25 Questions

PRM Logistic Regression

Questions and Answers

- 1.Logistic Regression (LR) is very similar to ___ except the predictors do not need to be ____.
- A.
MANOVA, normally distributed

- B.
DFA, normally distributed

- C.
ANOVA, normally distributed

- D.
Factor Analysis, normally distributed

- E.
DFA, linear

- 2.In LR, the DV is a ___ variable, even though it represents a ___ probabilistic occurence of an event occurring.
- A.
Categorical, continuous

- B.
Continuous, categorical

- C.
Dichotomous, categorical

- D.
Interval, continuous

- E.
None of the above

- 3.The 5 assumptions are (1) independence (2) linearity (3) normality (4) homogeneity of variance & (5) non-multicollinearity. LR does not require:
- A.
1,2,3

- B.
2,3,4

- C.
3,4,5,

- D.
Is flexible with all 5

- E.
Is strict with all 5

- 4.Which one of the following is TRUE?
- A.
In MANOVA, the dv has to be categorical but an underlying continuous distribution

- B.
In DFA, the dv has to be categorical but an underlying continuous distribution

- C.
In LR, the dv has to be categorical but an underlying continuous distribution

- D.
Standard, statistical & heirarchical regression are all variable selection methods of LR.

- E.
C and d

- 5.The reduction in uncertainty or the degree of variance we have to explain is represented by: multiple regression uses ____, MANOVA and ANOVA use ____, and LR uses ____.
- A.
R squared, ss, -2LL

- B.
Ss, ss squared, -2LL

- C.
Ss, r squared, -2LL

- D.
-2LL, ss, r squared

- E.
-2LL, ss squared, r squared

- 6.Is the following True or False: In research, the significance of a finding is very dependent on sample size, but not effect size. Effect size will stay reliable always.
- A.
True

- B.
False

- 7.The nagelkerke R squared is a recommended effect size estimate to use in which statistical modelling?
- A.
MANOVA

- B.
ANOVA

- C.
DFA

- D.
LR

- E.
MR

- 8.In LR, the exponent of B, EB, tells us that something is NOT a significant predictor, if the EB is ___.
*In other words, it is equally probable that the skier fell or did not fall.*- A.
1

- B.
Higher than 1

- C.
Lower than 1

- D.
Close to 1

- E.
A fraction of 1

- 9.In the textbook example: In LR, if the EB is ____ then it is ____ that the skier will fall.
- A.
Below 1, less likely

- B.
Below 1, more likely

- C.
Close to 1, less likely

- D.
Close to 1, more likely

- 10.The use of EB is particularly pertinent in ___, whereby you are 4.3 times as likely or half as likely to do something.
- A.
Marketing

- B.
Environmental research

- C.
Health research

- D.
Tourism

- E.
Hospitality

- 11.Nagelkerke's statistic is the ____ which can achieve a value of 1.
- A.
Proportion of variation in the response variable explained by the specific model

- B.
Proportion of variation in the explanatory variable explained by the specific model

- C.
It is used as a measure of effect size

- D.
B and c

- E.
None of the above

- 12.If in LR you are predicting gender & are using dummy variables 0 and 1 for female and male respectively, the EB understood as a poor predictors will have an EB of ____.
- A.
Close to 1

- B.
Of 1

- C.
Higher than 1

- D.
Less than 1

- E.
None of the above

- 13.With reference to ques 12, an EB of .9 on 'information' is interpreted as:
- A.
You are less likely to have a high information score if you are male.

- B.
You are less likely to have a high information score if you are female.

- C.
You are equally likely either way.

- D.
The highest reading before 1.

- 14.In LR it is important to keep an eye on which dummy value is 0 and which is 1. This is:
- A.
True

- B.
False

- 15.LECTURE 4: MULTIPLE REGRESSION Predicting some dichotomous outcome with variables that are either continuous or dichotomous is:
- A.
MANOVA

- B.
ANOVA

- C.
DFA

- D.
LR

- E.
MR

- 16.The use of multiple variables to predict 1 variable and the variable being predicted or the DV is a continuous scale variable is:
- A.
ANOVA

- B.
MANOVA

- C.
DFA

- D.
LR

- E.
MR

- 17.The predictors (IV's) in MR can be:
- A.
Continuous

- B.
Dichotomous

- C.
A and b

- D.
Categorical

- E.
Whatever you like

- 18.The number of predictors do matter in terms of bringing up the
*R*squared and bringing down the*F*value in MR is:- A.
False

- B.
True

- 19.The usefulness of any single predictor will be gauged by its semi-partial correlation (SR). Is this true or false?
- A.
False

- B.
True

- 20.The general linear model, is:
- A.
The observed data times what you can model plus eror

- B.
Quite simply, the observed data is the result of what you can model plus error

- C.
Error plus any data and model varianace

- D.
Is an interplay between variance and error variance

- E.
None of the above

- 21.The difference between an ANOVA and a correlational design is that the ANOVA is ____ and ____ whilst the correlational design is ____.
- A.
Structured, involved, relative.

- B.
Experimental, looking for an effect, trying to prove relationships.

- C.
Experimental, looking for an effect, not looking for an effect.

- D.
Non-experimental, causal, not looking for an effect.

- E.
Non-experimental, looking for an effect, trying to prove relationships.

- 22.Standard MR uses ____ predictors, statistical regression uses ____, whilst hierarchical regression uses ____.
- A.
All, some, few.

- B.
All, minimum, subset.

- C.
All, some, subset.

- D.
All, best, subset.

- E.
They all use the same number of IV's.

- 23.Efficiency of predictors is a pretty important consideration when using MR. This is predominantly because of its placement in the numerator.This is:
- A.
True

- B.
False

- 24.In MR, what order do we check out (1) descriptive stats (2) inferential stats of predictors & (3) inferential statistics of equation?
- A.
3,2,1

- B.
1,2,3

- C.
2,1,3

- D.
3,1,2

- E.
1,3,2

- 25.LECT 4: Unique contributions of particular predictors (ie X1 alone & X2 alone) are ___ and joint contributions ___ assigned to any particular predictor.
- A.
Semi-partial correlation, are not

- B.
Partial correlation, are not

- C.
Semi-partial correlation, are

- D.
Partial-correlation, are