FAQ on ARIS REDS


1.  Unlike the 69-71 "merged" data, the redsecon.raw file in 1982 is in 
a format called "flat".   I want to convert these into stata format.  So 
is it correct to first divide these data into "decks" (it seems that 
each new line in redscon.raw refers to a different "deck"), and then 
write the appropriate dictionary using  the coding in redsecon.pdf?  I 
did this, but there seems to be information in redsecon.raw for decks 
that are not in redsecon.pdf.  For eg.  redsecon.pdf  lists decks 1-30, 
but using the method above, redsecon.raw seems to have (in addition to a 
bunch of typos) information on decks 42-45 as well.

You understand correctly what a flat file is. This is all the documentation we have. The codes for 42-45 are in redseadd--I have fixed the readme. Yes there are typos. At least this way we know the data haven't been doctored.

2.  Is there an economic questionnaire for 1982?  I only see the 
codebook for 1982 in redsecon.pdf.

Not scanned and I actually don't have a copy. We can probably get one. 

3.  For the 69-71 "merged" data, arismq.pdf has the detailed 
description/information for each variable.  But this stops after the 
total expenditure variable (column 1271).  However, arismcf.pdf, which 
has the dictionary for the merged file lists additional variables (upto 
column 1414).  I converted 69-71 to stata using the dictionary in 
arismcf.pdf, but it would help to have the detailed description on 
additional variables which seem to be missing from arismq.pdf.  Are 
these somewhere else?

Yes, wouldn't it be nice. This is all we have.

4.  Is there a file that gives a general description of the data over 
years?  That is, how many of the original 69-71 households were 
re-sampled/tracked in 1982, and how many were interviewed/tracked in 
1999?  I seem to have matched 3636 1969-71 households to the 1999 data, 
but have no clue whether this is the correct number.

Not really. You can of course check our papers and see how well we do. 
The paper by Vashishtha that we reference in some of those has a bit more information. I will try to get that scanned and added. 
Vashishtha, Prem S. Changes in structure of investment of rural
households: 1970-71-1981-2 Journal of Income and Wealth, July 1989 10(2)
21-45. 


1.  Just to double-check, the original sample of HHs and villages that
were first interviewed in 1969 are again interviewed in 1971, 1982, and
1999.  So these are the same HHs and individuals (net of migration etc.)
at 4 points in time.

They were also interviewed in 1970

2.  In the public69-71 directory, there are no 69 files.  Could these
have been mislabeled as 73,74 and arisdk96 raw files, which are also
present in public69-71?

those numbers 73,74, 96 are deck numbers from the aris 70-71 survey.
There is no separate  file 69 or 70  survey data. But the arism file has
all three years for the "merged households".

3.  In the secure 99 files, there is information on the links to 1971
and 1982 data.  Is there a link to the 1969 data as well, or are these
data not included at all?

If you can link to 70-71 you can also link to the arism file which
has the 69 and 70 data