Don't Let The Man Tell You How To Sleep

Or: The Case For Self-Experimentation

a portrait of the author under extreme sleep deprivation

My whole life, people have told me that I need eight hours of sleep a night. I remember parents, teachers, doctors, journalists, and podcasters professing that “any less than eight hours is undersleeping, which has severe negative effects on developing brains.” But I cannot figure out why they’re all convinced of this eight hour number, and I’m quite confused about where it came from.

I have never slept eight hours a night, every night, for any consistent period of time. The amount of sleep I need varies substantially from night to night. One night it might be 10 hours, the next 0, the next 6. 

People talk like I am an outlier and that most people need eight hours of sleep a night, but I don’t believe them. Needing 8 hours of sleep may be the average experience, but it is far from the majority experience.

How much sleep do average people need?

To answer this question we need fancy empirical methods. The NIH (the [American] National Institute of Health) recommends that adults sleep between 7 and 9 hours a night. It also found that ~27.75% of US adults got less than 7 hours of sleep a night, that the amount of sleep people get has not changed that much in the past few decades (for any age group), and that sleeping less than 7 hours a night is positively correlated with various health issues (along with many other factors).

In short, the majority of Americans (72.25%) have been sleeping for an average of 7+ hours a night for the past few decades, and there are numerous studies that use “below 7-9 hours of sleep” as a proxy for “bad sleep” or “not enough sleep”.

This explains why people have been telling me that I need 8 hours of sleep. They seem to imagine that sleep works like this:

Therefore, my job is to make sure that I get “enough sleep”, which is about 8 hours. But think about it: is this how sleep actually works?

Distributions baby!

Most human attributes fall on some kind of distribution. Here’s a graph showing the distribution of human grip strength:

And here’s a (very scientific) graph showing the distribution of human height:

I believe that human sleep has any amount of variation in it, where some people need more sleep and some need less. This contradicts the idea that everyone needs eight hours of sleep a night since it would mean some people are longer sleepers, and some are shorter sleepers.

Faith S Luyster et al crunched the numbers from the National Health and Nutrition Examination Survey (NHANES), cycles (2005-2006 and 2007-2008) for a total of 10,896 respondents aged ≥ 20 and plotted the responses to the question “How much sleep do you get (hours)?” on the following graph:

As you can see, Americans sleep an average of about 6-8 hours a night. Less than 10% of us sleep above 9 hours a night, or below 5 hours a night. If you assume that most Americans are getting close to the amount of sleep they need every night then you’ll conclude that Americans need different amounts of sleep, clustered around 7 hours of sleep per night.

However, this still assumes that every American needs the same amount of sleep every night, which is certainly not the case for me. 

What is MY sleep like?

The idea that you need exactly eight hours of sleep a night doesn’t fit my life at all. But neither does the idea that everyone needs some other, fixed number of hours of sleep each night! Last night I slept for at least 10 hours, and the night before I slept for about 5. But both today and yesterday I felt good and well rested (except for a slump around 3pm to 4pm today).

Part of this variance in sleep can be explained by my days being different lengths. Today I woke up at 6:30am and plan on staying awake ‘til 11:00pm or later (16.5 or more hours). Yesterday I started my day at 9:00am and went to bed at 8:00pm (11 hours).

However this doesn’t explain all the variance. The main reason my sleep amount changes so much is that I enjoy sleeping at different times. If I didn’t vary my sleep, I’d miss out on napping through lunch, waking up for sunrise, or midnight stargazing. Such a fate is unacceptable.

In order to maintain this lifestyle I need to stay attuned to when I need to sleep and roughly how much sleep I need. Luckily my history of sleep related anxiety and insomnia mean that I’ve tried so many strategies to allow me to go to sleep. I slept bi-phasically (where I would sleep from 1am to 6am and from 4pm to 7pm) for a few months, I slept for about 4 hours a night for almost a year, and I slept for 8ish hours consistently for a few months. All of these are things that I have done — the idea of my default sleep doesn’t really exist.

What does exist is a strong correlation between being asleep for about 1 hour and feeling well balanced (capable of understanding that my feelings are temporary and feeling peaceful about that) for about 2 hours. This means that if I’m asleep for one hour, I can be awake and feel good for about two hours (capping out at around 16 hours, with a slump in the middle where I feel like taking a nap), but this isn’t quite the same thing as sleeping for 8 hours every night.

Because I don’t sleep a consistent amount every night, it is hard for me to map my experiences onto a “sleep need graph” or even simply a “sleep amount graph.” So I am, in a way, outside of the available data on sleep.

What’s the point of average?

Some results are only apparent in large datasets. For example, this study by Ojukwu, Anyanwu, and Nwafor (2017) analyzes pregnant women’s foot arch index and foot, knee, and lower back pain. This study found no significant correlation between foot arch and any type of pain. 

This doesn’t mean that you can’t have foot pain and a low foot arch, it means that, if you are pregnant, your foot arch is probably not a source of your foot knee and lower back pain.

This is the rationale behind measuring most human attributes. Rosenbush and Parker claim there is a relationship between height and heart failure though they do not confidently state a reason why this is the case. I’ve heard some people speculate that this is because tall people’s hearts need to work harder to pump all that tall person blood, but the place I heard this was from other kids in elementary school so there’s a good chance it’s false (my brother Charlie also heard this theory in elementary school).

So when it comes to sleep, it’s probably a good idea to measure it and see what’s going on, especially since we don’t know why we sleep.

Most of the data that articulates the dangers of getting less than 8 hours of sleep, at least every study I’ve read, is entirely correlational! This means it’s unclear if lower than average sleep is caused by health issues, if health issues are caused by lower than average sleep, or if some secret third thing (a hidden moderator) causes health issues and lower than average sleep. So there is little reason to focus on getting average sleep or above!

If everyone is trying to get eight hours of sleep, it will be hard to know how much sleep each individual needs (assuming it falls on a distribution). Imagine if everyone tried to be average height because they didn’t want to have pesky tall person health issues. 

But focusing on finding “where you fall on a sleep graph” doesn’t quite make sense either, since people’s sleep needs vary substantially over their lifetime, and are often dependent on things like wintertime, when your work starts, or if you have to catch a train that morning. 

I cannot answer the question “how much sleep do I get a night?” because of how much my sleep varies, but if I need to plot myself on a sleep graph I’d think these thoughts: I like to spend roughly 2 hours awake to 1 hour asleep, and when this balance is upset I feel tired, groggy, mopey, and (eventually) quite irritable. Most days I am not very tired, groggy, mopey, or irritable. Therefore I probably sleep for an average of 8 hours per 24-hour cycle, maybe a little less.

Luckily, we have a very sophisticated system designed specially for you to measure how much sleep you need. This system is called tiredness.

You are the ideal lab rat

Nobody else can tell you how much sleep you need. You can still get advice from friends and scientists, but they don’t know how tired you are.

Here’s a list of things that have impacted my sleep and that might impact yours:

1. Division of space. If my bedroom is a place where I hang out, or worse use my phone or computer, my sleep gets worse. Bedroom is for sleep, therefore being in my bedroom makes me sleepy.

2. Sleepiness comes in cycles (for me they’re about 30 minutes). So if I do something that I will get worse at as I get sleepy, I will be more aware of when I should go to bed. Painting, drawing, playing an instrument (ideally not phone things) are all good metrics.

3. Sleep is kind of like imagination all by yourself. This means that sometimes I’m trying to sleep, but can’t, because I’m imagining things all by myself. This isn’t the same thing as sleep, but it can be close enough some nights.

4. Stressing about how much sleep I am going to get if I don’t fall asleep right now makes it impossible to sleep. 

5. There’s this fun meditation technique where you start at your head and move your attention down your body and slowly relax your muscles. Then you do it again and again and again until you sleep.

6. I can’t sleep if I don’t have a blanket on me. I assume this is because I’ve always slept with a blanket.

7. Being excited about my dreams (even if I don’t remember them) makes me excited to sleep. Basically this just reframes sleep from something I have to do to something I get to do.

8. I’m really susceptible to substances and hormones, so if all else fails, melatonin works great for me.

The main thing I’ve learned about my sleep is that if I can’t sleep, I probably don’t need to sleep all that much. I don’t know why I sometimes don’t need to sleep, but I don’t care.

Your experiences are probably different! I’ve heard that having a consistent bedtime and waketime is good, but it doesn’t do that much for me. Embrace your inner scientist and start experimenting with your sleep and stop listening to people who say you ought to sleep the average amount! 

Who do you think the scientists are?

When I talk to scientists about how they do science, I get two types of answers. The first is the kind of answer that sounds goofy, ludicrous, and low effort. Examples: “we got a bunch of people in a room, fed them pasta, and we asked them which pasta they preferred.” Or “we put particles in the big machine and it makes them go really fast and bump into each other.” 

The second is something that sounds very hard and complicated, but when you stop to think about it, it’s pretty simple. This is how most scientific journal articles are written (and it’s why they suck to read). This second group of people seem to believe that science (and the scientists who do it) are fundamentally different from everything else. They often believe that science only exists in big science boxes called academies, colleges, or universities, and they're a bunch of schoolheads.

If you look to scientific academies for answers about how much you should sleep, you are assuming you are the average person, in spite of all evidence to the contrary. People who say “everyone ought to get 8 hours of sleep” are assuming that everyone is the average person, in spite of the fact that that is obviously not true.

These are logical failures based on the assumption that a given person’s individual experience is not valuable because we have a big dataset to use instead, but of course this is incorrect! Your feelings are relevant to your sleep even if they have very little effect on everyone’s sleep.

My best guess for why people fall into these logical failures is that they are naive to the idea that there is another way to be. These people can’t imagine life outside of school, and they have no interest in even attempting to imagine such a thing. People who try to live like the average person avoid remembering that there is another way to be since reckoning with their naiveté would be too embarrassing.

But you don’t have to be like them! You don’t have to be a schoolhead. You don’t have to live a boring life out of fear. You can break out of naiveté!

How do we break out of naiveté

We go back to basics and acknowledge how little we know. There are many situations, such as what amount of sleep is best for you specifically, where you can see the results clearly for yourself without any need for big datasets and science machines.

If we all experiment on our own sleep, we won’t be confused by studies and experts who say how much sleep we need. It will also be really fun, and it might even lead to important discoveries or revolutionary ideas (since scientists still don’t have a consistent theory to explain sleep).

Then we write down what we learn and compare notes. The other day my brother said to me, “I can choose when I wake up without using an alarm.” So I’ve asked him to track what time he goes to bed, what time he wants to wake up, and what time he wakes up. I explained this to a friend who said “I do that too, but it’s an anxiety thing” (we would have pursued this further, but neither of us had an interesting idea for an experiment). And I usually wake up from 5 to 6 and then choose whether or not to go back to sleep (although I ought to track this for at least a week in order to verify it).

All of these conversations are practice for shifting your perspective and snapping out of naiveté. By saying “here is how sleep is for me, how is sleep for you?” and allowing each other to answer however we please, we allow ourselves to imagine more than we ever could on our own. Attempting to conform to an idea of “normal” or “average” ends this process.

That is not to say that there is no value in figuring out averages, in fact they can be quite helpful for identifying small effects or effects that take a long time to appear (disease spread and health risks from smoking are great examples of this), but averages and large datasets are just one tool of many.

Much of popular science from the past few decades has been about figuring out what an average human is like: what is average human sleep, what is average human hunger, what is average human height, etc. And this makes sense! By figuring out what most people are like we can better accommodate the needs of most people (a necessity in big societies).

But you are not an average human. All of your experiences lie somewhere on a distribution of human experiences. Therefore you cannot simply rely on what you think the average person needs and you must experiment on yourself in order to learn what you need.

Stop living like a failed attempt at average and start trying to figure out where you fall on the distributions of human experience!

I suspect that the question “how much sleep do you get a night (hours)?” is less useful than the question “how much sleep did you get last night?” when attempting to learn how much sleep people get on average. So I’ve made a survey! Please please please take my survey, I want sleep data!!!!!!

(take my surveyyyyyyy!!!)