Three stories dominate India’s data landscape as the 2026 interviews approach: Census 2027 is underway with its houselisting phase running right now and caste enumeration coming, the Digital Personal Data Protection Rules were notified in November 2025, and artificial intelligence has entered both the questions and the machinery of statistics. Real boards have asked about census delay, artificial intelligence against machine learning, exit polls, and the limits of statistics itself.
Boards read the same headlines you do, and they enjoy asking a statistician to judge the numbers in them. Day 13 arms you on the three stories most likely to walk into your interview, every fact verified against official releases.
Census 2027: the biggest statistical event in the room
The verified picture as of August 2026. The Census was notified in the gazette on 16 June 2025 and approved by the Union Cabinet in December 2025. It runs in two phases: the houselisting and housing census from April to September 2026, which is in the field as you read this, and population enumeration in February 2027 with a reference date of 1 March 2027, held earlier, in September 2026, for Ladakh and the snow bound areas of Jammu and Kashmir, Himachal Pradesh and Uttarakhand, with a 1 October 2026 reference date.
Three facts elevate an answer from adequate to sharp. This is India’s first digital census, collected on mobile applications with a self enumeration option for households, run by around thirty lakh field functionaries. It is the sixteenth census in the series and the eighth since independence, arriving after the 2021 round could not be held due to the pandemic. And its second phase will enumerate caste for the whole population, a decision of April 2025, making it the first full caste count since 1931.
If the older question comes, the one a real board asked about problems from the census delay, answer it as a statistician: a fifteen year gap means sampling frames, weights and per capita denominators across the system aged, and every survey built on census foundations carried that uncertainty. Then close with the update: the repair is now in the field. Problem, consequence, machinery. That shape again.
DPDP: the privacy law a data officer must know
The verified status. The Digital Personal Data Protection Act was enacted in August 2023, and its operating rules, the DPDP Rules 2025, were notified in November 2025, with implementation phased over the following months and a Data Protection Board established. India now has a functioning personal data protection regime coming into force in stages.
Why does this belong in an ISS interview? Because the board may ask where the line sits between data privacy and data for policy, and the strong answer now has a legal anchor. The frame: policy needs data about people, people need protection from misuse, and the working resolution in statistics has long been aggregation and confidentiality, individual records protected while collective truth is published. The DPDP framework strengthens the citizen’s side of that bargain, and official statistics has practised its own version of it for decades. An answer that connects the new law to the old statistical ethic of confidentiality is an answer only a statistician can give.
AI: the question that is really about you
A 2024 board asked the difference between artificial intelligence and machine learning, and whether AI threatens jobs, including the candidate’s own future. The clean distinction: machine learning is the family of methods where systems learn patterns from data, and artificial intelligence is the broader ambition of machines performing tasks that need intelligence, with machine learning as its main engine today.
On the threat question, the confident statistician’s answer runs: AI shifts the work rather than deleting it, because models are built on data, and the quality, representativeness and honesty of data is precisely the statistician’s craft. The official system itself is modernising in this direction, with digital collection now inside the new CPI and the Census. A candidate who welcomes the tools while insisting on the discipline sounds like an officer; a candidate who fears the tools sounds like a bystander.
The limits question
One more verified board question ties the day together: what are the limitations of statistics, and how do we reduce them? The honest inventory: statistics speaks in aggregates and can miss the individual, it depends on the quality of collection, and it can be misread or misused. The reductions are the profession itself: sound design, transparency of method, and the humility to publish margins of error. Ending a limitations answer with the remedies is the difference between criticising the tool and mastering it.
Tomorrow, Day 14: Week 2 closes with rapid fire. Twenty five technical questions from real ISS interviews, in one post.
For the free ISS Interview Starter Kit and the DAF Question Mapper, send the word INTERVIEW on WhatsApp to +91 842591 5355.
Frequently Asked Questions
What is the schedule of Census 2027?
Houselisting runs April to September 2026, and population enumeration takes place in February 2027 with a reference date of 1 March 2027. Snow bound regions enumerate in September 2026 with a 1 October 2026 reference date.
Will Census 2027 include caste enumeration?
Yes. Caste will be enumerated in the population enumeration phase, the first full caste count since 1931, under a decision taken in April 2025.
What is the status of the DPDP Act?
The DPDP Act 2023 became operational with the notification of the DPDP Rules in November 2025, implemented in phases, with a Data Protection Board established.
What is the difference between AI and machine learning?
Machine learning is the set of methods by which systems learn patterns from data. Artificial intelligence is the broader goal of machines performing intelligent tasks, with machine learning as its principal engine today.
Sources: PIB and Registrar General releases on Census 2027, 2025 and 2026; PIB and MeitY releases on the DPDP Rules, November 2025.
Practice. Improve. Repeat.