This pseudo- panel database was constructed from successive cross-sections
data. Cohorts of individuals over time are defined by date of birth [Angus
Deaton (2000) The Analysis of Household Surveys: A Microeconometric Approach
to Development Policy].
For example: for year 1981:
Three cuts were made to break down the sample into the following classes:
The Cohorts of individuals over time are defined by date of birth:
id=1 corresponding to individuals in the 22 to 32 age group;
id=2 corresponding to the 33 to 43 age group;
id=3 corresponding to the 44 to 54 age group;
Cohorts were constructed so as to include all individuals of the same sex
(either male
or female) in a specific age group. The year is the year of the survey, and
members are the members of individuals (id) in each cohort sample in each
survey year.
For example: for year 1982:
Three cuts were made to break down the sample into the following classes:
The Cohorts of individuals over time are defined by date of birth:
id=1 corresponding to individuals in the 22 to 32 age group;
id=2 corresponding to the 33 to 43 age group;
id=3 corresponding to the 44 to 54 age group;
Cohorts were constructed so as to include all individuals of the same sex
(either male
or female) in a specific age group. The year is the year of the survey, and
the members are the members of individuals (id) in each cohort sampled in
each survey year.
I constructed Cohorts for each year following 1981 until the year 2001. I
then gathered all the Cohorts into one database, one on top of another (in
stacked form). So the database is read like this: For the same year
(repeated years), I have different individuals. In each line id=1 there is
one individual who is different from the individual in the following line
(in terms of years of education, son, wage, etc.), but they belong to the
same group.
I would like to build the database like this:
I have to follow each individual over a long time. In other words, I would
like to put my data in time series: for different years, the same
individual. The first individual of the first line of the year 1981 is the
same individual in the first line of 1982. He would be the same individual
of the first line of the year 2001. The first line (individual) of id=2 of
1981 is the same individual in the first line of id=2 in 1982. The following
years would be displayed in the same manner. This is valid for id=3.
I am familiar with the literature on panel-data. I have all the principle
books concerning the subject. My lack of knowledge regards the programming
of Stata. What I intend to do with Stata is transform repeated data into
non-repeated data.
