Você está na página 1de 2

HOMEWORK 1 BASIC ECONOMETRICS

Problem 1:
The data are obtained from a real study and their use is allowed just for teaching purposes!!!

Consider the Eviews work-file diabetes_study in which you have the following variables, obtained from a
sample of chronic disease patients:

COST_DIRECT_Y the direct costs related to the chronic disease, for each patient, during one year (e.g.
obtained based on estimating the medication costs, inpatient visit costs, outpatient visit costs, job loss,
change and workload reductions due to the illness, etc.)

COST_INDIRECT_Y the indirect costs related to the chronic disease, for each patient, during one year
(e.g. obtained based on estimating the costs of time spent by the family and others with caring the patient
in his outpatient and inpatient visits, job loss, change and workload reductions of the caregivers from the
family, etc.)

The dependent is the TOTAL_COST_Y= COST_DIRECT_Y+ COST_INDIRECT_Y.

SOCIODEMOGRAPHICS CHARACTERISTICS

AGE

EDUCATION DUMMIES:

ED_GRADUATE
ED_SENIOR_MIDDLE_SCHOOL
ED_JUNIOR_MIDDLE_SCHOOL
ED_PRIMARY_SCHOOL
ED_NO_EDUCATION

ETHNICITY:

ET_ROMANIAN
ET_HUNGARIAN
ET_OTHERS
ET_ROMA

GENDER

MARITAL STATUS:

MS_MARRIED
MS_COHABITATION
MS_DIVORCED_SEPARATED
MS_NEVER_MARRIED
MS_WIDOWED

MONTHLY_INCOME

NO_PERSONS_LIVING_WITH

RETIRED

CLINICAL CHARACTERISTICS
NO_YEARS_SINCE_DIAGNOSTI

OTHER_HEALTH_PROBLEMS

SELF-ASSESSMENT OF CHRONIC DISEASE SEVERITY:

SELFASM_UNDECIDED
SELFASM_NOT_SERIOUS
SELFASM_LITTLE_SERIOUS
SELFASM_SOMWHAT_SERIOUS
SELFASM_VERY_SERIOUS

DEPRESSION DIAGNOSTIC:

DIAG_MAJOR_DEPRESSION
DIAG_MILD_DEPRESSION
DIAG_NO_DEPRESSION

A. Try to setup a regression, using the strategy from general to specific, in order to show on what
factors depends the total cost with the chronic disease? Use a logY --> X specification. Can you
give some explanation for the reason of introducing certain variables in the regression?

Remark: It is known that too many variables in the regression may represent a heavy burden for
the data. Therefore, you can combine the dummies, but in a way that has some socio-economic
logic behind.
E.g.: you may want to consider the group
MS_MARRIED+MS_COHABITATION versus
MS_DIVORCED_SEPARATED+MS_NEVER_MARRIED+MS_WIDOWED.
The logic behind is to compare those patients that have a spouse with those that are single for
one reason or another.

B. Which are the variables that are significant and at what significance level? Can you find an
explanation for the sign of each of the significant variables?

C. It is usually assumed that depression influences the direct and indirect costs, increasing both of
them, because of overusing the treatment system. Please check the hypothesis of an increase in
the total cost with the chronic disease (together with increased depression), on the data you are
provided.

Remark: It is known that self-assessment of the severity of the chronic disease is usually
correlated to the depression status. You may want to consider only one of these two clinical
factors. Do a redundancy test for the exclusion and interpret.

D. Report R^2 and R^2 adjusted and interpret. What other way to measure the goodness of fit of
your regression do you know? What does it show this measure?

WRITE A REPORT OF MAXIMUM 2 PAGES AND A HALF. PRINT THE REPORT AND SUBMIT IT AT THE
COURSE OR SEMINAR, WITHIN THE DEADLINE PERIOD. YOU CAN USE FOR THE ECONOMETRIC
ANALYSIS ANY OF THE STATISTICAL PACKAGES: EVIEWS, STATA, SPSS, GAUSS, MATHLAB, SAS, R, ETC.

Você também pode gostar