correlation two variables with normal distribution and interval
pearson correlation
correlation of two variables interval ratio or ordinal
spearman
to measure the strength of a relationship and direction in terms of dependence and independence
regression
to make predictions
regression
dependence of independent scale variables in an ordinal dependant variable
not regression
how much depend a nominal variable from other nominals
not regression
dependence of a nominal independent on an interval dependent
regression
skewness 4,54
high to left. not normal
normal/close to normal distribution skewness
-0.5 to 0.5
approximately symmetric skewness
-0.5 to 0.5
high skewed
more than 1 or less than -1
moderately skewed
between (-)0.5 and (-)1
skew -0.6
moderate
normal kurtosis
3
nonparametric needs normal distribution?
no
examples of nonparametric
Mann and Whitney U test, Kruskal-Wallis H test, Wilcoxon test, Friedman test
Kruskal-Wallis H test, 0.54 sig
no statist difference of means
Friedman test 0.04
stat diff of means
Cronbach alpha 0.8
scales relate
inter item correlation matrix what to observe
values under 0.2, or negative.
KMO what does it do?
tell if factor analysis is suitable
Bartlett’s test of sparsity what is?
suitable for factor analysis has to be 0.05 or less. then there is a diagonal matrix.
communalities what are?
factors
extraction communialities suitable?
more than 0.2
principal axis factoring
for exploring new variables
rotated factor matrix
at least 0.4
kolmogorov normal distribution
if is more than 0.05
standarized coeficient beta
which independent variable has more influence on the dependent
VIF problematic
more than 2 can be problematic, more than 4 very problematic
VIF what indicates?
multicollinearity
Condition index, what means?
collinearity
condition index problem
more than 15
