This study includes three sets of questions that can be used to measure the extent to which people prefer to keep their lives religiously segregated from others: the religious composition of respondents’ friend circle; how important, if at all, they feel it is to stop interreligious marriage of men and women; and whether they would be willing to have people of other religions as neighbors.

These three sets of questions are highly correlated with each other (Cronbach’s alpha is higher than 0.7 for all six religious groups) and can reliably be combined into an index of religious segregation.³¹

The study displays the results of the index ranging from 0 (no preference for religious segregation) to 6 (highest preference for religious segregation). The higher the score on the index, the higher the respondent’s preference for religious segregation.

The individual components of the scale are as follows:

1. Composition of the respondents’ friend circle: How many of your close friends are (respondent’s religion)? (e.g., If respondent is Hindu: are Hindu? If respondent is Muslim: are Muslim?, etc.). All of them, Most of them, Some of them, Hardly any of them, None of them.
2. Views on stopping interreligious marriage:How important is it to stop women in your community from marrying into another religion? Very important, Somewhat important, Not too important, Not at all important.
   1. How important is it to stop men in your community from marrying into another religion? Very important, Somewhat important, Not too important, Not at all important.

(The two separate questions on interreligious marriage of men and women are highly correlated with each other and can be reliably combined into one variable.)

3. Willingness to accept people of other religions as neighbors: Would you be willing to accept (a Hindu/Muslim/Christian/Sikh/Buddhist/Jain) as a neighbor? Yes, No. (Respondents were not asked if they would be willing to accept members of their own religious group as neighbors, but they were asked about all other groups.)

Respondents could also volunteer a range of nonresponses such as they “don’t know” or that “it depends on the situation,” or refuse to answer the question. See questionnaire for full question wording and response options.

Each set of questions was coded as follows:

1. Composition of the respondents’ friend circle:
   1. All friends belong to the same religion (2)
   2. Most friends belong to the same religion (1)
   3. Some/Hardly any/No friends belong to the same religion/Don’t know/Refused (0)

2. Views on stopping interreligious marriage:
   1. Of women
      1. Very important to stop (2)
      2. Somewhat important (1)

• Not too important/Not at all important/Depends on the situation/Don’t know/ Refused (0)

1. Of men
   1. Very important to stop (2)
   2. Somewhat important (1)

• Not too important/Not at all important/Depends on the situation/Don’t know/ Refused (0)

The separate scores were then added to a total score ranging from 0 to 4, and recoded as follows: (0=0) (1,2=1) (3,4=2).

3. Willingness to accept people of other religions as neighbors:
   1. For each religious group asked about:
      1. No, not willing to accept as a neighbor (1)

Yes/Other/Both/Neither/Depends/Don’t know/Refused (0).³²

5. Each person was asked about five religious groups (not including their own) resulting in a total score ranging from 0 to 5. The total score was then recoded as follows: (0=0) (1,2=1) (3,4,5=2).