quickie experiment ideas

[Psyche]

So, I study psychology, especially the cognitive and social stuff. And I've recently come upon quite a bit of funding for my research and at the same time have become adept at running computer based experiments (participant sits at a computer, takes input, responds with their computer) really cheaply and with large sample sizes thanks to learning javascript and how to use Mechanical Turk.

I've sort of been overwhelmed lately by the possibilities that all this presents for me, and wonder if you guys can help me identify some of particular interest. I'd like to field ideas you have for simple, or even complicated experiments people can participate in by sitting at a computer for up to around 30 minutes. If the question they answer is interesting and hasn't been addressed in a better way yet by someone else, then I'll put it together and run it within the week. If it becomes a paper, you'll be acknowledged or even included as an author of the result, depending on your preference and level of contribution.

To give you an idea of how simple an experiment idea can be, let me describe one that's already been done, and only took a few minutes to put together. One experiment presented people several squares with different shades of grey - lighter shades and darker shades. In one condition, they were asked, "Which of these is the best example of the color gray?" In the other condition, "Which of these is the best example of the color grey?". Just the spelling of the word "grey" was changed, and everything else about the experiment left the same. It turns out that people significantly prefer lighter shades for "gray" than they do for "grey". That's an interesting quirk, isn't it?

Now, that's not necessarily a great experiment. It's actually far on the low end of experiments I'd like to do in terms of complication and relevance. But now it's so cheap and easy to run that we can do it anyway. And that's the magic motivating this thread.

I'm not totally optimistic about how much traffic this thread will get, but at the very least it'll be a spot where I can brainstorm ideas for myself. So, whatevs.

[Allomancer]

okay? what do you expect us to say?

[Psyche]

"run an experiment that finds out if cat people are dumber than dog people"

[Psyche]

"are people less likely to like the color blue if they live further inland"

[Allomancer]

"do carrots cause car crashes"

[inte]

"can jet fuel melt steel beams"

[Psyche]

One thing I study a lot is randomness perception. Explicitly, a lot of people have the mistaken belief that, if something happens more frequently than normal during some period, it will happen less frequently in the future, or that, if something happens less frequently than normal during some period, it will happen more frequently in the future (presumably as a means of balancing nature). And even those who don't subscribe to the gambler's fallacy still exhibit a sort of over-alternation bias. In a series of coin tosses, the likelihood that a Heads will follow a Tails is 50% - that's what a random alternation rate looks like. However, people think that sequences where the alternation rate is more like 55% or even 60% look

more

random than truly random sequences. And no one seems to know for sure why. I'm going to try to survey the literature for hypothetical explanations and think up short experiments that would falsify each of them.

[Allomancer]

One thing I study a lot is randomness perception. Explicitly, a lot of people have the mistaken belief that, if something happens more frequently than normal during some period, it will happen less frequently in the future, or that, if something happens less frequently than normal during some period, it will happen more frequently in the future (presumably as a means of balancing nature). And even those who don't subscribe to the gambler's fallacy still exhibit a sort of over-alternation bias. In a series of coin tosses, the likelihood that a Heads will follow a Tails is 50% - that's what a random alternation rate looks like. However, people think that sequences where the alternation rate is more like 55% or even 60% look

more

random than truly random sequences. And no one seems to know for sure why. I'm going to try to survey the literature for hypothetical explanations and think up short experiments that would falsify each of them.

I've noticed that, too! Like, if you flip 3 heads in a row, people think, "okay, theres no way we can flip heads again"

[Psyche]

A weird thing is that in a lot of situations, the bias is the exact opposite - they expect a streak to continue instead of ending. For example, suppose a dude named Paul has won a few blackjack hands in a row; a lot of people see he's on a winning streak and expect it to keep going, at least for a while. That's called the "hot hand fallacy" and comes up in lots of contexts where random events correspond to winning or losing.

[Raskolnikov]

Maybe it was more useful to classify things as either continuing or alternating than random, as either is a pattern of sorts?

[shaft.ed]

However, people think that sequences where the alternation rate is more like 55% or even 60% look more random than truly random sequences. And no one seems to know for sure why.

I thought this was just due to a misunderstanding of what random means in a lay sense. People think random just means 'well mixed'

Given only one example ANY sequence could have been randomly generated (with a relatively representative odds distribution)

[shaft.ed]

"are people less likely to like the color blue if they live further inland"

i thought there are tribes in Africa that don't perceive the color blue.
But they do have numerous shades of green and are better at perceiving slight differences in green hues

Also isn't blue the last primary color to get 'named' in most languages?

[Nahdia]

Are people in an annoyed/frustrated state more likely identify with extremist views?

[shaft.ed]

effects of procrastination on performance:

Give people a task and tell them it takes most people roughly 10 minutes to complete although inform them they will have to be at the computer station for 30 minutes regardless of if they complete it early.

Give one group access to an internet browser on the computer and not the other.

[Nahdia]

(not sure if there's an ethical dilemma with this one. may be difficult to achieve with online respondents) Does offering an impossible to achieve incentive lower overall performance?
For example, give a simple puzzle to solve that's designed to take at least five minutes no matter what. In the experimental group, tell them they'll be compensated extra if they do it in under three.

[Raskolnikov]

Well, unless/until the person knows it's impossible to achieve, how would it be any different in influencing them as an incentive?

[Nahdia]

once they realize they've failed the incentive, do they start performing worse than if they never had the incentive at all?
if you find a clever enough way to mask it, you could also add another dimension and input a self-report at the end to see if they would want to try again if given the opportunity to see if they figured out that the incentive is impossible or not.

[Psyche]

I thought this was just due to a misunderstanding of what random means in a lay sense. People think random just means 'well mixed

That's one hypothesis. Not clear why a .6 alternation rate sequence would look more well-mixed than a .5 one on average, though.

[Psyche]

"are people less likely to like the color blue if they live further inland"

i thought there are tribes in Africa that don't perceive the color blue.
But they do have numerous shades of green and are better at perceiving slight differences in green hues

Also isn't blue the last primary color to get 'named' in most languages?

idk man

Are people in an annoyed/frustrated state more likely identify with extremist views?

I don't think I can do this online. It's too easy to quit an experiment when it's online. Also they'll review me negatively and I'll have trouble.

[Psyche]

effects of procrastination on performance:

Give people a task and tell them it takes most people roughly 10 minutes to complete although inform them they will have to be at the computer station for 30 minutes regardless of if they complete it early.

Give one group access to an internet browser on the computer and not the other.

Mmmmm, this is sort of doable by making the experiment full screen vs not, but in an online context it's hard to keep disengaged participants from using their time efficiently. I could use the honor system, I spose. Or generate some placeholder task for after participants finish? But...

[Psyche]

(not sure if there's an ethical dilemma with this one. may be difficult to achieve with online respondents) Does offering an impossible to achieve incentive lower overall performance?
For example, give a simple puzzle to solve that's designed to take at least five minutes no matter what. In the experimental group, tell them they'll be compensated extra if they do it in under three.

I'm guessing the incentive condition should be very obviously unachievable. Or I can go along a spectrum, making the condition as low as 1 minute for some partipants and 4 for others. Post-experiment surveys can assure that I'' achieving the desired effect, as you partially noted.

This is an experiment I can do! Only question is whether it's already been done before. I'll check it out.

[Psyche]

i wonder what it would mean if i gave people a randomness perception task instructing them to produce/identify patterns typical of concepts related and different from randomness. For example groups may be variously told: "Choose the random sequence", "Choose the well-mixed sequence", "Choose the orderly sequence", "meaningful" and so forth, also including concepts that are variously mistakable for randomness, like "interesting", "checkered", "simple"

argument would be that much research on randomness perception is actually about X

of course i'd have to follow up then with another experiment that evince that human responses are consistently aimed at capturing X instead of randomness

and on the other hand you're left with an unconvincing explanation of cases where people definitely have randomness in mind, such as when they expect a Tails after a series of coin tosses flipping Heads

[Brandi]

I'm pretty bat at wording things so maybe you can understand the idea of my question here:

Are people more or less likely to be discouraged to pursue certain "dream"* fields of work if they are generally thought of as creative ones?

*(As in something someone cares about and wants to do- typically not something like working at mcdonalds)

[Brandi]

sorry if im too late, i really like the concept of this thread but am not sure if Id be good at contributing

i hope to see your posts about your research though, even if we scummers aren't helpful to you

[Chevre]

>blow up a balloon
>cut a hoop of material from a plastic grocery sack
>rub both with wool or fake fur
>hover the hoop above the balloon like you're gosh-dang magic or somethin
>profit!