Story · Culture
Who marries across caste
How many marriages in India cross caste, and how many people say such marriages should be stopped?
Key takeaways
- A peer-reviewed study of the India Human Development Survey’s 2011-12 round puts inter-caste marriages at 5.82% in 2011, by women’s own answers; another, of the National Family Health Survey, puts couples from different caste groups at 11.8% in 2005-06, 11.2% in 2015-16 and 14.0% in 2019-21. Exhibit 1
- In that NFHS reading, 15 of 36 States and Union Territories had a higher share than the all-India 14.0% in 2019-21. Exhibit 2
- By year of marriage, the NFHS share was 6.97% for the cohort married before 1971 and 13.3% for that married in 2019-21, in the study’s female column (7.99% and 13.1% in its male column). Exhibit 3
- In Pew’s 2019-20 survey, 64% of adults said it was very important to stop women in their community from marrying into another caste, and 62% said so of men; among Scheduled Caste adults, 60% and 59%. Exhibit 4
- 71% of adults with no formal education said so of women, against 50% of college graduates. Exhibit 6
How many marriages in India cross caste, as the India Human Development Survey (IHDS), run by the National Council of Applied Economic Research (NCAER) and the University of Maryland, and one peer-reviewed study of its 2011-12 round give it; how many couples in the National Family Health Survey (NFHS), run for the Ministry of Health and Family Welfare, are from different caste groups, nationally, by State and by year of marriage, as another peer-reviewed study of its 2005-06, 2015-16 and 2019-21 rounds gives it; and how many adults told the Pew Research Center’s survey of 29,999 adults (17 November 2019 to 23 March 2020) that it is very important to stop people in their community from marrying into another caste. Each figure carries its tier and its sample; none is a count from an official record.
- of adults said it was very important to stop women in their community marrying into another caste (survey, Pew Research Center, 17 November 2019 to 23 March 2020)
- 64%
- of marriages were inter-caste in 2011, in a peer-reviewed study of the India Human Development Survey
- 5.82%
- of couples in the National Family Health Survey 2019-21 were from different caste groups, a broader measure, in a peer-reviewed study
- 14.0%
Exhibit 1
A peer-reviewed study of the IHDS puts inter-caste marriages at 5.82% in 2011; another, of the NFHS, finds 14.0% of couples in 2019-21 from different caste groups, a broader measure; an IHDS brief says 95% of marriages take place within caste
What each document reading the India Human Development Survey (IHDS) or the National Family Health Survey (NFHS) says about marriage within and across caste, with who answered and the document’s tier. Figures as each document prints them.
Table of 5 rows from 3 documents: the IHDS research brief, a survey, 95% of marriages within caste; Ray, Roy Chaudhuri and Sahai, a peer-reviewed study, 5.82% of marriages inter-caste in 2011, from IHDS-II; Pradeep and Goli, a peer-reviewed study, couples from different caste groups in the NFHS: 11.8% in 2005-06, 11.2% in 2015-16, 14.0% in 2019-21.
Document Tier Who answered Marriages within caste Inter-caste marriages, by the woman’s answer Couples from different caste groups IHDS Research Brief No. 2 (NCAER and University of Maryland) Survey 35,145 ever-married women aged 15-49, IHDS-I (2004-05) and IHDS-II (2011-12) 95% – – Ray T, Roy Chaudhuri A and Sahai K, Journal of Economic Behavior and Organization, 2020 Peer-reviewed study One ever-married woman aged 15-49 in each household, IHDS-II (2011-12); the rate for 2011 – 5.82% – Pradeep and Goli, Journal of International Development, 2026 Peer-reviewed study Couples in NFHS-3 (2005-06) – – 11.8% Pradeep and Goli, Journal of International Development, 2026 Peer-reviewed study Couples in NFHS-4 (2015-16) – – 11.2% Pradeep and Goli, Journal of International Development, 2026 Peer-reviewed study Couples in NFHS-5 (2019-21) – – 14.0% Note: The columns measure different things and are not added up or compared as one series; a dash means the document gives no such figure. Ray, Roy Chaudhuri and Sahai count a marriage as inter-caste when the woman answered no to “Is your husband’s family the same caste as your natal family?”, which the IHDS asks with the word jati: “the question actually uses the word “jati” in Hindi (and its equivalents in all the other eleven languages in which the survey was administered), and not caste.” Pradeep and Goli count couples whose partners reported different caste groups, one of four (Scheduled Castes, Scheduled Tribes, Other Backward Classes and general castes): “As NFHS does not provide direct estimates of intercaste marriages, an indirect estimate of the couples who belong to different caste groups is taken as the indicator of intercaste marriage.” A marriage between two jatis in the same group is not counted, and one between groups is. The brief does not give its question or say which of its two rounds the 95% comes from, and it is a scan, read by eye.
Sources: Survey: Desai S, “India Human Development Survey Research Brief No. 2”, National Council of Applied Economic Research (NCAER) and University of Maryland, undated, pages 1 and 2. A sample, not a national count. Peer-reviewed study: Ray T, Roy Chaudhuri A and Sahai K, “Whose education matters? An analysis of inter caste marriages in India”, Journal of Economic Behavior and Organization 176: 619-633, August 2020, pages 619, 621 and 622. A sample, not a national count. Peer-reviewed study: Pradeep R and Goli S, “The Equalising Effect of Marital Exogamy: Intercaste Marriages and the Reduction of Child Nutrition Inequality in India”, Journal of International Development 38(5): 905-931, 2026, page 910 and Table A5, pages 926-927. A sample, not a national count. Analysis by The Dalit Project, as on 12 Oct 2026. Exhibit 2
In the NFHS 2019-21, as one peer-reviewed study reads it, 14.0% of couples across India were from different caste groups, and more than that in 15 of the 36 States and Union Territories
Couples whose partners reported different caste groups (Scheduled Castes, Scheduled Tribes, Other Backward Classes, general castes), %, in each round of the National Family Health Survey, by State and Union Territory, in the order and regions the study prints them. 160,646 couples, pooled over the three rounds.
Table of 36 States and Union Territories, 2019-21: Chandigarh (few couples) 8.8%; Delhi 19.8%; Haryana 14.7%; Himachal Pradesh 9.2%; Jammu and Kashmir 6.6%; Ladakh (few couples) 2.0%; Punjab 15.3%; Rajasthan 13.6%; Uttarakhand 15.7%; Chhattisgarh 12.0%; Madhya Pradesh 16.0%; Uttar Pradesh 13.7%; Bihar 18.7%; Jharkhand 19.0%; Odisha 15.5%; West Bengal 12.0%; Arunachal Pradesh 9.2%; Assam 12.0%; Manipur 19.5%; Meghalaya 2.2%; Mizoram 4.2%; Nagaland 3.1%; Sikkim (few couples) 27.6%; Tripura 9.6%; Dadra and Nagar Haveli and Daman and Diu (few couples) 20.7%; Goa 6.8%; Gujarat 18.7%; Maharashtra 18.2%; Andaman and Nicobar Islands (few couples) 13.5%; Andhra Pradesh 13.4%; Karnataka 22.3%; Kerala 14.9%; Lakshadweep (few couples) 12.7%; Puducherry 9.3%; Tamil Nadu 6.5%; Telangana 12.1%. All India 14.0%.
State or Union Territory Region 2005-06 2015-16 2019-21 Pooled samplecouples Chandigarh (few couples) North – 9.7% 8.8% 91 Delhi North 12.8% 16.5% 19.8% 1,641 Haryana North 18.5% 9.3% 14.7% 3,049 Himachal Pradesh North 12.8% 6.4% 9.2% 903 Jammu and Kashmir North 5.4% 5.0% 6.6% 1,520 Ladakh (few couples) North – – 2.0% 8 Punjab North 20.7% 6.4% 15.3% 3,138 Rajasthan North 3.0% 10.2% 13.6% 8,265 Uttarakhand North 13.3% 10.6% 15.7% 942 Chhattisgarh Central 3.6% 6.4% 12.0% 3,550 Madhya Pradesh Central 4.4% 13.2% 16.0% 9,204 Uttar Pradesh Central 11.6% 9.0% 13.7% 24,295 Bihar East 6.4% 13.0% 18.7% 9,417 Jharkhand East 12.3% 15.2% 19.0% 3,486 Odisha East 13.1% 21.2% 15.5% 4,949 West Bengal East 9.8% 10.3% 12.0% 12,710 Arunachal Pradesh Northeast 8.5% 9.0% 9.2% 127 Assam Northeast 13.4% 7.5% 12.0% 3,726 Manipur Northeast 13.0% 12.0% 19.5% 315 Meghalaya Northeast 4.1% 1.6% 2.2% 298 Mizoram Northeast 0.3% 4.9% 4.2% 100 Nagaland Northeast 6.9% 2.2% 3.1% 185 Sikkim (few couples) Northeast 15.7% 29.3% 27.6% 63 Tripura Northeast 16.0% 8.1% 9.6% 540 Dadra and Nagar Haveli and Daman and Diu (few couples) West – 16.7% 20.7% 62 Goa West 25.8% 11.9% 6.8% 331 Gujarat West 15.6% 19.0% 18.7% 9,372 Maharashtra West 17.2% 16.2% 18.2% 17,526 Andaman and Nicobar Islands (few couples) South – 35.5% 13.5% 39 Andhra Pradesh South 9.1% 11.3% 13.4% 10,954 Karnataka South 20.3% 21.9% 22.3% 9,591 Kerala South 19.7% 13.3% 14.9% 3,689 Lakshadweep (few couples) South – 11.0% 12.7% 5 Puducherry South – 4.3% 9.3% 141 Tamil Nadu South 3.0% 7.4% 6.5% 13,011 Telangana South – 9.1% 12.1% 3,403 All India 11.8% 11.2% 14.0% 160,646 Note: A broader measure than a woman’s own report of her husband’s family’s caste: a marriage between two jatis in the same caste group is not counted. The pooled sample of couples is shown for each State; it is under 100 couples for Chandigarh (91), Ladakh (8), Sikkim (63), Dadra and Nagar Haveli and Daman and Diu (62), Andaman and Nicobar Islands (39), Lakshadweep (5), marked “few couples” in the table, so their shares can swing a lot from round to round. The study’s table prints “NA” for a State with no figure in a round and does not say why; it is shown as a dash. The study’s text gives Karnataka 20.2% in 2005-06 (page 911); its table gives 20.3%, which is shown.
Source (peer-reviewed study): Pradeep R and Goli S, “The Equalising Effect of Marital Exogamy: Intercaste Marriages and the Reduction of Child Nutrition Inequality in India”, Journal of International Development 38(5): 905-931, 2026, Table A5, pages 926-927. A sample, not a national count. Analysis by The Dalit Project, as on 12 Oct 2026. Exhibit 3
By year of marriage, the share of couples from different caste groups in the NFHS was 6.97% for the cohort married before 1971 and 13.3% for that married in 2019-21, in the female column of one peer-reviewed study’s table
Couples whose partners reported different caste groups, per cent, by when they married, in the female and male columns as the study prints them. Three rounds of the National Family Health Survey (2005-06, 2015-16 and 2019-21) pooled.
Table of 18 marriage cohorts: Prior to 1971, female 6.97%, male 7.99%; 1971-73, female 9.1%, male 10.18%; 1974-76, female 9.58%, male 9.74%; 1977-79, female 10.88%, male 11.35%; 1980-82, female 10.99%, male 10.82%; 1983-85, female 10.8%, male 11.18%; 1986-88, female 11.36%, male 11.24%; 1989-91, female 12.16%, male 12.01%; 1992-94, female 12.88%, male 13.47%; 1995-97, female 12.79%, male 12.75%; 1998-00, female 13.15%, male 12.88%; 2001-03, female 13.13%, male 13.17%; 2004-06, female 13.37%, male 13.57%; 2007-09, female 14.42%, male 14.04%; 2010-12, female 13.44%, male 13.29%; 2013-15, female 13.26%, male 13.24%; 2016-18, female 15.24%, male 15.17%; 2019-21, female 13.3%, male 13.1%.
Married in Femaleper cent Maleper cent Prior to 1971 6.97 7.99 1971-73 9.1 10.18 1974-76 9.58 9.74 1977-79 10.88 11.35 1980-82 10.99 10.82 1983-85 10.8 11.18 1986-88 11.36 11.24 1989-91 12.16 12.01 1992-94 12.88 13.47 1995-97 12.79 12.75 1998-00 13.15 12.88 2001-03 13.13 13.17 2004-06 13.37 13.57 2007-09 14.42 14.04 2010-12 13.44 13.29 2013-15 13.26 13.24 2016-18 15.24 15.17 2019-21 13.3 13.1 Note: The study names the columns female and male and says of them: “from 7% for the cohort married before 1971 to 13.3% for the 2019–21 cohort”, and “A similar trend is evident in the marriage cohorts of males.” It does not say more about how the two columns differ. Per cents keep the decimals the table prints.
Source (peer-reviewed study): Pradeep R and Goli S, “The Equalising Effect of Marital Exogamy: Intercaste Marriages and the Reduction of Child Nutrition Inequality in India”, Journal of International Development 38(5): 905-931, 2026, Table 2, page 911. A sample, not a national count. Analysis by The Dalit Project, as on 12 Oct 2026. Exhibit 4
In a 2019-20 survey, 64% of adults in India said it was very important to stop women in their community from marrying into another caste, and 62% said so of men; among Scheduled Caste adults, 60% and 59%
Adults who said it is very important to stop women, and men, in their community from marrying into another caste, %, all adults and each caste category as Pew names them, in the order Pew prints them. 29,999 adults aged 18 and over, interviewed face to face in 26 states and three union territories, 17 November 2019 to 23 March 2020.
Table: General population, men 62%, women 64%; General Category, men 59%, women 61%; Scheduled Caste, men 59%, women 60%; Scheduled Tribe, men 66%, women 68%; Other/Most Backward Class, men 67%, women 69%.
Adults To stop women To stop men All adults 64% 62% General Category 61% 59% Scheduled Caste 60% 59% Scheduled Tribe 68% 66% Other/Most Backward Class 69% 67% Note: These are answers to a survey question, not a count of marriages stopped: Pew asked whether it is very important, somewhat important, not too important or not at all important to stop men and women in their community from marrying into another caste. 80% said it was at least somewhat important for women and 79% for men. Each group’s figure is of adults in that group. Pew gives a median margin of sampling error of 1.7 percentage points for the whole sample and none for each caste category.
Source (survey): Pew Research Center, “Religion in India: Tolerance and Segregation”, Pew Research Center, 29 June 2021, page 29. A sample, not a national count. Analysis by The Dalit Project, as on 12 Oct 2026. Exhibit 5
By region, the share who said it was very important to stop women marrying into another caste ran from 37% in the South to 83% in the Central region
Adults who said it is very important to stop women in their community from marrying into another caste, per cent, by region of India as Pew defines them. 29,999 adults aged 18 and over, interviewed face to face in 26 states and three union territories, 17 November 2019 to 23 March 2020.
Bars: North 72%; Central 83%; East 64%; West 67%; South 37%; Northeast 51%. Reference line: all adults 64%.
- NorthAll adults, 64%72
- Central83
- East64
- West67
- South37
- Northeast51
Note: Pew’s report says: “These regions reflect zonal council divisions.” Median margins of sampling error differ by region: North ±4.4 points (7,238 interviews), Central ±8.7 points (2,034 interviews), East ±4.9 points (6,852 interviews), West ±6.1 points (6,679 interviews), South ±4.3 points (5,570 interviews), Northeast ±7.4 points (1,626 interviews). For men the shares were North 71%, Central 82%, East 62%, West 66%, South 35%, Northeast 49%.
Source (survey): Pew Research Center, “Religion in India: Tolerance and Segregation”, Pew Research Center, 29 June 2021, pages 29 and 227. A sample, not a national count. Analysis by The Dalit Project, as on 12 Oct 2026. Exhibit 6
71% of adults with no formal education said it was very important to stop women marrying into another caste, against 50% of college graduates; 69% in rural areas and 53% in urban
Adults who said it is very important to stop women, and men, in their community from marrying into another caste, %, by education and by rural or urban area, as Pew prints them. 29,999 adults aged 18 and over, interviewed face to face in 26 states and three union territories, 17 November 2019 to 23 March 2020.
Table: No formal education, women 71%, men 69%; Primary through secondary, women 62%, men 60%; College graduate, women 50%, men 48%; Urban, women 53%, men 52%; Rural, women 69%, men 67%.
Adults To stop women To stop men No formal education 71% 69% Primary through secondary 62% 60% College graduate 50% 48% Urban 53% 52% Rural 69% 67% Note: The education rows are from the report’s overview (page 29) and the rural and urban rows from its chapter on caste (page 106); both print the same 64% and 62% for all adults. Pew gives no margin of error for these groups.
Source (survey): Pew Research Center, “Religion in India: Tolerance and Segregation”, Pew Research Center, 29 June 2021, pages 29 and 106. A sample, not a national count. Analysis by The Dalit Project, as on 12 Oct 2026.
What’s next
The next update
We will update these figures on release of a new round of the India Human Development Survey or the National Family Health Survey, or of a study of either that reports marriage across caste. This is version 1.
What the data can’t tell us yet
A count from an official record: every figure here is a survey answer or a study of a survey. The IHDS and NFHS figures measure inter-caste marriage differently, and no State figure from the IHDS was found.
What you can do
- Cite this piece.
- Found an error? Tell us how to correct it.
About this data
- What this shows
- How many marriages in India cross caste, in two readings of the India Human Development Survey and one of the National Family Health Survey (with its figures by State and by year of marriage), beside how many adults told Pew’s 2019-20 survey that it is very important to stop people in their community from marrying into another caste.
- Source
- Pew Research Center; National Council of Applied Economic Research (NCAER) and University of Maryland; Indian Statistical Institute repository record of the article (its abstract); Journal of Economic Behavior and Organization 176: 619-633; Journal of International Development 38(5): 905-931; Ideas for India (the authors’ summary of their study), 6 documents, Abstract and sections 2, 3 and 8 and Table 2 and Table A5. Analysis by The Dalit Project, as on 12 Oct 2026. Every source
- What it does not show
- A count from an official record: every figure here is a survey answer or a study of a survey. The IHDS and NFHS figures measure inter-caste marriage differently, and no State figure from the IHDS was found.
- Last updated
- 12 Oct 2026 · Next expected update: on release of a new round of the India Human Development Survey or the National Family Health Survey, or of a study of either that reports marriage across caste
- Corrections
- Changelog (version 1) · Corrections policy