April 2003
Department of Biostatistics
Postgraduate course in biostatistics
April 2003
Obligatory exercise
In a study of quality of life, data were collected from 652 persons. The data contain information on sex, self-assessed health, age, and self-assessment of emotional problems. Moreover, two blood samples were collected with an interval of one week from each of 40 men among the 652 persons. The blood samples were independently assayed for the concentration of compound X.
The data, which can be found on http://www.biostat.au.dk/teaching/postgraduat.htm in two formats (data.dta in STATA format and data.sav in SPSS format), contain the following variables:
Sex (0: male, 1: female)
Health (1: very good, 2: good, 3: acceptable, 4: bad)
Age (in years)
Agegroup (1: < 50 years, 2: ³ 50 years)
Problems (0: few, 1: many)
Conc_1 (concentration of compound X in first blood sample)
Conc_2 (concentration of compound X in second blood sample)
Question 1.
a) Describe the risk of having many emotional problems for people aged 50 years and above, and for people less than 50 years. 95% confidence limits should be provided.
b) Compare the two age groups with respect to the risk of having many emotional problems. Is age a risk factor for having many problems?
c) Does sex have an effect on the association between age group and the risk of having many problems? What is the relative risk adjusted for sex?
Question 2.
a) Describe the age distribution for each of the four groups defined by self-assessed health.
b) Describe the association between age and self-assessed health.
Question 3.
a) Describe the agreement between the two measurements of concentration of compound X.
b) Describe the association between the concentration of compound X (first blood sample) and age.
c) Calculate a 95% prediction interval for the concentration of compound X in a male patient 50 years old.
All your answers should be motivated, and 95% confidence limits or standard errors should be provided where appropriate.
A personal solution of this exercise should be sent to
Department of Biostatistics
Vennelyst Boulevard 6
8000 Århus C
no later than 12.00 AM on Monday, May 12, 2003.
1