"univariate analysis"

[excal]

But this is the quintessential grey area. Making judgments based on race serve no useful purpose in anything OTHER than a one-dimensional analysis. You'd have to literally know NOTHING else about their character for it to work.

It's purely academic, and when you try to make a case for it on a wider level, such as applying it to a multi-dimensional scenario AS YOU ARE TRYING TO DO. You are using it to justify it on the basis of the article you presented, your argument falls to pieces COMPLETELY and it just makes you look like a racist and a bigot.

That post was addressed to both of you

My post was to attack only the severe generalisation you make that all one-dimensional analyses are a failure by design.

"It's sharp, therefore I shouldn't touch it.". Does that univariate analysis fail, even when there are other factors involved (cost of hospital stay, need to touch blade etc)?

[enwiabe]

Should depend on how sharp the object is, no? Do you not touch a knife because it isn't sharp? No, you grab it by the handle. A pencil is sharp but people still touch it. How sharp is the pencil?

[excal]

Should depend on how sharp the object is, no? Do you not touch a knife because it isn't sharp? No, you grab it by the handle. A pencil is sharp but people still touch it. How sharp is the pencil?

You make a decision (hold by the handle) because the blade of the knife is sharp enough to be harmful (your single variable).

You make a decision (to be able to touch the sharp of the pencil) because the pencil is not sharp enough to be harmful (your single variable).

The point is, your premise is based on a overly generalised assumption that all univariate analyses are faulty by design. I'm merely providing an example to prove your generalisation to be incorrect.

[enwiabe]

That's not univariate though... the sharpness is one variable, how much sharpness YOU can tolerate is another variable. That's a two-dimensional analysis.

Fail.

Also, I said applying one-dimensional APPROXIMATIONS to multi-dimensional problems failed.

... Again, you simply fail. Methinks you should have just stayed out of this convo sweet'ums.

[dekoyl]

Sorry to interrupt but..
^Ouch.

[excal]

That's not univariate though... the sharpness is one variable, how much sharpness YOU can tolerate is another variable. That's a two-dimensional analysis.

Fail.

Also, I said applying one-dimensional APPROXIMATIONS to multi-dimensional problems failed.

... Again, you simply fail. Methinks you should have just stayed out of this convo sweet'ums.

How did it fail?

Do you often think about how sharp a blade is before running your finger down it? I think it's pretty common sense that because knifes are sharp, we do not touch the blade in that manner. It may be multi-variate as you and I pointed out, but we apply (quite successfully in this case too) that univariate approximate (the assumption that all knife blades are sharp) to avoid getting hurt.

To apply a typical enwiabe ending to an argument: Stop grasping at straws, you fial.

[enwiabe]

But that's not univariate. How sharp is too sharp? You need something to scale it against. I.E. A second variable detailing how much sharpness you can tolerate.

What quantifies sharpness?

And as in the scenario before. What quantifies productivity? Is it religion? No, it's the causal upbringing of said person and the values with which they were instilled INDIVIDUALLY that quantifies productivity.

[excal]

Nothing's univariate. That's a given. I'm saying, however, a univariate approximation is not destined to fail by design. It can, and does, work and we do it all the time in our day to day lives without even realising it.

I am trying to show you that your generalisation is overly broad and incorrect.

[enwiabe]

Nothing's univariate. That's a given. I'm saying, however, a univariate approximation is not destined to fail by design. It can, and does, work and we do it all the time in our day to day lives without even realising it.

I am trying to show you that your generalisation is overly broad and incorrect.

You tried but you failed and haven't provided a follow-up example. I just proved that the univariate approximation failed without considering the second variable. Where is your counter-example?

[brendan]

i lost both of you guys when you started talking abt univariate stuff.

[excal]

Nothing's univariate. That's a given. I'm saying, however, a univariate approximation is not destined to fail by design. It can, and does, work and we do it all the time in our day to day lives without even realising it.

I am trying to show you that your generalisation is overly broad and incorrect.

You tried but you failed and haven't provided a follow-up example. I just proved that the univariate approximation failed without considering the second variable. Where is your counter-example?

But that's not univariate. How sharp is too sharp? You need something to scale it against. I.E. A second variable detailing how much sharpness you can tolerate.

There is no need for a counter-example has you've haven't actually said anything to counter the example provided earlier. I never said the situation ('how you should hold knife') was a situation that was one with only a single variable. It is certainly a situation with multiple considerations involved within it as I said (and you reiterated: 'But that's not univariate. How sharp is sharp?').

Basically you're saying that the situation I describe is a multivariate situation and therefore my example fails.

Evidently, you're misinterpreting what I've been saying. I'm providing an example where a single variable assumption (blade is sharp) can apply to a multivariate (how sharp? how much will it cost for a hospital visit if I cut myself? what's the risk?) situation and be successful (risk of being cut by blade reduced by not touching it). I am NOT saying the knife situation is univariate as that would make the example irrelevant anyway. I am demonstrating a successful use of a single variable assumption (the blade is sharp - do not touch) in a multi-variable situation (how sharp the blade is / consequences and risks of an accident).

In other words, I contend that people do make a single variable assumptions ('it's a blade, therefore it will be sharp and hurt me; therefore I will not touch the blade'), without considering the 'amount' of sharpness however it is quantified. And, contrary to your prior assumption/generalisation that all single variable analyses are failures when applied to multivariate problems, they can be successful assumptions.