Methodology

The The Kerala Gen Z Report 2026 is a single-wave, voluntary-response survey conducted in English and Malayalam. The analysis below is intentionally conservative: we describe the sample as it is, not as we wish it were.

Fieldwork, coding and analysis were carried out by the Community Leadership and Social Change Program Research Scholars, Department of Graduate Institute, DHI University (DHIU), India, and published open access by Moyan AI.

Instrument

The questionnaire contained 38 close-ended questions across five thematic blocks: digital behaviour and wellbeing; education, employability and AI use; migration intent; civic engagement; and entrepreneurship, pluralism and media trust. Questions were bilingual, with Malayalam translations shown in brackets after each English item. No personally identifying information was collected.

Sample

N = 399 valid responses. The sample is young (83.2% aged 20–24, 16.3% aged 16–19, 9.3% aged 25–29), heavily male (72.7%), and overwhelmingly Muslim (97.0%). Almost half (48.6%) live in Malappuram district, and 80.7% report a rural Grama Panchayat residence.

Educational distribution: 38.1% in higher secondary, 44.1% in general arts and science degrees, 10.3% in professional/technical degrees, 5.5% in post-graduation or research, and 2.0% SSLC.

Fieldwork

Responses were collected between 21 May – 28 June 2026 through an online form. The link was distributed via WhatsApp groups and personal networks, which explains the geographic concentration. Because participation is voluntary and online, the sample is not random and should not be treated as representative of all Kerala youth.

Key limitations every reader should note

• Religious concentration. With 97% Muslim respondents, the pluralism and media-trust findings should not be read as the view of a religiously mixed cohort. They are the view of a single religious group.
• Gender skew. Males make up nearly three quarters of the sample. Findings on entrepreneurship, politics and migration may look different with gender balance.
• Malappuram weighting. Half the sample is from one district. State-level generalisation requires caution.
• Self-report. Screen time, anxiety, intent to migrate and future behaviour are all reported by the respondents themselves, not measured objectively.
• Volunteer bias. Online, network-driven distribution tends to attract people with stronger opinions and more internet access than the average young person.

How percentages are computed

Each percentage is the count for a response option divided by the total number of valid responses for that question. Because some questions were not asked of all respondents or had missing answers, denominators may vary slightly; the exact n for each figure is shown on its statistic page and in the open data.

Open data release

Aggregate data is released as CSV and JSON in the open data page. Row-level microdata is not released because the sample is small and highly concentrated; releasing individual rows would make identification easier. Custom cross-tabs can be requested at communityjournaldhiu@gmail.com.