The DHS Program User Forum
Discussions regarding The DHS Program data and results
Home » Countries » India » Caste correlation table
Caste correlation table [message #4061] Wed, 25 March 2015 01:54 Go to next message
krishnab75 is currently offline  krishnab75
Messages: 5
Registered: October 2013
Location: Los Angeles
Member
Hello.

So I am comparing the caste information between the IA2, IA3, and IA5 India DHS surveys. So I was trying to correlate the caste data from the IA2,IA3 surveys with the caste data in IA5. I understand that the caste data in IA5 is in the S117 and S118 variables and that there is a separate castedata.zip file. However, the castedata file in IA5 identifies the specific caste names like "Boya" or "Lambadi," while the caste variables v131 in IA2,IA3 just indicate "schedule caste" and "other backward caste," etc. So I was wondering if there is a table that correlates the caste names like Boya to the appropriate designation of schedule caste or backward caste, etc.
Thanks
Re: Caste correlation table [message #4084 is a reply to message #4061] Fri, 27 March 2015 09:19 Go to previous messageGo to next message
Bridgette-DHS is currently offline  Bridgette-DHS
Messages: 3017
Registered: February 2013
Senior Member
Following is a response from Senior DHS Specialist, Fred Arnold:

"In the 1992-93 survey (NFHS-1), there are three questions related to caste/tribe of the head of the household in the household questionnaire, with the same questions about the head of the household repeated in the woman's questionnaire. The first question asks whether the head of the household belongs to a scheduled tribe (ST). If yes, the second question asks for the name of the tribe. If the head of the household does not belong to a scheduled tribe, the third question asks for the caste name of the head of the household. There are no questions about scheduled castes (ST) or other backward classes (OBC).

The 1998-99 survey (NFHS-2) and the 2005-06 survey (NFHS-3) include the same two questions on caste/tribe in the household questionnaire. The first question asks for the name of the caste or tribe of the head of the household, and the second questions asks whether the caste or tribe is an SC, ST, OBC, or none of these. NFHS-2 does not have any caste/tribe questions in the woman's questionnaire. In addition, NFHS-3 asks the same two questions that are in the household questionnaire in the woman's questionnaire and the man's questionnaire for the respondent for those questionnaires, not for the household head.

In NFHS-2 and NFHS-3, the respondent self-reported their caste/tribe status as SC, ST, OBC, or none of these, and no attempt was made to verify which category the caste/tribe they mentioned was actually in. It is not possible to match up the caste names in NFHS-1 to the SC, OBC, or none of these categories since the schedules of castes and tribes are state-level designations not national designations. In other words, a particular caste than is an SC in one state may not be an SC in another state. There are further complications when a person moves between states that have different caste schedules. Some states even break down their lists of scheduled castes and scheduled tribes by region within the state, so a particular caste/tribe designation may be different in different parts of the same state."
Re: Caste correlation table [message #4170 is a reply to message #4084] Sat, 11 April 2015 21:45 Go to previous messageGo to next message
krishnab75 is currently offline  krishnab75
Messages: 5
Registered: October 2013
Location: Los Angeles
Member
Hey Bridgette, this is very helpful. I just had one thing to follow up on. So I was looking through the caste variables like in the India NFHS-1 survey table IAHR23FL. So the caste variable is SH035 which is the answer to the question "Which caste head HH belong?" So the answer seems to be an integer value that probably corresponds to a set of actual caste names. Is there a table that will match the index numbers to the actual caste name--like a string/text value?

I am good on NFHS-2 and NFHS-3 withe the caste names. In the case of NFHS-2 the IAHR42FL table does provide actual caste names in variable SH40CT, so that is good. In the NFHS-3 survey the caste names are provided in the CASTETRIBE table variable CTRESP provided along with IAHR52FL.

So I am just missing caste names for NFHS-1. On the other hand, let me know if the indexes for the caste names are the same in NFHS-1 as they are in NFHS-2, NFHS-3?

Thanks.
Krishna

[Updated on: Sat, 11 April 2015 21:57]

Report message to a moderator

Re: Caste correlation table [message #4200 is a reply to message #4170] Wed, 15 April 2015 20:37 Go to previous messageGo to next message
Liz-DHS
Messages: 1516
Registered: February 2013
Senior Member
Dear User,
I emailed the address you provided on the forum. I had an attachment, but it was too large for the forum. It went fine by email.
Thank you for your post.
Re: Caste correlation table [message #4406 is a reply to message #4200] Thu, 21 May 2015 13:34 Go to previous messageGo to next message
dhruschk is currently offline  dhruschk
Messages: 5
Registered: May 2015
Member
Hello,
I would also be interested in receiving a copy of the caste names file for NFHS-1. I have the pdf file for the E. Kitts book that comes with the downloadable zip folder, but it ends at value 2889 and the values for the variable S148 in the dataset also include values from 2890 to 7370. Thank you!
Re: Caste correlation table [message #4416 is a reply to message #4406] Fri, 22 May 2015 16:37 Go to previous messageGo to next message
Bridgette-DHS is currently offline  Bridgette-DHS
Messages: 3017
Registered: February 2013
Senior Member
Please check the following post on Caste Codes.
Re: Caste correlation table [message #4423 is a reply to message #4416] Sun, 24 May 2015 09:03 Go to previous messageGo to next message
dhruschk is currently offline  dhruschk
Messages: 5
Registered: May 2015
Member
Thank you for quick reply! The pdf that comes with the IAIR23 files only gives codes for values up to 2889. However, there are values in the dataset up to about 7000. Also, for some reason, the pdf that is available for downloading in the referenced post appears to be corrupted. When I download it, it always says the file has been damaged. Thank you for your time. Dan
Re: Caste correlation table [message #4428 is a reply to message #4423] Tue, 26 May 2015 06:44 Go to previous messageGo to next message
Bridgette-DHS is currently offline  Bridgette-DHS
Messages: 3017
Registered: February 2013
Senior Member
Please see the attached.
Re: Caste correlation table [message #4432 is a reply to message #4428] Tue, 26 May 2015 09:09 Go to previous messageGo to next message
dhruschk is currently offline  dhruschk
Messages: 5
Registered: May 2015
Member
Thank you for sending these. I think these are the same tables that appear in Eustace Kitts 1885 and are available in the IAIR23 zip folders. They are missing the labels for values 2890 to 8000 (about 2500 cases). Thanks again for delving into this issue. I completely understand if at this point those labels don't exist, but I thought I would check. Best, Dan
Re: Caste correlation table [message #5469 is a reply to message #4432] Thu, 28 May 2015 06:31 Go to previous messageGo to next message
Bridgette-DHS is currently offline  Bridgette-DHS
Messages: 3017
Registered: February 2013
Senior Member
Unfortunately, this is all we have available.
Re: Caste correlation table [message #5470 is a reply to message #5469] Thu, 28 May 2015 08:31 Go to previous messageGo to next message
dhruschk is currently offline  dhruschk
Messages: 5
Registered: May 2015
Member
Thank you so much for your efforts to find this. All the best
Re: Caste correlation table [message #5485 is a reply to message #5470] Fri, 29 May 2015 08:37 Go to previous messageGo to next message
Bridgette-DHS is currently offline  Bridgette-DHS
Messages: 3017
Registered: February 2013
Senior Member
You are welcome.
Re: Caste correlation table [message #5491 is a reply to message #4061] Fri, 29 May 2015 15:22 Go to previous messageGo to next message
krishnab75 is currently offline  krishnab75
Messages: 5
Registered: October 2013
Location: Los Angeles
Member
Hey Brigette, I am not completely comfortable with the absence of supporting documentation on the caste codes. Essentially we have no way to confirm the validity of the underlying data--especially if codes between 2000-7000 are missing from the documentation. I mean this absence also casts doubt on whether the available document with 0-2000 codes is actually the correct document or the proper set of codes used to code the data.

I imagine it is not an easy situation to track down this info, but as scholars we are getting a lot more pressure to prove the validity of the data that we use. I am sure you are aware of the recent data falsification scandals in political science at UCLA. So I just want to make sure that if someone asks about the data, I have appropriate documentation and auditing to point back to. Also note that without the appropriate caste codes, that caste variable is not really useable in the survey. Is it possible to contact the original coders of the data to see if they might know the codes?

Re: Caste correlation table [message #5563 is a reply to message #5491] Mon, 08 June 2015 06:13 Go to previous message
Bridgette-DHS is currently offline  Bridgette-DHS
Messages: 3017
Registered: February 2013
Senior Member
Here is a response from Senior DHS Technical Specialis, Fred Arnold:

The caste names were not coded centrally. They were coded by numerous field agencies that conducted the NFHS-1 fieldwork for each of the 25 states. The field agencies were provided with a standard set of codes to use, which we have sent to you. This list includes all caste names for codes 0001 through 2899 that were used by the field agencies. We do not have caste names for the 104 caste codes above 2899. Those 104 codes constitute only 2.9% of all the caste codes. It is likely that the 104 caste codes were created by the field agencies for specific states since some caste names given could not be matched to the master list of codes from 0001-2899 from the 1931 Census. Given that the coding was conducted for 25 states separately more than 20 years ago, it is not possible to reconstruct the caste names for the 104 codes. It is not correct to say that the lack of caste names for 2.9 percent of cases renders the caste variable unusable. You can simply combine the 104 caste codes from 2900-6000 into one caste code of OTHER CASTE and use the variable in any analysis.

Previous Topic: Village identifiers in NFHS3
Next Topic: 2014-2015 DHS
Goto Forum:
  


Current Time: Thu Mar 28 20:00:36 Coordinated Universal Time 2024