Reading...
Play button
Play button
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;
46 Cards in this Set
- Front
- Back
|
'Goodness of fit' Chi suqare tells us...
|
how representative is the sample? [compares overall sample to overall population distribution]
|
|
|
standardized residuals tells us...
|
how close individual values are to expected values (breaks down chi square)
|
|
|
if standardized residuals are more than _______ in magnitude, they are significant
|
1.96 [treated like z-scores]
|
|
|
kolmogorov-smirnov test is related to chi square, but it is appropriate for what type of variables?
|
ordinal
|
|
|
you can use these tools to locate blunders and outliers
|
frequency distributions and histograms
|
|
|
cross tabulations use ______ scale variables.
|
nominal
|
|
|
in chi-square tests, the "sig" number tells us...
|
the probability you will be wrong if you reject Ho. [if less than alpha .05, then reject]
|
|
|
two variables are statisticall independent if...
|
knowledge of one would offer no information to the identity of the other
|
|
|
the null hypothesis for statistical independence is..
|
Ho: the variables are statistically independent.
|
|
|
cross tab is used to...
|
find associations between 2 variables (not necessarily causation)
|
|
|
for continuous variables, the probably of the true mean = some exact number is..
|
zero!
|
|
|
the alternative hypothesis for ONE-WAY anova is...
|
h1: at least 2 of the means are different
|
|
|
t-tests is to compare 2 means, anova is to compare...
|
3+ means
|
|
|
for ANOVA, if h0 is true, the ratio of 'between treatement' and 'within treatment' should be close to...
|
1
|
|
|
TWO-WAY anova tests...
|
effects of 2 or more factors (independent variables) on a dependent variable
|
|
|
factors are also called...
|
independent variables
|
|
|
response variables are also called...
|
dependent variables
|
|
|
treatments are also called
|
levels (ie - different responses for independent variables)
|
|
|
what are the null hypotheses for ANOVA, where A is one factor, and B is another factor?
|
1) h0- no differences between treatment1a and treatment 2a
2) h0- no differences between treatment1b and treatment2b, 3) h0- no interaction between A and B |
|
|
univariate anova checks..
|
the INTERACTION effect
(if A is one factor and B is another factor, it tests the sig of A*B) |
|
|
univariate anova checks..
|
the INTERACTION effect
(if A is one factor and B is another factor, it tests the sig of A*B) |
|
|
True or False: When using regression, the dependent variable is on the X axis.
|
False - dependent variable is on the Y axis, independent variable is on the X axis
|
|
|
When using regression, what is the null hypothesis when testing significance of independent variables?
|
There is no linear relationship between the independent and dependent variables.
|
|
|
When using regression, what is the alternative hypothesis when testing significance of independent variables?
|
There is a linear relationship between the independent and dependent variables.
|
|
|
What is the null hypothesis when testing the significance of the intercept in regression?
|
Ho: The intercept = 0.
|
|
|
What is the null hypothesis when testing the significance of the intercept in regression?
|
Ha: The intercept is not zero.
|
|
|
R-square is also called...
|
Coefficient of Determination
|
|
|
What does coefficient of determination tell us, and what is its name in an SPSS output?
|
Tells us % of the whole picture you can explain with the given independent variables [regression model's ability to predict]. Name in SPSS output = R-Square!
|
|
|
the formula for R-square in simple terms is...
|
explained variation / total variation
|
|
|
correlation measures...
|
th degree to which there is an association between two or more variables
|
|
|
True or False: in correlation analysis, if r = 0, there is an absence of linear association
|
True.
|
|
|
True or False: in correlation analysis, if r = 0, there is no association.
|
False -- there may an association that is square or some cubed.. ( but r=0 says there is no LINEAR association)
|
|
|
Abdominal abscesses and peritonitis, pelvic inflammatory disease, diarrhea (ETBF strains) primarily in children 1-5yrs, and inflammatory bowel disease (ETBF strains in adults)
|
B. fragilis
|
|
|
What is multicollinearity?
|
When there are correlations amongst independent (predictor) variables.
|
|
|
What is the rule of thumb for reducing multicollinearity?
|
r between Xs cannot be larger than the largest r between X and Y
|
|
|
What is conjoint measurement used for?
|
understand consumer trade-offs [between price and quantity/quality of attributes] ; importance of attributes ; implications for design
|
|
|
Discriminant analysis is used when the ________ variable is _______.
|
dependent (y) ; categorical (x)
|
|
|
this analysis is used to classify individuals into one of two or more alternative groups on the basis of a set of measurements
|
Discriminant
|
|
|
the following function is considered what type of analysis: nonmetric dependent variable = f(metric independent variables)
|
discriminant
|
|
|
the following function is considered what type of analysis: metric dependent variable = f(metric independent variables)
|
regression
|
|
|
the following function is considered what type of analysis: metric dependent variable = f(nonmetric -dummy- independent variables)
|
conjoint (think attitude scores)
|
|
|
Factor analysis deals with data reduction in relation to the effect of _______ amongst variables.
|
multicollinearity
|
|
|
If testing correlations between factor A and factor B in factor analysis, what does "-1" in the "pearson correlation" signify?
|
Factor A and B are NOT correlated.
|
|
|
When testing correlations between factor A and factor B in factor analysis, and the correlation is close to 1, what does this mean?
|
Factor A and B are correlated
|
|
|
Cluster analysis is very closely related to...
|
market segmentation! (grouping people with similar interests together)
|
|
|
multidimensional scaling is commonly used with
|
perceptual mapping (measuring relative distances)
|
