By the time many Indian women emerge from their most demanding years—juggling paid work, child-rearing and household duties—they’ve racked up a deficit of years in sleep.
Work, for many Indian women, runs into two shifts: one in a formal setting, and the other at home. But the so-called ‘double shift’, it turns out, may be eating into their crucial sleep time—leaving them with fewer uninterrupted hours of rest than men through much of their adult lives.
An average Indian ‘working woman’ sleeps nearly eight and a half hours a day, which is 10 minutes short of what working men put in, Mint’s analysis of official government data shows. Worse, their sleep is more likely to span multiple stretches throughout the day. The crucial single stretch—night-time sleep—lasts less than eight hours, unlike men, who manage to clock in nearly 20 minutes more.
This comes from what is possibly India’s biggest-ever analysis of sleep patterns, by size and scope, covering over 10.2 million rows and 41 columns of data, which Mint has conducted on the basis of a large government survey held last year, called the Time Use Survey. The survey’s sample of over 450,000 Indians was spread across ages, genders, and states, and was representative of the full population.
The trends look different for homemakers. Burdened by household chores, they, too, sleep less—much less—than men in their households at night, but more than make up for it with daytime napping. This means they end up sleeping beyond nine hours in total, which is quite a luxury for working men. But that’s not to say this helps: scientific wisdom says uninterrupted night-time sleep is more valuable for a healthy lifestyle than sleep split over stretches.
How much sleep is sufficient? India does not have an official guideline, nor does the World Health Organization. The UK’s National Health Service recommends seven to nine hours for adults; so do some other US official sources. But for Indian women in prime youth, homemakers or otherwise, fewer than half meet the average figure of eight hours. Among men, more than 60% make it.
In fact, large-scale research around sleep habits is limited in India’s context, and even globally, gender and socio-cultural aspects of sleep have not been studied sufficiently. The Mint analysis you are reading aims to plug this gap.
What our analysis is all about
At a time when scientific research on sleep was still rare, 19th-century German philosopher Arthur Schopenhauer wrote: “... Sleep is to a man's whole nature what winding up is to a clock.”
Studies ever since, which mushroomed only in the 20th century, have echoed this maxim in more scientific terms: sleep deprivation can lead to health risks, ranging from obesity to diabetes, stroke and depression.
Schopenhauer himself lived up to 72 in an era of far shorter lives, and attributed his fitness to always getting sufficient sleep. But in 21st-century India, women’s “sleep tax”, elderly sleeplessness, and pre-bedtime addiction to screen time could be threatening our ability to meet that kind of bar, our analysis shows.
Academics have conducted such studies in countries with a longer tradition of time-use surveys, such as the US. But the first pan-India Time Use Survey by the statistics ministry wasn’t held before 2019; the second took place in 2024, and its data was released earlier this year. Our analysis relied on respondent-level granular data collected by the 2024 survey.
Hold your breath: the survey undertook the massive task of asking its respondents to list down everything that they did in the 24 hours preceding their interview—an activity log of sorts, from 4am the previous day to 4am on the day of the survey. Participants were required to say which activity (out of a list of 166, based on international standards, ranging from sleeping, eating and gardening to working, chatting and praying) they performed in all the 48 half-hour slots of the day.
Mint has earlier reported several other findings of the survey, once highlighting how marriage burdens women’s lives with unpaid household and caregiving work, a life-changing kind of impact that eludes men almost entirely.
The story now awaiting you will delve into the different sleep routines for various demographics—from the young to the elderly, from metro-bred kids to rural youth, and from students to homemakers. Most parts of the analysis, unless mentioned, exclude the children (ages 6 to 14) and the elderly (60 and above), who tend to sleep more and can skew the averages.
Some assumptions were made while parsing through the data, and have been explained separately. ‘Night-time sleep’ and ‘essential sleep’ have been used interchangeably, and it may actually be during the daytime for those who work in night shifts. Naps, also described as ‘incidental sleep’ by the survey, have been treated differently. All figures are averages for the demographic in focus.
Research suggests that men underplay how much they sleep: they want to be seen as more macho, less juvenile; challenging the senses to ward off sleep could be more masculine. But since the Time Use Survey asked for the full half-hourly activity log, respondents were less likely to underplay or overplay their sleep time.
Importantly, the survey was not equipped to measure the quality of sleep. If someone was awake for half an hour in the middle of the night while lying on the bed—or perhaps just sat up doing nothing—they may not have found it important enough to disclose to the surveyor, and the whole period between their bedtime and their time of waking up would be recorded as time spent sleeping.
Miles to go before I sleep
Did you know that humans are the only mammal that can delay and control sleep at will? No other mammal has this superpower. And it seems some sections of our society are using this superpower more than others.
In rural India, working-age men and women go to bed almost at the same time, around 9.50pm on average, our analysis showed (see chart 1). But women wake up a full 26 minutes earlier than men: at around 5.40am. In urban India, the day starts at 6.10am for women, and 6.38am for men, again a half-an-hour gap (women go to bed only slightly earlier).
This refers to the longest uninterrupted stretch of sleep the survey recorded. Count all stretches of essential night-time sleep, and we see that women still sleep 21 minutes less than men in rural areas on average, and 17 minutes less in urban areas (see chart 2). But overall, women sleep more than men over the course of a day, as many are homemakers and can afford daytime naps. (That’s not so for women in the workforce, as discussed earlier.)
Popular Bollywood songs at their cheesiest—of the “Neend churayi meri, kisne o sanam” ("Who stole my sleep, darling?") vibe—routinely give the sense that it’s romance that impacts sleep for the youth. But the survey suggests otherwise. The metaphorical sleep loss gets quite real and disturbing, with public health consequences, during the peak years of familial responsibilities—child-rearing, earning for the family, or attending to household chores (see chart 3).
The young and middle-aged are the most frugal sleepers at night (see charts 4 and 5). For men, the nadir of sleep time comes in their 40s, just about crossing the eight-hour mark for urban men of that age group. For women, the lowest point is in their 30s, when it lasts just around 7.6 hours a day on average.
How gender roles bend sleep
The gender gap in night sleep time is insignificant in childhood. Boys and girls aged six to 14 have similar sleep schedules. But the gap starts to swell as differentiated household duties kick in during adolescence. By the 30s, the sleep gender gap expands to half an hour. Men in all age groups, in both rural and urban areas, clock an average night sleep time of eight hours or more. For women, the age group of 20 to 50 are effectively years of lost sleep.
Things improve again in the 60s and beyond, as both men and women start getting more leisure time. In fact, the elderly sleep more than 10 hours a day, when you also include daytime naps.
Ironically but tellingly, the gender gap is not much different even whether it’s a male-led or a female-led household. Women get less night-time sleep, and make up for it with naps, even when they describe themselves as the head of the family. We also compared sleep times of men who engage only in household duties, versus similar women (male vs female ‘homemakers’). Such men sleep a lot more—and nap more, too.
Health experts say women face more sleep deficit due to the fatigue from household duties, and may also need more sleep due to hormonal differences. The only widely-cited study on this, conducted in 2016 by Jim Horne of Sleep Research Centre at the UK’s Loughborough University, said women need 20 minutes more sleep than men because of more “complex” brains, and because they tend to “multi-task”, using more of their brain than men, as per the lead author’s statement to several Western media outlets at the time. (The original study could no longer be traced online.)
Earlier this year, the American Academy of Sleep Medicine set up a task force to promote women’s sleep in the US, claiming that about 30% of US women failed to get sufficient sleep on a regular basis. “Some sleep disorders, such as chronic insomnia and restless legs syndrome, are more common in women,” the academy said. “Sleep problems in women also can be related to both cyclic alterations in the production of hormones and significant changes in hormone production that occur across the lifespan.”
The murder of sleep
Overcome by guilt after murdering a sleeping King Duncan, Shakespeare’s protagonist Macbeth hears a voice say to him, “Sleep no more!” Literary analysts famously interpret the king's murder as the murder of sleep itself: they admire the innocence of Duncan’s sleeping face and believe that Macbeth’s act had killed the innocence of sleep.
The common quibble of urban couples who are parents to teenagers suggests that the mobile phone is the 21st-century Macbeth, delaying and hurting sleep time. Data strongly backs their complaint (except that it puts them in the dock as well).
Over 55% of urban kids go to bed only after 10pm, as compared to 28% in rural areas. Among older teens, bedtime comes by 10.30pm on average, with 80% in urban areas and 55% in rural areas sleeping after the 10pm mark. This late sleep cycle gets even worse in metro cities, where bedtime for older teens averaged 10.47pm for boys and 10.34pm for girls (see chart 6). They wake up at 7.14am and 6.51am, respectively. In smaller towns and cities, the bedtime is 15-20 minutes earlier; in top cities like Delhi and Mumbai, it’s even later. In all cases, girls go to bed and wake up earlier than boys.
The survey data helped us identify the activity immediately before bedtime for each respondent, which we grouped for each major demographic group. The two most common activities that Indians end their day with are watching television or videos (which we interpret as “screen time”), and socialization or talking (see chart 7).
Around 34% of older teens in towns and cities watch TV or videos before sleeping, against 25% in the overall population. The share is higher among boys. The data doesn’t allow for a clearer classification into television or videos, but a large share is likely to be referring to the phone. Another 20% engage in conversation before sleeping, and this, too, could imply phone use.
But the teens could hold their parents to account, too, as they are hardly setting up a good example. Watching television or videos is nearly as prevalent among urban men in their fathers’ age group.
There is a consensus among health researchers that exposure to the blue light that today’s screens emit can delay sleep onset. A recent study on American teenagers, led by Dr Benjamin Zablotsky and published by the US Centers for Disease Control and Prevention, linked high screen time to later bedtimes, insufficient sleep duration, reduced sleep efficiency, insomnia, and excessive daytime sleepiness.
Screen time or kitchen time?
Meanwhile, bedtime rituals become remarkably different for men and women as they go past youth, presumably due to different levels of access to technology, and of course, the different imperatives to attend to at home.
Consider this: while around 30% of rural working-age men watch TV or videos before going to bed, the share is just 17% among women. In urban areas, this is 37% and 26%, respectively. Screen time before bedtime is more common among the more educated.
Watching TV or videos is the most common pre-bedtime activity in south India: in these states, it is true for nearly 35% of the population. In general, this activity before bedtime is more prevalent among tech professionals, older teens in metros, particularly boys, and very rich urban men. In Bengaluru, in fact, this goes up to 44%.
For women, another ritual that takes up time before they retire to the bed is cleaning up the kitchen after dinner (see chart 8). This was the last activity of the day for around 18% of rural working-age women and 16% of such urban women—and it was the third most common pre-bedtime task. Among men, this activity is very rare: less than one in 100 engage in this task the last thing in the day.
One interesting trend differentiates the poorest classes from the richest: going to bed immediately after dinner. Nearly a quarter of the population living in the poorest households in rural India reported no activity between dinner and bedtime, this share dropped to 9% among the richest. In urban India, the shares were 15% and 6%. In eastern India, this trend is unusually common, possibly because of a higher share of families falling into the bottom quintile on the national benchmarks.
Less than 0.5% of Indians read for leisure before bedtime, with little difference among those who are more educated. Many children and adolescents study, though.
Sleepless days and nights
Screen use is less prevalent among older people, with more time spent on talking instead before bedtime. In rural areas, many of the elderly also end their day with dinner.
But here’s the more worrying observation. The survey recorded an activity called “sleeplessness”, presumably time spent wanting to sleep and failing. India’s elderly face the biggest burden on this aspect among all age groups. Around one in 10 Indians aged above 70—more so in rural areas—reported at least half an hour of sleeplessness.
Among other demographics, the sleeplessness rate is less than 2-3%. In Uttarakhand, sleeplessness rates are particularly high, with close to 15% of the population reporting at least half an hour of sleeplessness, even among the working-age population (see charts 9 and 10). This exceeds 5% in Tripura and Nagaland as well. The elderly find sufficient time for daytime naps, with the average time spent on it near 1.5 hours.
Research conducted in the Indian context suggests that sleep disorders are common among the elderly, due to worsening health in general, comorbidities, and post-retirement lifestyle changes.
Bedtime on the map
Some time in the first decade of the 20th century, the Royal Society of London recommended to India’s colonial government that the Empire’s crown jewel adopt two time zones, one for the western part of the country and the other for the east. This would be in line with how the sun shines: given the vast west-to-east expanse of India, the sun rose and set more than 90 minutes later in the west than in the east.
The government, back then, rejected it in favour of a single time zone—a decision that impacts parts of the country even today. The survey’s analysis found that residents of the 13 states and Union territories that lie to the east of the 82.5°E longitude that dictates the national time zone tend to wake up around 12 minutes earlier than the western half—around 5.46am in rural parts, and 6.14am in urban parts (see chart 11). However, they go to bed around four minutes (rural) and eight minutes (urban) later, meaning they sleep less than those in the west.
The schedules are particularly on the later side in the prosperous westernmost states: Gujarat, Maharashtra and Goa (average bedtime: around 10.20pm, average wake-up time: around 6.25am).
This doesn’t directly say much about the positive or negative impacts of a single time zone, since the east and west have different prosperity levels, cultural practices and climatic and topographical conditions as well. But a 2019 study by economist Maulik Jagnani had shown that later sunset times (i.e. in western India) can be correlated with a decline in quality of sleep, especially of poorer children, which could end up reducing the quality of their education. This is because while the sun sets later, delaying bedtime, they must get up the next morning to be able to attend school, a system that has largely synchronized schedules across the country (as do government offices).
Overall, 70% of India’s rural working-age population and 64% of the urban population gets eight hours of sleep. This is the highest in some of the north-eastern states, and the lowest in Delhi, Bihar, Assam, Odisha, and Kerala (see chart 12). In each of these states, less than half of working-age women get eight hours of essential sleep; Delhi lags the most at 42% (see chart 13).
Napping is the most common in eastern states and western states (around 50 minutes a day), while north-eastern states nap the least (just about half an hour). Nap time is also high in West Bengal, Odisha, and Gujarat (see chart 14).
Most working-age Indians go to bed (84% rural, 66% urban) between 9pm and 11pm. The rest are a smaller share, but they struggle more with clocking the required hours on bed. Those who go to bed after midnight tend to sleep for an average of 7.7 hours, as compared to nine-plus hours for those who are asleep by that time. Their naps are not enough; their total sleep time during the day doesn’t cross the benchmark in many cases.
This indicates a sleep deficit for those who work in late evening or graveyard shifts: despite getting the daytime to take rest, they are unable to make the most of it. Those who work in normal daylight hours are able to sleep the most.
The night-watchwoman
British historian Roger Ekirch published a book in 2005 called At Day's Close: Night in Times Past. In a rare account of how night-time activities evolved over the centuries, he wrote that it was common to sleep in two phases at night once upon a time. He found evidence and accounts of this practice in not just Great Britain, but also in faraway places, including Asia.
When people in the medieval ages woke up in the middle of the night, they would be quite active, according to a 2012 BBC article titled “The myth of the eight-hour sleep”, which quoted Ekirch’s research. “They often got up, went to the toilet or smoked tobacco and some even visited neighbours. Most people stayed in bed, read, wrote and often prayed.”
The so-called “biphasic” sleep died down with the Industrial Revolution, as efficiency during the work day became top priority. But the data from India’s Time Use Survey allows us some glimpse into what causes disruptions to sleep at night in this country.
Only around 1.5% of the respondents in the survey provided enough details to gauge what disrupted their night-time sleep. Admittedly, this does not offer a full picture, since not everyone would have found it consequential enough to mention a 20-minute midnight snack or a late night phone call that interrupted their sleep: these might have simply been recorded as uninterrupted sleep. Nor is this a sizable enough sample. But what the sample shows is telling.
Among the 1.5% who reported sleep disruptions, we found notable gender-wise differences that again reflect differentiated household duties. For example, tending to children was a common reason why women in both rural and urban areas had their sleep disrupted for an hour or more. This wasn’t so for men: in fact, the share of such men was less than 3% (see chart 15). Preparing a meal was also a common disruptor of sleep for women. Among men, the most common reason for waking up at night was related to eating and drinking. Religious practices, talking, and personal hygiene were also common reasons.
A 2008 study, led by University of Surrey sociologist Venn Susan, called this women’s “fourth shift”, a phenomenon that is unique to them, since fathers generally were found not to tend to childcare disruptions at night. (The third shift was “emotional consciousness of the family”.) The study was based on interviews with 26 couples in the UK, and was published in the British Journal of Sociology.
Sleep disruptions in general can have health costs. Sleep quality is vital for our overall health, and people with poor sleep quality are at a higher risk for diabetes, heart disease, stroke, and mental health issues like anxiety and depression, Harvard University research shows.
A beginning
Mythology, literature, and cinema have long explored the profound relationship between sleep, love, and the human soul. Yet, in our quest to meet the eight-hour standard and power through the day with coffee, the social and cultural dimensions of this fundamental human need have been little understood. For an activity occupying a third of our lifetime, our understanding of sleep's disparities—particularly those related to gender—remains alarmingly limited.
The recent rise of the sleep industry, fuelled by growing awareness of lifestyle-related disorders, signals a positive shift. Companies now offer fat paychecks to ‘professional sleepers’ to perfect their products, and the market is booming. But as this health trend blossoms, it may be important to explore how Indians across demographics and regions experience rest in different ways. This study hopes to provide a start.
***
Methodology note: How this study was conducted
The statistics ministry held the survey between January and December 2024, with 454,192 individuals aged six and above, across 139,487 households. They were asked to list down what they did in each of the 48 half-hour slots between 4am the day before the survey and 4am on the day of the survey. In each time slot, up to three activities, performed for at least 10 minutes each, were recorded.
The full list of activities had 166 items, as per the International Classification of Activities for Time Use Statistics. Each activity was recorded in a separate row, resulting in 10.13 million rows after removing non-responders. Several adjustments were made to these records to be able to do this analysis.
1. There were slots in which a person performed multiple activities one by one. In such slots, we allotted each activity equal duration and assigned it a unique start and end time to eliminate overlaps. For example, if a person reported eating, praying, and sleeping (in that order) in the 9-9.30pm slot, we assumed ‘eating’ during 9-9.10pm, praying during 9.10-9.20pm, and sleeping during 9.20-9.30pm, implying a bedtime of 9.20pm. Another example: if an individual reported an activity A during the 8.30-9am slot as well as the 9-9.30am slot; then A and B successively in the 9.30-10am slot; and B, C, and D successively in the 10-10.30am slot, this was interpreted as activity A from 8.30am to 9.45am, B from 9.45am to 10.10am, C from 10.10am to 10.20am, and D from 10.20am to 10.30am.
2. Though recorded for a single 24-hour window, activities were assumed to represent a person’s general routine for the sake of simplicity. This means that if, in a 4am-4am activity log, one slept between 4am and 6am and again from 11pm to 4am (two separate activities in the log, the first and the last), it was assumed that their usual sleep duration is 11pm to 6am (treated as one single activity). If one slept from 4am to 6am on a Monday, and was studying from 3am to 4am on the next day (Tuesday), it was assumed that they study immediately before their bedtime, even if they may not have done so on Monday.
3. The average bedtime and wake-up time was calculated only for each person’s longest stretch of “essential/night-time sleep” in a day. If a person had two such sleep slots that were equally long, and there was no other longer slot, the one appearing later in the day was used for this purpose. To identify the stretch that came later, a modified time cycle was used instead of the 4am-4am cycle (due to the adjustments made in step 1 and 2), starting with the time that marks the beginning of their first non-sleep activity. E.g. if one’s maximum sleep stretch is five hours long—5-10am and 8pm-1am—the second one was used. But if these slots were 2-5am and 8-11pm (both three hours), the 2-5am slot was used to identify the bedtime and wake-up time.
4. The average bedtime or wake-up time for a particular demographic was calculated by plotting time as a cyclical phenomenon on a Cartesian plane. This is because person X’s bedtime of 11pm and person Y’s bedtime of 5am would have an average of 2pm on a linear scale, whereas intuitively it should be 2am given the cyclical nature of time. All respondents’ bedtimes/wake-up times were imagined as points on a circle representing the clock on a Cartesian plane, and the average location of these points was identified (appropriate weights were applied to cosines and sines of the relevant angles).
5. Sleep disruptions were defined as instances when two stretches of sleep-related activities were up to 59 minutes apart. These two stretches could have any number of activities in this one-hour period. When two stretches were 60 minutes or more apart, it was not treated as a disruption.
6. Around 12% of the sample was interviewed on a day that was not a ‘normal’ day for them, i.e. it was a weekly off or a holiday. However, the averages in the analysis do not account for this difference. Actual sleep times on a ‘normal’ day would be lower than what has been reported in this analysis, since people sleep more on rest days.
7. Activities are based on self-reporting of respondents, which can be subject to individual biases and over-reporting or under-reporting. E.g. if one wakes up for 15 minutes at midnight, they may not consider it important enough to disclose to the surveyor, or another person may report even the slightest disruptions.
8. Consumption quintiles were identified on an individual basis, not household basis, even though spending was recorded at a household level. For this, each individual's “per capita monthly expenditure” was calculated first (it was the same for each member of a given household). The exercise was done separately for rural and urban respondents.
9. All figures are based on weighted averages based on the weights reported in the raw survey data.
10. The urban parts of these districts were considered as ‘metro cities’: all districts of Delhi, Mumbai Suburban, Mumbai, Bengaluru, Chennai, Kolkata, Hyderabad, Medchal-Malkajgiri, Rangareddy, Ahmedabad and Pune. Three districts were considered from the Hyderabad metropolitan area due to the tiny sample size in Hyderabad district.