Mixed Methods Research in India: Recruit Interviews Around a Survey
Mixed methods research pairs a survey with interviews, boards or diaries. Settle the link before fieldwork: re-contact respondents who agreed in the questionnaire, or recruit a matched sample. This guide gives a pass-or-fail checklist and sizing maths for India.
By Jitendra Kumar Nirala, Founder · Updated 9 Nov 2026

Key takeaways
- Mixed methods research works when the qualitative participants are linked to the survey on purpose: either re-contact respondents who agreed in the survey, or recruit a fresh sample defined by the same survey groups.
- Re-contact is simplest when the questionnaire carries a consent question and a contact field. After the survey closes it depends on whether the vendor can re-invite respondents under the original consent, so get that in writing first.
- In our model, 12 interviews through re-contact need about 334 survey completes when 25% of respondents fall in the target group, 30% of them opt in, 60% of invitees qualify and 80% of booked people attend. At 15% opt-in the need rises to 667.
- Plan interviews around the number of distinct groups you must explain, and run the opt-in check at 20% of target survey completes.
In this article
- 1What is mixed methods research, and where does the qualitative part fit?
- 2How should interview participants be linked to a survey sample?
- 3Why does the linking choice get made too late?
- 4How many survey completes do you need to reach your interviews?
- 5What should the questionnaire carry so the qualitative part can be recruited?
- 6Which qualitative format fits which survey result in India?
- 7How do you combine survey and interview findings?
- 8How do we recruit and sequence the qualitative part at Qualfacto?
- 9Frequently asked questions
Your survey shows that fewer people are buying your product, but it cannot say why. So you plan interviews, bulletin boards or diaries to find the reason. Then the problem appears: the questionnaire has already gone out with no consent question and no contact field, and nobody can say who holds the respondents’ phone numbers.
That mistake is costly. You may have to recruit strangers from scratch, lose the link between each person’s interview and their own survey answers, and rebuild a timeline you never planned. Re-running the survey is rarely an option.
The fix is to choose the link before the survey goes out: re-contact respondents who agreed in the survey, or recruit a matched sample defined by the same groups.
This page is for researchers, agency planners and brand teams in India. It gives three ways to link the two, a pre-field checklist with pass and fail tests, and a sizing calculation.
What is mixed methods research, and where does the qualitative part fit?
Mixed methods research collects numbers and open-ended accounts in one study and combines them to answer one question. Designs differ mainly in timing: convergent designs collect both datasets around the same time and merge them, while sequential designs collect one dataset first and build the other on it.
In commercial work, there are four patterns:
- Explanatory sequential: the survey runs first, and interviews explain the result.
- Exploratory sequential: interviews, boards, or diaries run first, and the survey then measures what they found.
- Convergent: both run at once and are compared afterwards.
- Embedded: one method is the main evidence and the other plays a supporting part, such as a survey with a few interviews added to illustrate it.
The qualitative part is what you commission around the survey. It can be in-depth interviews (IDIs), which are one-to-one conversations. It can be an online bulletin board, where participants answer prompts in their own time. It can be a diary study, which is multi-day photo and video tasks. Or it can be opt-in digital behavioural tracking, which starts only with the participant's explicit permission.
Mixed methods research gives you both scale and explanation: the survey says how many, and the interviews say why. The cost is real. It takes more time and money than either method alone; it needs people who can run both, and the two parts can disagree. If you are still choosing between a survey and interviews, read qualitative vs quantitative research: how to choose first.
How should interview participants be linked to a survey sample?
Choose one of three links, and choose it before the survey is fielded. We use "group" for a set of respondents defined by their answers, such as women aged 25 to 34 in Pune who shop online weekly.
| Link | Use it when | Must be true before the survey goes out | Who invites |
|---|---|---|---|
| Re-contact survey respondents | You need to ask people about their own answers | The questionnaire has a re-contact consent question and a contact field, and the survey vendor agrees in writing to release them | You or your survey vendor |
| Matched fresh recruit | The vendor cannot release contacts, the survey and interviews run at the same time, or the target group is small | The recruiter has the exact group definitions: age band, city, behaviour, language | The recruiter |
| Qualitative first, survey after | The topic is new and you do not yet know the answer options | Nothing about the survey sample. The themes feed the questionnaire | The recruiter |
A matched sample means the same definitions, not the same people. You compare interviews and survey at group level, and you give up the link between one person's interview and their own survey answers.
Published practice is mixed. We read the full text of four open-access studies that pair a survey with interviews, in November 2026. We calculated interviews as a share of survey completes analysed (interviews ÷ survey completes):
| Study | Survey completes | Interviews | Interviews as % of survey | How interviewees were found |
|---|---|---|---|---|
| 226 analysed (273 recruited) | 11 | 4.9% | Respondents who gave their details in the survey for follow-up | |
| 382 | 14 | 3.7% | Purposively sampled in a second phase, for variety in reliance type, AI literacy, first-generation status and discipline | |
| 107 | 11 | 10.3% | Recruited separately through university networks, LinkedIn, Google and email, not necessarily survey respondents | |
| 121 retained from 548 | 18 | Not applicable: interviews came first | Recruited via a university mailing list and Reddit. The survey was built from the interview themes |
One study recruited from survey opt-ins, one sampled purposively in a second phase, and two recruited separately. These are four academic studies, so they show what researchers do, not a target for your ratio.
Why does the linking choice get made too late?
The choice gets made too late because the survey and the qualitative part are often bought separately, so nobody owns the link between them until the questionnaire is already signed off. By then the consent question and contact field are hard to add, and teams are left with a matched fresh recruit and a timeline they did not plan.
Two things make it worse. The first is sampling: each phase needs a sample sized for its own job, and when the two run together the samples need to be comparable, which is why the group definitions must reach the recruiter early. The second is weight. "Don't compare 12 interviews to 500 survey responses without a plan."
We would add one rule: interviews explain a survey, they do not estimate it. Before fieldwork, write one sentence per group saying what its interviews are for, such as explaining a low rating or describing how a habit works. If you cannot write the sentence, that group does not need interviews.
How many survey completes do you need to reach your interviews?
For re-contact, you need interviews wanted ÷ (show rate × qualify rate × opt-in rate × group share) survey completes. This is our model, not an official figure. The four rates are:
- Show rate: the share of booked people who attend, which is why recruiters over-recruit by booking more people than the sessions they need.
- Qualify rate: the share of invited opt-ins who pass a short check and accept a slot.
- Opt-in rate: the share of respondents in the group who agreed to follow-up.
- Group share: the share of all survey completes that sit in the target group.
Worked example, using planning assumptions that you should replace with your recruiter's and your own survey's numbers: 12 interviews, 80% show rate, 60% qualify rate, 30% opt-in, 25% group share.
- 0.80 × 0.60 × 0.30 × 0.25 = 0.036
- 12 ÷ 0.036 = 333.3, which rounds up to 334 survey completes
- Check: 334 × 0.25 × 0.30 = about 25 eligible opt-ins, and 25 × 0.60 × 0.80 = 12 interviews
At 15% opt-in the product is 0.018, and 12 ÷ 0.018 = 666.7, so 667 completes. A small group or a low opt-in rate can make re-contact impractical.
Do not apply one opt-in rate to every group. In a 2025 analysis of UK Multi-Ethnic Lifestyle Study data, Kayani et al. (PLOS One, 2025) found that 54% of 6,147 participants consented to be re-contacted about future health research, and that older and more educated participants were more likely to say yes.
A matched fresh recruit does not depend on survey size. Using the same logic as incidence rate maths, and assuming the same 80% show rate and 60% qualify rate, plus 25% incidence (the share of people screened who fit the group): 12 ÷ (0.80 × 0.60) = 25 people to book, and 25 ÷ 0.25 = 100 completed screeners. This is our model.
What should the questionnaire carry so the qualitative part can be recruited?
The questionnaire should carry six things, each a test you can run before launch. A pass means you can answer yes without guessing.

- A separate re-contact question.Pass: a respondent can answer “no” and still finish the survey with nothing lost. Fail: consent is bundled into the survey’s general terms.
- A contact field.Pass: it asks for the mobile number the person actually uses and validates the Indian format. Fail: it asks for an email only.
- Group variables as closed questions.Pass: every variable you will select interviewees on is asked before the end. Fail: you will infer it from open text later.
- City and language as closed answers.Pass: you can count completes per city and language. Fail: you cannot.
- A plain description of the follow-up.Pass: it states the format, the length, and that rewards depend on the study's terms. Fail: it promises a reward amount you have not confirmed.
- Written agreement on the contact file.Pass: you hold a written yes from the survey vendor on who receives it, for what purpose and for how long. Fail: "we will ask later".
Phone numbers are personal data. Collect them for a narrow, stated purpose, in line with the Digital Personal Data Protection Act, 2023 and the Digital Personal Data Protection Rules, 2025 (see the Ministry of Electronics and Information Technology’s data protection page). Our privacy policy sets out how we handle participant data.
Add a seventh check during fieldwork. At 20% of target completes, project the eligible opt-ins. For the example above you need at least 25. If the projection is short, switch to a matched fresh recruit while there is still time.
Which qualitative format fits which survey result in India?
Use in-depth interviews to explain a specific result, bulletin boards to put the same prompts to several groups, and diary studies or opt-in tracking to test a claimed habit against recorded behaviour. These are our recommendations.
- In-depth interviews: one-to-one sessions of 30 to 120 minutes, online or in person. Use them to explain a specific result in a specific group. See the in-depth interview method for sizing and briefing.
- Online bulletin board: use it to put the same prompts to several groups without scheduling each person. The online bulletin board page explains how one runs.
- Digital diary study: use it when the survey claims a habit over days, such as how often someone orders groceries. Participants complete photo and video tasks on a mobile app or the web, and you compare the diary with the claim.
- Opt-in digital behavioural tracking: the same comparison, using recorded behaviour. It needs explicit permission, so the design must cope with people who say no.
Online reach is high for households and young adults. In the Comprehensive Modular Survey: Telecom, fielded January to March 2025, the Ministry of Statistics and Programme Implementation (MoSPI) (press note of 29 May 2025) reported 86.3% of households with internet access within the premises and 85.5% with at least one smartphone. Among 15 to 29 year olds, 92.7% in rural areas and 95.7% in urban areas used the internet in the last three months.
Those figures are for households and young adults only. For older, rural or lower-income groups, check feasibility before you fix the sample, and consider in-person sessions.
How do you combine survey and interview findings?
Combine the two parts group by group. For each group, put the survey result, the interview themes that explain it and the decision it affects in one row. Report the survey as the measurement and the interviews as the explanation.
If the two disagree, do not pick a winner. First check that the interviewees match the group definition. Then check that the survey question and the interview prompt asked about the same thing. Only then report the gap as a finding, because a gap between what people say in a survey and what they describe in an interview can be useful. A diary study or tracking can then test which account is closer to behaviour.
Never turn interview counts into percentages. Twelve interviews cannot size a group.

How do we recruit and sequence the qualitative part at Qualfacto?
We recruit the qualitative part from your brief, while you run or source the survey separately. Here is the sequence:
- Before the survey:
send us your brief by email at support@qualfacto.io or on WhatsApp. Include the decision, the group definitions, cities, languages, number of interviews and dates. We aim to reply within one business day (Monday to Friday, 10:00 am to 6:00 pm Indian Standard Time); detailed briefs may take longer.
- Feasibility and quote:we assess whether the groups can be recruited and send a tailored proposal. We have no fixed rate card, because price depends on method, audience, sample, geography and how hard the profile is to recruit.
- Survey fieldwork:you run the survey, with the checklist above built in.
- Recruitment:for matched recruits, we recruit by city and target by household and demographic profile. We identity-check participants before inviting them. Sessions can run in languages other than English, and we can supply moderators.
- Fieldwork and deliverables:deliverables depend on scope and are agreed before the study starts. They can include recordings, transcripts, structured data exports and reports.
We offer a free pilot, so you can test a profile before committing. For how long each stage takes, see the qualitative research timeline.
Send your brief to support@qualfacto.io.
Taking part rather than commissioning? Registration is only an expression of interest, and you can see how to join and get invited.
Registration does not guarantee selection. Invitations depend on each study's requirements and your profile, and rewards depend on the specific study and its participation terms.
Get your qualitative part recruited around your survey
Send your brief to support@qualfacto.io and we will assess feasibility and send a tailored proposal.
Frequently asked questions
Sources
- 1.Results of Comprehensive Modular Survey: Telecom, 2025, Ministry of Statistics and Programme Implementation, Government of India, 29 May 2025
- 2.Lessons Learnt from a 2FA roll out within a higher education organisation, Tetlay, Treharne, Ascroft and Moschoyiannis, arXiv, 5 November 2020
- 3.Four Types of LLM Reliance and Their Predictors Among Undergraduate Writers, arXiv, 2026
- 4.A Mixed-method Study on Security and Privacy Practices in Danish Companies, arXiv, 2021
- 5.Modeling Health Video Consumption Behaviors on Social Media, arXiv, 2023
- 6.Mixed Methods Studies That Mix, Trevor Calabro, Substack, 26 July 2025
- 7.Data Protection, Ministry of Electronics and Information Technology, Government of India
- 8.Which demographic characteristics are associated with willingness to take part in recontact studies?, Kayani, Willis, Gharibzadeh, Khunti and Routen, PLOS One, 4 November 2025
- 9.Digital Personal Data Protection Rules, 2025: Operationalising consent, security, and governance obligations, JSA, JSA Prism, November 2025



