It's All Linear
Logistically Speaking
Puttin' the "No" in ANOVA
Transformers Michael Bay Hasn't Used Yet
Potpourri
100
β in the population model for linear regression is the same as what value in the sample regression model?
b (coefficient of the variable)
100
The biggest advantage to using logistic regression is that it always returns a value between 0 and what?
1
100
A two-way ANOVA analysis requires how many predictor variables?
2
100
What is the order of the following regression model? y=b0+b1x+b2x2
2
100
The final exam will be made up of two sections: 2/3 short-answer modeling questions, and 1/3 what?
Multiple choice questions
200
To put a categorical variable into a regression, we first have to change it into what type of variable?
Dummy/indicator variable
200
A positive coefficient for a variable means that as that variable increases, what happens to the probability that Y=1?
It increases
200
R2 is calculated using SSR and what other value?
SST
200
For a polynomial regression, including x3 in a regression means you must also include x and what other variable?
x2
200
Backward elimination always removes the variable with the highest what?
p-value
300
We can use R2 adjusted to compare models with different numbers of variables because which value in the ANOVA table is not guaranteed to either increase or decrease with the addition of a variable?
MSE
300
If we know as x increases by one unit, the odds of an event occurring increase by a factor of e, what is the coefficient of x?
1
300
If you use 100 data points to predict a model with 5 variables, what is your dferror?
94 (100-5-1)
300
You know you should consider a logarithmic transformation of a variable when what is true about its distribution?
Its distribution is right-skewed
300
If you have a data set where you aren't sure whether differences exist in the distributions of Y=1 and Y=0, which version of discriminant analysis should you use?
Quadratic (QDA)
400
Linear discriminant analysis gets its name because it creates a linear what?
(Decision) Boundary Line
400
A "Prob>ChiSq" value in the output for a logistic regression is equivalent to what more general term (usually used in the output for a linear regression)?
p-value
400
A significant p-value for a one-way ANOVA procedure means what must be true for all the mean values of the X categories?
Nothing (we can only say two of the means are not equal)
400
You should always compare a reciprocal transformation with what other specific transformation?
Second-order polynomial (an x2 transformation)
400
If you run a regression to predict how much sleep a BMGT 430 student will lose before the final exam using age, hours of studying, whether 430 is the student's first exam or not, and class rank (Freshman, Sophomore, Junior, or Senior), how many separate regression lines will you ultimately have?
8 (2 categories for first exam/not, 4 class rank categories; 2*4=8)
500
Mallow's Cp is used to check for what problem with linear regression?
Bias (specifically omitted variable bias)
500
If you have an odds ratio of 3/5 of getting an A on the 430 final exam, what is your probability of not getting an A on the 430 final exam?
5/8 (probability of 3/8 you will get an A)
500
When running the comparable dummy variable regression for a one-way ANOVA procedure, the intercept (b0) of the regression represents the mean of what?
Y (not the mean of the baseline group!)
500
After doing a logarithmic transformation of both X and Y, you find the coefficient for ln(X) is 3.05. This means as X increases by 1%, Y increases by a factor of what?
e0.0305
500
You want to predict how many hours straight you can expect to sleep once finals are over using data from past semesters on the number of finals you had and the number of hours you spent studying. If you run all possible regressions for this problem, how many models will you have to run?
4 (1 with no variables, 1 with number of finals, 1 with number of hours, 1 with both)
M
e
n
u