Different technical domains use this term differently, but in causal inference, spurious correlations are understood as structural relationships in the data and are distinct from chance (where random error happens to produce an apparent pattern). The correlation between bed sheet deaths and cheese consumption is a great example of a classic and common source of spurious correlations, time series data, where the confounding arises from shared dependence on calendar time. The confounding is mediated by latent (and often unmeasured) variables, but it has precisely the structure of confounding, not chance. One could hypothesize a number of latent variables. Economic conditions and age structure of the population are two possibilities, but whatever they are, they are indexed to time, and occur upstream of both variables of interest. Depending on your lens, you could invoke collider bias in the search for relationships that would sound ridiculous to the casual observer and make for good fodder on a website devoted to them.
Not me just sitting here trying to ponder if an increase in temperature could lead to more cheese consumption, heavy fat caloric intake to keep warm, and also people being in bed under their sheets more for warmth, wondering if C could indeed cause A and B separately lol...
Lots of great theories in this comment section on what could cause a correlation between increased cheese consumption and increased death by entanglement in bedsheets.
I'll throw out another one... What if people who eat cheese tend to drink wine while enjoying their cheese? If drinking wine before heading to bed puts one in a non-sober state, maybe they're more likely to end up tangled in their sheets?
My first thought regarding the cheese and death by bedsheets correlation was that cheese can contribute to sleep disturbances. These disturbances may cause individuals to have more vivid nightmares. Hence, this could explain how A (higher cheese consumption) causes B (# of people who die by becoming entangled in their bedsheets). I thought the "duh it's wrong" attitude was not helpful, as there actually could be an explanation that links the two claims together. Additionally, I think this is wrongfully leading people to give up if the explanation isn't a given (which is exactly how the LSAT wants to trick you). #feedback
About the cheese-bed entanglement correlation, it could be that wealthier populations eat more cheese; wealthier populations also live longer, and the elderly are more likely to get seriously hurt and die by becoming entangled in their bedsheets. So it could be hypothesis 3: C causes A and B.
The last example could probably be removed from this page as it isn't really helpful.
Hypothesis 1: A causes B. Cheese consumption causes death by entanglement in bedsheets. How? Um, I have no idea.
You not having any idea doesn't explain anything nor does it help me understand. It's you shrugging your shoulders and walking away. You do the same for the other hypotheses except the right answer.
The sort of lazy explanation that goes "That's wrong, duh. That's wrong, duh. That's wrong, duh. And here's the right answer, obviously. Done."
But just as a point, and maybe you can point out the problem with the explanation:
Cheese consumption causes death by entanglement in bed sheets by increasing the likelihood of brain activity associated with "tossing and turning" which, when consumed near bedtime, results in a higher risk of becoming entangled by your bed sheets. Those who consume cheese but do not sleep or lay in bed might have the same brain activity but there is no risk of dying since they're not in bed.
A good (3) explanation for C causes A & B for the firefighters and fire example:
Rich neighborhoods cause both. the increased funding/taxes leads to bigger responses to fires in general so more fire fighters show up. the wealth in the area also means people have bigger houses in general, which leads to bigger fires as there is more to burn.
Definitely request more lessons like this where the main focus is just going through additional examples. Helps a lot!!! Especially for those of us who typically learn something through pattern recognition.
If all there is a correlation, are we evaluating these to show the best possible causal relationship? Because if correlation does not equal causation, are we just trying to figure out what the most likely causal relationship is even though we cannot be certain? #help.
Subscribe to unlock everything that 7Sage has to offer.
Hold on there, stranger! You need a free account for that.
We love that you want to get going. Just create a free account below—it only takes a minute—and then you can continue!
Hold on there, stranger! You need a free account for that.
We love that you came here to read all the amazing posts from our 300,000+ members. They all have accounts too! Just create a free account below—it only takes a minute—and then you’re free to discuss anything!
Hold on there, stranger! You need a free account for that.
We love that you want to give us feedback! Just create a free account below—it only takes a minute—and then you’re free to vote on this!
Hold on there, you need to slow down.
We love that you want post in our discussion forum! Just come back in a bit to post again!
Subscribers can learn all the LSAT secrets.
Happens all the time: now that you've had a taste of the lessons, you just can't stop -- and you don't have to! Click the button.
71 comments
ice cream causes drownings because you're supposed to wait to go swimming after you eat
Different technical domains use this term differently, but in causal inference, spurious correlations are understood as structural relationships in the data and are distinct from chance (where random error happens to produce an apparent pattern). The correlation between bed sheet deaths and cheese consumption is a great example of a classic and common source of spurious correlations, time series data, where the confounding arises from shared dependence on calendar time. The confounding is mediated by latent (and often unmeasured) variables, but it has precisely the structure of confounding, not chance. One could hypothesize a number of latent variables. Economic conditions and age structure of the population are two possibilities, but whatever they are, they are indexed to time, and occur upstream of both variables of interest. Depending on your lens, you could invoke collider bias in the search for relationships that would sound ridiculous to the casual observer and make for good fodder on a website devoted to them.
Not me just sitting here trying to ponder if an increase in temperature could lead to more cheese consumption, heavy fat caloric intake to keep warm, and also people being in bed under their sheets more for warmth, wondering if C could indeed cause A and B separately lol...
Can we all appreciate how funny and well-understanding the examples are?
The last example about bedsheets and cheese consumption was just too good to forget
Lots of great theories in this comment section on what could cause a correlation between increased cheese consumption and increased death by entanglement in bedsheets.
I'll throw out another one... What if people who eat cheese tend to drink wine while enjoying their cheese? If drinking wine before heading to bed puts one in a non-sober state, maybe they're more likely to end up tangled in their sheets?
I geeked out over this lesson. Everything clicked!!!
what if the fire station was on fire #law
The third scenario is the source or signal for the fire in the first place. C the sound of the alarm -> Fire Fighters -> size of fire
The third hypothesis in the first example is perception. You perceive there to be a lot of firefighters and you also perceive the fire to be very big.
Maybe i've read too much descartes
You might not mourn by eating ice cream, but I sure do.
My first thought regarding the cheese and death by bedsheets correlation was that cheese can contribute to sleep disturbances. These disturbances may cause individuals to have more vivid nightmares. Hence, this could explain how A (higher cheese consumption) causes B (# of people who die by becoming entangled in their bedsheets). I thought the "duh it's wrong" attitude was not helpful, as there actually could be an explanation that links the two claims together. Additionally, I think this is wrongfully leading people to give up if the explanation isn't a given (which is exactly how the LSAT wants to trick you). #feedback
C, the firefighter's were given an anonymous tip from a pyromaniac who started the large fire as the firefighters arrived.
About the cheese-bed entanglement correlation, it could be that wealthier populations eat more cheese; wealthier populations also live longer, and the elderly are more likely to get seriously hurt and die by becoming entangled in their bedsheets. So it could be hypothesis 3: C causes A and B.
The last example could probably be removed from this page as it isn't really helpful.
You not having any idea doesn't explain anything nor does it help me understand. It's you shrugging your shoulders and walking away. You do the same for the other hypotheses except the right answer.
The sort of lazy explanation that goes "That's wrong, duh. That's wrong, duh. That's wrong, duh. And here's the right answer, obviously. Done."
But just as a point, and maybe you can point out the problem with the explanation:
Cheese consumption causes death by entanglement in bed sheets by increasing the likelihood of brain activity associated with "tossing and turning" which, when consumed near bedtime, results in a higher risk of becoming entangled by your bed sheets. Those who consume cheese but do not sleep or lay in bed might have the same brain activity but there is no risk of dying since they're not in bed.
A good (3) explanation for C causes A & B for the firefighters and fire example:
Rich neighborhoods cause both. the increased funding/taxes leads to bigger responses to fires in general so more fire fighters show up. the wealth in the area also means people have bigger houses in general, which leads to bigger fires as there is more to burn.
people in Wisconsin def get real kinky with the cheese
Definitely request more lessons like this where the main focus is just going through additional examples. Helps a lot!!! Especially for those of us who typically learn something through pattern recognition.
This is such a great lesson. Thank you
This lesson HELPED a ton!!!!
#feedback it's a crime that you would put so many banger paragraphs in here and NOT have them voiced over with a video.
If all there is a correlation, are we evaluating these to show the best possible causal relationship? Because if correlation does not equal causation, are we just trying to figure out what the most likely causal relationship is even though we cannot be certain? #help.
This makes a lot of sense thank you
People in Maine need to stop consuming margarine.
Are the writers ok? These examples are starting to get a little weird lol. I mean the ice cream one got dark real fast.
A causes B
B causes A
Some secret third thing causes A and B
I lied, there's nothing