let's say I want to make a statistical analysis of how many customers there are in a supermarket, but I only count them when I go to the supermarket. Should I count myself as a costumer or will that fuck up the statistics?

>>16731006 (OP)
If you're buying something then you are a customer but if you are going to do this more than once, you will bias the sample

>>16731006 (OP)
Depends. Do you only go there when you would go otherwise for your groceries and the analysis is made on top of that, or do you go there only with the specific purpose of doing your analysis? Ideally you'd select a supermarket you never go to and count yourself out.

>>16731010
how is it bias? everytime I go there I'm there, so shouldn't it be okay?

>>16731013
>Ideally you'd select a supermarket you never go to and count yourself out.
why would I go there though? My reasoning is I want to know if the supermarket I go to has a lot of clients or few clients depending on what hour I go to, if I never go there it defeats the purpose of knowing

>>16731014
That's the problem, you only have data for when you are there.
You are severely going to inflate the statistics for gay porn mags and gin

>>16731020
then how do I get data for when I'm not there? If I'm not there I can't count how many people there are.
This is a problem that affects all the field of statistics. No statistic prediction is valid because of this bias

>>16731019
>why would I go there though?
For your analysis?
>I want to know if the supermarket I go to has a lot of clients or few clients depending on what hour I go to
In that case you might as well just go to the store as you normally do and count yourself in.

>>16731006 (OP)
Is one client statistically significant for the purpose of your analysis?

>>16731026
but if I go there for my analysis I'm there, counting that I'm not there is falsifying the results
>>16731028
I live in a rural town, not in San Francisco

>>16731006 (OP)
>>16731024
An example of how you could do it:
> Want to count average # of customers at each hour of the day, from 8 AM to 6 PM (or whenever the supermarket opens/closes)
> Block out [math]n[/math] days (ex, 100) to do the experiment
> Use a computer to randomly pick [math]m<n[/math] of those days (ex, 15)
> On each selected day, go to the supermarket for the whole day and count customers not including yourself

You're getting a good random sample of days, not biased by yourself as a customer. If you want to account for >>16731010 then just go grocery shopping at the times you normally would, and if it coincides with the days you're watching the store then count yourself. Alternatively you could get a friend to watch the store on the days you go shopping, which would probably be easier but none of us really have this option.

>>16731030
>but if I go there for my analysis I'm there, counting that I'm not there is falsifying the results
Are you interested in how many people go to the supermarket for the purpose of groceries, or for the purpose of other occurences? Why aren't you breaking in at midnight? Why won't you pay 1000 people to come into the store with you? What is the purpose of the analysis?

>>16731030
>I live in a rural town, not in San Francisco
Not an answer to my question.

>>16731031
what if counting the number of customers in that shop becomes a hot topic in scientific magazines and hundreds of scientists come to the shop to count customers? should you count them as well? if there are more scientists than normal customers it will skew the statistics

>>16731033
I wanna know how many people gobto the store for any reason, of course counting customers counts as a reason to come to the store. I want to know when I'm less likely to have to wait in the queue for the cashier to attend me. I can't go at night because the cashier isn't there and I would have to wait until the store opens again

>>16731037
Not only that, It will clearly inflate the gin and gay porn sales even higher

>>16731039
>I want to know when I'm less likely to have to wait in the queue for the cashier to attend me.
Well then that settles it. Here's a tip for you, wow often do you wait for yourself in the queue?

>>16731046
the problem is that every customer in the shop is doing the same statistics as me, and they will come when they percieve there are less people on the shop. so should I count them as well, or should I ignore them because they are counting how many people come to the supermarket?

>>16731049
Why don't you just camp outside the door in a tent and record how many enter or exit every 15 minutes?

>>16731046
also, the times I don't go to the supermarket the wait is lower because I'm not there and it's less people to attend

>>16731053
if everyone did that, no customers would ever enter the shop

>>16731046
how often do you wait in the queue when you're not buying anything?

>>16731055
Or they would all wait for the optimal time and then try to all enter at once.

>>16731058
I don't. The queue is for people who are buying stuff, I have no reason to go there when I'm not buying stuff

>>16731049
The number of people in the shop is irrelevant, the length of the queue is what you need to measure. Would a family of four with one cart count as one or four customers?

>>16731060
the point is that changing my behaviour would change the statistics of the problem since I'm a datapoint. If I refuse to enter and wait outside, the waiting times will be shorter because I'm not there

This is a problem of evolutionary fitness, not statistics. Regardless of queue length a person who is most able to move their cart to the end of the line will check out before someone who is less able to do so.

You're only minimizing queuing time, not exit time.

>>16731064
>If I refuse to enter and wait outside, the waiting times will be shorter because I'm not there.

If you refuse to enter and wait outside you're just adding to your wait time by exactly that much.

>>16731006 (OP)
As long as you already frequent that store, and make sure to buy something that time, I think you can count yourself.

>>16731075
but then I'll never measure 0 customers, because I'll always be there

>>16731049
>the problem is that every customer in the shop is doing the same statistics as me
I think you're grossly overestimating how much thought other customers give to this sort of thing.

Outside of end of business times and near holidays, there is rarely a time of day when things are so busy you're going to experience a significant backup at the queues.

>>16731076

Are you measuring "customers" or "people who impede your checkout?"

>>16731085
both
>>16731082
if I do it, what prevents them from doing it too?

OP is severely underestimating the contribution from checkout girl provision and till throughput rate, not even starting on the trolley vs handbaskets.
All of these factors are going to have the greatest impact on his wait times.
If a lot of guys go in to buy a pack of beer and cigarettes, minimal chitchat, in and out in 5 minutes their ultimate impact will be completely different to the mother doing her weekly shop with a creaking trolley's worth of groceries who takes half an hour to clear the checkout.

>>16731091
even if you take that into account, the problem of whether to count yourself still remains

>>16731089
>both
Ok so if you conflate two things don't be surprised when your data is fucked up and useless then?

>>16731091

>>16731096
I don't get this meme

>>16731107
People buying lottery tickets are often very slow and annoying about it. They don't know what tickets they want before getting there, they change their minds frequently, they often tend to be more talkative with the cashier and babble at them the whole time about some mundane shit, and some lower life forms will stand at the counter in the way of the next person and scratch the tickets right after paying for it in the hopes that they can have the cashier scan it and immediately cash in a small prize for more tickets. Fucking cunts.

>>16731096
Do they gots a fake $20?

>>16731107
NTA, but lotteryfags are the worst. had to pay inside last time i filled up my car at the gas station and ended up behind some fucking leathery ass bitch who clearly just cashed her welfare check and immediately went out and bought very specific and large numbers of different $1, $5, and $10 scratch cards (it took the better part of 20 minutes and she spent ~$300 in all), then just had the attendant immediately turn around and feed them all into the machine, lost basically all of the money, then bought a couple packs of newports, and left.

like fuck man at least the old cunts smashing video poker machines at the casino have the excuse of getting a dopamine buzz from all the lights and bells and whistles... this shit doesn't even have that but they still excuse dropping hundreds of dollars a month into a black hole. they're obnoxious fucks who hold everyone up for basically nothing.

>>16731058
I sometimes do it just to artifically lenghten the queue and make people more annoyed, then when i arrive at the cashier i do a very long, audible yawn, say "nevermind' and leave squeezing through the queue.

>>16731191
if you don't actually spend time with the cashier then you don't bother anyone. In fact, I'd be glad if someone in front of me on the queue walks away

>>16731330
>if you don't actually spend time with the cashier then you don't bother anyone.
Did you skip the yawn and nevermind part?

>>16731330
>if you don't actually spend time with the cashier then you don't bother anyone.
There's a guy by the door with a clipboard who is losing his mind