Professional Documents
Culture Documents
Study Programme
BScB, 2nd semester
and level
Term S17o
Supplementary
All X Specified No
material/aids
Hand-in of hand-writ-
Yes X No Comments: Graphs, formulas and text
ten material allowed
Number of pages
4
(incl. front page)
Your exam paper must comply with the following format requirements (read carefully):
Add page numbers and total number of pages on all pages of your paper (e.g. 1 of 15, 2 of 15 etc.)
Your exam paper MUST be handed in as one PDF file, but additional material/appendices may be
uploaded in other file formats.
The file name must be the sequential number allocated to you in WISEflow AND the name of the
exam. Please insert this in the page header on all pages as well.
This exam is anonymous, so please do not write your name or student ID number anywhere.
Other instructions:
Regardless the way results are obtained the calculations and methods used must be clearly stated.
Side 1 af 4
This assignment consists of seven questions, all expected to be answered. In the evaluation, the
following weights will be applied: 10, 15, 15, 15, 15, 15 and 15% for question 1 to 7.
In the evaluation, it is emphasized that assumptions are listed and discussed, and that conclusions
are listed and commented upon. Without comments, JMP output will not be given any points.
The total length of your answer may be a maximum of 20 standard pages. Please use standard
font size (12 pt.), line spacing and margins so that the physical number of pages do not exceed 20
pages.
The data set EXAM2017_labels.jmp (check list of variables below) must be used to solve the as-
signment. All variables are defined as continuous in the data set. It is your job, in relation to each
question, to choose the right data type.
The data set contains 868 employed persons aged 18-59 years.
Variable list:
Side 2 af 4
Question 1: (weight 10 %)
The variable J_Exercise is a categorized variable describing how many weekly hours the person is
exercising. It is expected that each of the five categories will contain the same number of persons.
Question 2: (weight 15 %)
Test whether there is a connection between alcohol consumption and the level of exercising.
Two analyses are expected. The first analysis must assume that both variables are nominal/ordi-
nal, and the second analysis must assume that both variables are continuous. Note that there are
two variables describing alcohol consumption in the data set.
Question 3: (weight 15 %)
Assume that the variable describing alcohol consumption, I_Alcohol_7, is continuous. Formulate a
relevant ANOVA model analyzing potential differences in average alcohol consumption between
gender (F_Male) and the level of exercising (J_Exercise). Potential interaction effects must also be
analysed.
Discuss whether the assumptions are met, reduce the model if necessary and interpret the results
found. Which groups are statistically different with respect to alcohol consumption?
Question 4: (weight 15 %)
a) Using simple linear regression, describe the relation between A_Income og B_YearsEduca-
tion. Estimate the model, check assumptions and interpret the results.
b) Conduct a test for expanding the model with a polynomial second-order expression.
c) What is the expected income of a person with 12, 15 and 17 years of education, respec-
tively? Describe the uncertainty of the three estimates using prediction intervals for the
expected values.
Side 3 af 4
Question 5: (weight 15 %)
a) Formulate a relevant model explaining personal income. Age, education, civil status and
job type are all expected to influence the personal income, but other factors may also be
relevant to include.
b) Should the model be formulated as a Log-Level or a Level-Level model?
Question 6: (weight 15 %)
a) Using your own words, explain what the following model describes and indicate what the
expected signs (positive or negative) should be in your view.
b) Estimate the model, do not reduce the model, assess if relevant assumptions are fulfilled
and comment on the results.
c) Expand the model with a polynomial second order expression of income, interaction be-
tween gender and education, and interaction between gender and age. If relevant, reduce
the model and comment on the results.
d) Is the final model better than the original model (6.a)?
Question 7: (weight 15 %)
Formulate a model explaining whether a person is a public employee (K_Public=1). Argue for the
selected methodology, variables included and assumptions for the methodology. As a minimum
gender, age and children living at home must be included in the analysis.
Reduce the model if necessary and assess the model: Is the model good at predicting whether a
person is a public employee or not?
What is the probability for public employment for (i) a man, 30 years old and with children living at
home, and (ii) a woman, 50 years old without children living at home?
Side 4 af 4