What did you use for ‘identity’ in Hindi… or Kannada…or Assamese?

(Asked by numerous participants at workshops)

Research aims

We had four clear aims in designing this research:

  • Surface and convey the experience of managing identities and identity artifacts, from the perspective of people in lower-income communities in India
  • Uncover pain points and under-articulated user needs around identity management including the implications of identity systems from voter IDs to Aadhaar cards and social media
  • Suggest principles to improve user experiences and the design of identity services
  • Integrate (1–3) to protect/promote individual privacy, agency and dignity

However, even equipped with these aims, we had to think through many of the on-the-ground realities of what it meant to “surface and convey experiences” or understand pain points. We share these here.

Expert interviews

In addition to our kick-off meetings where we designed our research questions, chose target demographics and other details, we started by interviewing expert informants working in the “digital identity” space. We set out the following questions:

  • What do you consider the key policy and implementation issues around identity systems in general?
  • Given your role as a [ABC] organization, what are the specific challenges you face?
  • Which actors are shaping the conversation around identity systems?
  • What do you consider the current knowledge gaps, areas where the sector needs to learn more in order to advance?
  • How does your organization approach user privacy with regards to digital data and/or identity systems?
  • Do you know of any other work that explores the user perspective, either within identity or in other sectors, that you found valuable?
  • What questions would you want to ask users?

We consulted the following experts in interviewing ranging from half an hour to an hour, either in person or over Skype:

A1: Expert Informants
Surname Forename Organization Role
Abraham Sunil CIS Founder
Adinol Shailee BanQu VP Partnerships
Agarwal Pravin BetterPlace Founder
Bansai Rajesh ex-UIDAI, now BFA Assistant Director General
Bhadra Subhashish Omidyar Network Associate, Digital Identity
Blagsveldt Sean Babajob Founder
Chandran Pinky RadioActive Founder
Desai Vyjayanti World Bank ID4D Program Manager
Dubbudu Rakesh Reddy Factly Founder
Fisher Tom Privacy International Research Officer
Hosein Gus Privacy International CEO
Madhukar CV Omidyar Network Director, Digital ID
Madon Shirin LSE Professor
McCann Neill UNDP Lead Electoral Advisor
Medhi Thies Indrani Microsoft Research Researcher
Mhatre Seb DfID Data Innovation Lead
Nagpal Himanshu BMGF Senior Program Officer
Parthasarathy Balaji IIITB Dean, Professor
Porteus David BFA Founder
Reid Kyla GSMA Head of Digital Identity
Seth Aaditeshwar GramVaani Founder
Subramanyam Kalyani Naz Foundation Program Director
Varghese Anupam Eko VP, New Products
Wensley Mark Mastercard Foundation Senior Program Manager, Financial Inclusion
Whitley Edgar LSE Professor
Yadav Anumeha Journalist

Focus groups and roundtables

We also conducted two focus groups with IIITB students to test our interview guide, as well as a radio show focus group with RadioActive in Bengaluru (including spokespeople from transgendered communities, autorickshaw drivers, sex workers, domestic helps who RadioActive works with) to surface some of the discussions around identity credentials. We held regular workshops throughout the course of the research, starting with a kick-off workshop in November 2016, an internal post-pilot workshop in February 2017 with IIITB faculty, broader roundtables in Delhi and Bengaluru in April 2017, and panels in Washington DC, at the Stockholm Internet Forum, and London.

Photo of people sitting in the Radio Active studio

A radio show focus group

User interviews


Our interest was in low-income demographics, with a sample of 10% in each site as middle-income smartphone users (on average two people per site as middle income). The occupational profile of respondents included street vendors, domestic helpers, factory workers, auto drivers, and security guards, among others, interviewed in public spaces, or sites of identity-based transactions. We did not consciously seek those without identity credentials and only encountered two respondents out of 150 who had no credentials whatsoever (the Nepalese cooks in Bengaluru mentioned in the report) through our snowball method.

Table A2: Age-Gender Distribution of sample
  10–19 20–29 30–39 40–49 50–59 60–69 Grand Total
Female 1 23 20 17 4 1 68
Male 1 35 20 10 11 3 80
Trans 2 2
Grand Total 2 58 44 27 15 4 150


Calculating income was more challenging. Although we looked to the Indian National Sample Survey we could not find income classifications easily and so based our criteria on Pew (2015)1 categorization:

Monthly income less than Rs. 4,000 (under $2 a day)
Monthly income between Rs. 4,000 and Rs 20,000 ($2–$10 a day)
Monthly income between Rs. 20,000 to Rs. 40,000 ($10–$20 a day)
Upper Middle:
Monthly income higher than Rs 40,000 (higher than $20 a day). (We are aware there is a great diversity within this range, but as they were not our key demographic, we have used this broad scale.)
Table A3: Income group distribution of sample
  Delhi Karnataka Assam Grand Total
Low 25 34 16 75
Low-Middle 4 23 6 33
Middle 12 5 10 27
Upper Middle 4 1 10 15
Grand Total 45 63 42 150

(Note: Assam numbers for middle and upper middle demographics were higher as we were asking questions around privacy, which included smartphone usage.)

However, asking income questions is always problematic—first, rather than individual salaries, we asked for family incomes, as this is more relevant for homemakers, students and others. Further, not only do most find such questions intrusive, but a monthly salary estimation may not be apt for traders, seasonal workers or daily wage earners. In the absence of direct information, we calculated salaries (and also starting from expenditure as this was easier for many to talk about) based on disparate information. For example, although one may assume that a domestic worker (such as Shailaja, a single mother) may earn well, we calculated that even if they clean three houses a month at Rs. 1,500 each (as Shailaja did) they would earn less than a pani puri salesman (such as Ajit), selling 20 pani puris a day at Rs. 20 each (both were the sole earners in their families). Therefore, Shailja was in the low income category by these calculations, while Ajit fell into low-middle.

Table A4: Occupational distribution of sample
  Entrepreneur Organized Student Unorganized Grand Total
Agriculture   4     4
Artist       6 6
Govt employee   10     10
Homemaker       14 14
Non-Profit worker   2     2
Pvt employee   2     2
Roadside hawker       18 18
Senior citizen   1   1 2
Service   22   43 65
Small Scale Entrepreneur 18 1     19
Student     8   8
Grand Total 18 42 8 82 150


We decided to focus on six sites for comparison and contrast—an urban and peri-urban (rather than purely rural) location within three states. We chose the states of Karnataka in the south, the National Capital Territory (NCT) of Delhi in the north and Assam in the east.

Table A5: Fieldwork location
  Delhi Karnataka Assam
Urban Delhi Bengaluru Guwahati
Peri-urban and rural (all pseudonyms) Kesarpur Garudahalli Bilgaon, Kodolitol, Feharbari
Table A6: Participant distribution across states
  Delhi Karnataka Assam Grand Total
Rural 22 30 20 72
Urban 23 33 22 78
Grand Total 45 63 42 150

The choice of three states was determined by multiple factors. First, geographic range was key in obtaining a diversity across India. Second, although our study was not focused solely on Aadhaar, in terms of state-wise penetration, the UIDAI statistics2 capture Aadhaar penetration as spread evenly across these states, as Delhi ranking first out of 36 in Aadhaar penetration, Karnataka at 20 out of 36 and Assam as 36 out of 36 (Assam’s 7% penetration of Aadhaar is due to a focus on the NRC rather than Aadhaar).

Finally, our choice was influenced by practicalities— with IIITB based in Bengaluru, and the core research team fluent in Kannada, the state of Karnataka was a good first site to test our interview guide and methods. We then worked beyond Bengaluru in Garudahalli, a village of 2,000 around three hours north of Bengaluru (building on previous research contacts). However, in north India, explorations beyond Delhi led us to the peri- urban site of Kesarpur, rather than the exact corollary to Garudahalli in the south. We added to this the “edge” case such as Assam, where Aadhaar has achieved less penetration, but where issues of citizenship in belonging to the National Register of Citizens are just as problematic.

Question design

Our questions initially started with building rapport with the interviewee—if a street trader, for example, buying something from them, and then introducing ourselves as researchers, with a letter of authorization and cards with the research phone number. We then broadly followed the format as below:

Three themes 1. Inventory of identity artefacts 2. Exploring transaction stories 3. Surfacing themes of privacy, agency, dignity
  Diagram showing the three question themes
Sample questions What’s in your pockets that can prove who you are? For ABC, when was the last time you used it? Was it easy or hard? Who did you use it with? How did you feel? Have you ever needed to change your information on ABC? Do you trust the organization with your information?
Goals An inventory of the analog physical and digital credentials a person possesses At least one story detainling the who, where, when of an identity transaction for whichever artefact(s) seem to o(cid:7)er richest story Nuanced insights into how the user thinks about privacy and trust, based on their experiences managing their credentials
Figure A1: Goals and types of questions

We followed an initial protocol in Karnataka and Delhi (Table A7), and refined it further for privacy questions in Assam (Table A8, following):

Table A7: Initial interview guide for Karnataka and Delhi

Setting up the interview

Practices of using identity credentials


Name, phone number (for contact), age, gender, address, last education year completed, house owner or no?, mobile phone? primary income source

Establishing incentives for enrolling

Practices of using identity credentials

Pain points

Can you tell us what it was like to sign up for ABC?

Why did you sign up for ABC? (also listen out for peer group, friends, family?)

Understanding transaction stories

For all of these, the story of the transaction:

What did you use the last time to prove who you were?

Who did you interact with? How did you feel about it?


For government: utilities, buying land, loans, traveling, etc.

For private-sector: SIM, bank account, employer, purchases

For individual: Informal work, money lender, etc.

Exploring intermediaries and trust

What was the process like in obtaining ABC? Who helped you? Or did you help someone else?

Do you get help from govt officials/bank managers, etc? What was the experience like?

Are there some people (formal and informal) who you have trusted more and had a better experience with?

What kinds of experiences have you had—positive and negative?

Deconstructing privacy

How would you feel if any of your cards were lost (or if someone used your Facebook under your name)?

Has someone finding out about identity ever caused problems? e.g., problems with privacy on Facebook? Information on Aadhaar card (name (caste?)/address)?

Who would you trust most with your personal information? government, mobile operator, Facebook?

How would you feel about providing access to your social media information in exchange for financial services?

Exploring what agency might look like

Have you ever tried to change any information on a card? What was that experience like?

What aspects of information should be on an Aadhaar card? Or all other cards? What do people know about you on Facebook?

How do you manage all your ID cards, if you have to apply for one, renew one, etc.

What do you think about the control over the information in the identity credential—are you satisfied/worried about who has access to it?

Understanding how dignity is defined and experienced

How do you feel now about the process of getting ID cards? Do you ever feel challenged?

Do you have any friends or stories of anyone who has had a difficult experience with getting or using an ID card?

With reference to social media—do you know of anyone where their information has been misused?

These questions were extremely useful in giving us an insight into user practices. However, before conducting fieldwork in Assam, we felt we needed to refine our protocols to a) focus more on privacy b) on smartphone usage (the first two were related), and on male/female dynamics (following feedback from roundtables). Therefore, our questions became more streamlined and focused:

Table A8: Refined interview guide for Assam

Noting demographics, inventory, practices

What identity credentials do you have with you?/Do you have different names/ addresses on any of them (a reminder we are not from the government!) Interactions (agency): Which ones did you use over the last week/month—and with which institution (different officials)? Which institution asked for which credential? Do you ever show one ID in place of another? Why is that? Dignity: Were you able to decide which credential to show? How did these interactions make you feel?

Understanding male/female dynamics

Who keeps hold of the ID cards in the house—is it the husband/wife? Why? Who enacts most in transactions? Why?

Understanding smartphone usage

What kinds of identity-based actions or transactions do you have to do on your phone, e.g., to access a bank account/Facebook etc etc? How do you find them? Do you know what happens to the data you enter? How do you feel about sharing this information? (To a woman—do you think you might use a smartphone differently to a man and is there anything you are more careful about? To a man— how do you feel about your wife/sister/daughter, etc., using a smartphone?)

Deconstructing privacy

When you enrolled for a SIM/Aadhaar/bank account, what information did you give? Who/which institution holds this information? Specifically, who has access to your telephone number and your biometric data? Do you think the information is shared with anyone? Are you comfortable with them sharing it (with govt depts, banks, private companies)?

How would you feel if every time you used this XXX (card, SIM, bank passbook), a record of that was sent to the XXX government so that they could track your activity over time? What if it was the police who could track your movements and interactions—with hospitals, PDS etc?

Some companies are now offering lower prices to people who provide personal information. For example, some banks will offer better loans to people who can provide a history of their mobile money transactions, or even just their phone calls and airtime top-ups. What do you think about this kind of deal? Would you be willing to share your detailed mobile phone records with a bank or other company if it meant you could get a better loan?

How would you feel if your medical records were shared with your family/friends/ colleagues? Would you share your health information with a new health service if they gave you a discount or cash promotion for all of your previous health information?

What information would you be comfortable sharing? What would you never share?

Final note

Keep sensitively asking “why do you think that/why is that?/can you share an example?”, e.g., if respondent says “I’m not educated enough to know that,” ask “why do you think that?” or “I think the government can be trusted more than the private sector”—why?

We found sharing our own identity credentials (especially to verify our research role) a light-hearted and casual starting point to enter into discussions on artifacts to both introduce ourselves and the research on identities. Simple terms for documents were easy to translate, although understood in different ways in different states: For a voter ID card, most respondents used the phrase “voter ID,” though in Delhi, another term used was “pechaan patra” (literally document which recognizes you); for a birth certificate, the English was sometimes used, or sometimes translated to “janma-pramaan patraa”; public distribution system (PDS) cards were commonly referred to as “ration cards” or by color: BPL (or below poverty line) as green, AAY (Antodaya Anna Yojana) (AAY or poorest of the poor) as red/pink, etc. Other certificates we heard (not carried on person) were the caste certificate or “Jaati-pramaan patra”; proof of address was “paani-patra” (rural), “rent agreement” (urban), while in Garudahalli in Karnataka, respondents mentioned “vasa-sthalaa drudikarana patra” (closest translation = authorization letter for status of domicile).

The artifact break down in interviews was as follows:

Table A9: ID inventory
  Delhi Karnataka Assam
Aadhaar 43 57 9
Voter ID 38 56 35
Ration card 16 35 19
Driving License 11 15 10
Passport 4 3 8
PAN Card 20 24 25

Other demographic distributions were as follows (all self-reported):

Table A10: Occupation-gender distribution of sample
  Entrepreneur Organized Student Unorganized Grand Total
Female 3 20 5 40 68
Male 15 20 3 42 80
Trans   2     2
Grand Total 18 42 8 82 150
Table A11: State-wise gender distribution of sample
  Delhi Karnataka Assam Grand Total
Female 19 22 23 64
Male 26 39 19 84
Trans   2   2
Grand Total 45 63 42 150
Table A12: Marital status distribution of sample
  Assam Delhi Karnataka Grand Total
Married 27 35 43 105
Separated     2 2
Unmarried 15 9 17 41
Widow   1 1 2
Grand Total 42 45 63 150
Table A13: Religion distribution of sample
  Delhi Karnataka Assam Grand Total
Buddhist 2     2
Christian   1 3 4
Hindu 32 55 11 98
Jain 1     1
Muslim 10 6 28 44
Sikh   1   1
Grand Total 45 63 42 150
Table A14: Resident status of sample
  Delhi Karnataka Assam Grand Total
Migrant 34 31 28 93
Non Migrant 11 32 14 57
Grand Total 45 63 42 150
Table A15: Income distribution of sample on migrant status
  Low Low-Middle Middle Middle-Upper Middle Grand Total
Migrant 48 13 30 2 93
Non Migrant 29 15 10 3 57
Grand Total 77 28 40 5 150

Storage and analysis

We translated and transcribed the interviews to English, established a process and storage protocol using a tracking sheet to record details on the interviewees (demographics), process (transcription to coding to analysis, etc.) and on the interview itself (language of interview, by whom, etc.) We utilized a two-stage process for qualitative analysis of the transcripts. In the first stage, after our very first pilot research in Bengaluru, we held numerous working sessions to discuss core emerging themes and undertook a day of affinity mapping. After this, and before we began our Assam fieldwork, we then created a more formal coding guide which served to both analyze our Karnataka and Delhi data, as well as prepare us for Assam (codes available on request). We then coded all interviews, and field notes.

Screenshot of Dedoose software
Figure A2: Coding and qualitative software

  1. Rakesh Kochhar, “A Global Middle Class Is More Promise than Reality,” Pew Research Center, Global Attitudes & Trends, July 8, 2015.

  2. “State/UT Wise Aadhaar Saturation” (Delhi: Unique Identification Authority of India, May 22, 2017).