a relationship between the independent variable and the dependent variable the problem i have in an experiment is that there's always what are called potential confounds right there might be an alternative explanation for the effect that is different from my proposed independent variable so you have to do what's called controlling for the confounds you have to structure your experiment to rule out uh those alternative explanations if you don't your results are inherently confounded which means they're deeply ambiguous you you cannot derive any clear um uh uh conclusion from them so what what do what do we do well why is that a problem well here's the problem the alternative explanations again in number are right like so you're trying to see uh you know like you know does water dissolve no that's water the salt dissolve in water well of course it does and then you go to somebody says no it doesn't and you go to pour in the salt and just as they're pouring in they spray the salt with plastic so it doesn't dissolve they go aha salt does not dissolve in water you go no no i mean i don't mean salt that ha that has just recently been covered uh with plastic and they go okay do it again and they go and then just as you're about to put they flash freeze it and the salt stays on the surface and you say haha it doesn't dissolve in water you say no no i don't mean salt in frozen water i mean in room temperature water and just about before you're going to pour your salt in they pour in another chemical that prevents the salt from dissolving and they go aha these oh no i mean pure water and they're pure water and you see where this goes right this is a problem brought up a long time ago by hempel and then japanese published on it which is the number of possible counterfactuals is indefinitely large but we don't rule them all out right we can't because so which ones do we control for well we control for the ones that strike us as the most reasonable that make good sense that should be taken seriously that we should pay attent