Principia Discordia

Principia Discordia => Techmology and Scientism => Topic started by: Kai on January 08, 2009, 05:40:18 PM

Title: Experimental statistics...
Post by: Kai on January 08, 2009, 05:40:18 PM
...is going to suck as much now as it did the first time 5 years ago.

Discuss.
Title: Re: Experimental statistics...
Post by: Harlequin on January 08, 2009, 05:59:54 PM
Have you got a supporting model for your hypothesis?
Title: Re: Experimental statistics...
Post by: AFK on January 08, 2009, 06:00:56 PM
What does the trendline suggest? 
Title: Re: Experimental statistics...
Post by: Kai on January 08, 2009, 06:10:29 PM
I WILL KILL A MOTHERFUCKER  :x
Title: Re: Experimental statistics...
Post by: Fuquad on January 08, 2009, 07:34:59 PM
The numbers show a strong trend towards that response.
Title: Re: Experimental statistics...
Post by: Kai on January 08, 2009, 08:00:21 PM
Right now I want to kick the whole damn grad school thing in the balls, but thats what the burnout did to me. Burn out last semester, and I've only recovered in bits and pieces since then.

Whole fucking thing overwhelms me. Might as well have a pity party about it and let everyone know how fucking emo I am.
Title: Re: Experimental statistics...
Post by: Telarus on January 08, 2009, 08:24:00 PM
I fee; you Kai. that's why I took this last term off (PROTIP: don't go more than 90 days not enrolled). Getting back into it this week with a Lit class, a scultping course, and a digital sculpting 'hard surfaces' course (I'm pretty sure they mean architecture and machinery).

I may hit you up for basic insect physiology when I get to the 'organic surfaces' digital modeling.
Title: Re: Experimental statistics...
Post by: Cramulus on January 08, 2009, 08:40:49 PM
aaaaaah fucking Statistics!  :argh!:

I feel your pain there. I suck at math, bad, but had to take a lot of stats to compete in Experimental Psych. I hated every decimal of it. I'LL GIVE YOU THE FUCKING P VALUE.
Title: Re: Experimental statistics...
Post by: Kai on January 08, 2009, 08:46:04 PM
Quote from: Telarus on January 08, 2009, 08:24:00 PM
I fee; you Kai. that's why I took this last term off (PROTIP: don't go more than 90 days not enrolled). Getting back into it this week with a Lit class, a scultping course, and a digital sculpting 'hard surfaces' course (I'm pretty sure they mean architecture and machinery).

I may hit you up for basic insect physiology when I get to the 'organic surfaces' digital modeling.

Can't do it. I have no way to live outside of food service without this degree, and I have no way to live without the loans I'm getting for housing.
Title: Re: Experimental statistics...
Post by: Cain on January 08, 2009, 09:18:26 PM
I absolutely despised stats when I did psych.

If you want, I have an e-book copy of All the Maths You Missed, But Need to Know for Grad School.  It contains some more advanced mathematical ideas, such as Fourier Analysis (well, advanced for me, you might know this), but if you think it could help, give me a shout out.
Title: Re: Experimental statistics...
Post by: Kai on January 08, 2009, 09:27:58 PM
Quote from: Cain on January 08, 2009, 09:18:26 PM
I absolutely despised stats when I did psych.

If you want, I have an e-book copy of All the Maths You Missed, But Need to Know for Grad School.  It contains some more advanced mathematical ideas, such as Fourier Analysis (well, advanced for me, you might know this), but if you think it could help, give me a shout out.

Thanks Cain. I don't think this is going to be HARD persay, but rather, its gonna be a repeat of all the shit I had 5 years ago and forgot, which sucks. my research doesn't really deal with statistical analysis thank fuck, the only stats needed will be cladistics based.

we're also not doing any survey design in this class (also thank fuck).
Title: Re: Experimental statistics...
Post by: Cain on January 08, 2009, 09:30:04 PM
Ah, yeah.  Understandable, its always annoying doing something you don't like for the second time.  Especially when it has little relevance.
Title: Re: Experimental statistics...
Post by: ñͤͣ̄ͦ̌̑͗͊͛͂͗ ̸̨̨̣̺̼̣̜͙͈͕̮̊̈́̈͂͛̽͊ͭ̓͆ͅé ̰̓̓́ͯ́́͞ on January 08, 2009, 10:52:35 PM
Quote from: Cain on January 08, 2009, 09:30:04 PM
Ah, yeah.  Understandable, its always annoying doing something you don't like for the second time.  Especially when it has little relevance.

Experimental statistics has little relevance?
Title: Re: Experimental statistics...
Post by: Cain on January 08, 2009, 10:57:55 PM
Quote from: Net on January 08, 2009, 10:52:35 PM
Quote from: Cain on January 08, 2009, 09:30:04 PM
Ah, yeah.  Understandable, its always annoying doing something you don't like for the second time.  Especially when it has little relevance.

Experimental statistics has little relevance?

Relatively speaking.

Obviously slightly more important for other topics, not so much for studying insects.
Title: Re: Experimental statistics...
Post by: ñͤͣ̄ͦ̌̑͗͊͛͂͗ ̸̨̨̣̺̼̣̜͙͈͕̮̊̈́̈͂͛̽͊ͭ̓͆ͅé ̰̓̓́ͯ́́͞ on January 08, 2009, 11:00:56 PM
It was my impression that it's crucially important to making experimental data scientifically useful, regardless of the field.
Title: Re: Experimental statistics...
Post by: ñͤͣ̄ͦ̌̑͗͊͛͂͗ ̸̨̨̣̺̼̣̜͙͈͕̮̊̈́̈͂͛̽͊ͭ̓͆ͅé ̰̓̓́ͯ́́͞ on January 08, 2009, 11:06:23 PM
Also, burnout sucks. I hope you get through it soon Kai.

I've had my fair share and the main thing I've learned from it is to get help sooner rather than later.
Title: Re: Experimental statistics...
Post by: Kai on January 08, 2009, 11:06:48 PM
Quote from: Cain on January 08, 2009, 10:57:55 PM
Quote from: Net on January 08, 2009, 10:52:35 PM
Quote from: Cain on January 08, 2009, 09:30:04 PM
Ah, yeah.  Understandable, its always annoying doing something you don't like for the second time.  Especially when it has little relevance.

Experimental statistics has little relevance?

Relatively speaking.

Obviously slightly more important for other topics, not so much for studying insects.

Oh, its useful for any experimental design that requires data manipulation, and that includes all of the biological sciences


However, my focus is systematics, and for the particular taxonomic research I am working on it is not going to be necessary. The 'experiment' I am conducting doesn't require measurement sampling for estimation of parameters in any formal fashion.
Title: Re: Experimental statistics...
Post by: Kai on January 08, 2009, 11:09:03 PM
Quote from: Net on January 08, 2009, 11:06:23 PM
Also, burnout sucks. I hope you get through it soon Kai.

I've had my fair share and the main thing I've learned from it is to get help sooner rather than later.

Well, part of that is the latent depression hanging around. I have people who help with that, but in the end its up to me to pull myself out of it, and its not easy. Its the same amount of effort that goes into heavy labor, albeit reflected in a different kind of result.

Breathing and exercise helps.
Title: Re: Experimental statistics...
Post by: ñͤͣ̄ͦ̌̑͗͊͛͂͗ ̸̨̨̣̺̼̣̜͙͈͕̮̊̈́̈͂͛̽͊ͭ̓͆ͅé ̰̓̓́ͯ́́͞ on January 08, 2009, 11:16:38 PM
Man, I'm actually extremely interested in experimental statistics right now.

Wanna trade?

I've got Corporate Identity, Art Direction, Web Design II, and Aesthetics for you to choose from.
Title: Re: Experimental statistics...
Post by: Kai on January 08, 2009, 11:38:21 PM
Quote from: Net on January 08, 2009, 11:16:38 PM
Man, I'm actually extremely interested in experimental statistics right now.

Wanna trade?

I've got Corporate Identity, Art Direction, Web Design II, and Aesthetics for you to choose from.


Aesthetics sounds fun.

I'll switch ya stats for philosophy.
Title: Re: Experimental statistics...
Post by: Faust on January 09, 2009, 01:49:59 AM
I hate statistics. Sadly it is central to process modeling for industry.
The worst is the six sigma ethos were supposed to absorb.
Its pretty cool how precise quality systems have become, but I don't want to eat, think and breath it.
Title: Re: Experimental statistics...
Post by: Cain on January 09, 2009, 10:32:45 AM
Quote from: Net on January 08, 2009, 11:00:56 PM
It was my impression that it's crucially important to making experimental data scientifically useful, regardless of the field.


Not according to Kai, who said

Quotemy research doesn't really deal with statistical analysis thank fuck, the only stats needed will be cladistics based.

Which would be, uh, that context thing I was talking about.
Title: Re: Experimental statistics...
Post by: LMNO on January 09, 2009, 03:09:43 PM
Kai, is there a way you could re-think why you're taking statistics?

I mean, I know it doesn't apply to your field of study at school, but maybe it could apply to some other part of your life?  I dunno exactly what, but I used to try and re-purpose my boring classes so they applied to something different than what they were intended to address.


Wow, that sounds pretty vague and unhelpful.
Title: Re: Experimental statistics...
Post by: Kai on January 09, 2009, 04:41:24 PM
Quote from: LMNO on January 09, 2009, 03:09:43 PM
Kai, is there a way you could re-think why you're taking statistics?

I mean, I know it doesn't apply to your field of study at school, but maybe it could apply to some other part of your life?  I dunno exactly what, but I used to try and re-purpose my boring classes so they applied to something different than what they were intended to address.


Wow, that sounds pretty vague and unhelpful.

I don't know really. Its possible at some point in future research I will need to make some sense of a large set of data. I just don't work with numbers enough in a complex fashion in any aspect of my life, I mean, I'm not about to go into banking or the stock market any time soon, I'm not interested in sports or anything like that....
Title: Re: Experimental statistics...
Post by: AFK on January 09, 2009, 05:57:46 PM
Quote from: Kai on January 08, 2009, 09:27:58 PM
Quote from: Cain on January 08, 2009, 09:18:26 PM
I absolutely despised stats when I did psych.

If you want, I have an e-book copy of All the Maths You Missed, But Need to Know for Grad School.  It contains some more advanced mathematical ideas, such as Fourier Analysis (well, advanced for me, you might know this), but if you think it could help, give me a shout out.

Thanks Cain. I don't think this is going to be HARD persay, but rather, its gonna be a repeat of all the shit I had 5 years ago and forgot, which sucks. my research doesn't really deal with statistical analysis thank fuck, the only stats needed will be cladistics based.

we're also not doing any survey design in this class (also thank fuck).

Ooh survey design.  That's one of the things I do for a living.  You can always run something by me. 
Title: Re: Experimental statistics...
Post by: Vene on January 10, 2009, 01:23:36 AM
Quote from: Kai on January 09, 2009, 04:41:24 PM
Quote from: LMNO on January 09, 2009, 03:09:43 PM
Kai, is there a way you could re-think why you're taking statistics?

I mean, I know it doesn't apply to your field of study at school, but maybe it could apply to some other part of your life?  I dunno exactly what, but I used to try and re-purpose my boring classes so they applied to something different than what they were intended to address.


Wow, that sounds pretty vague and unhelpful.

I don't know really. Its possible at some point in future research I will need to make some sense of a large set of data. I just don't work with numbers enough in a complex fashion in any aspect of my life, I mean, I'm not about to go into banking or the stock market any time soon, I'm not interested in sports or anything like that....
Is there any chance you would work with insect genes?  The chi-square test was my best friend when I took genetics last semester.
Title: Re: Experimental statistics...
Post by: Kai on January 10, 2009, 04:51:48 AM
I so do not want to do the gene barcoding for this project. I'll do the sequence comparisons if I have to, just to establish associations, but I just do NOT want to do more than that.
Title: Re: Experimental statistics...
Post by: ñͤͣ̄ͦ̌̑͗͊͛͂͗ ̸̨̨̣̺̼̣̜͙͈͕̮̊̈́̈͂͛̽͊ͭ̓͆ͅé ̰̓̓́ͯ́́͞ on January 10, 2009, 08:53:42 AM
Quote from: Cain on January 09, 2009, 10:32:45 AM
Quote from: Net on January 08, 2009, 11:00:56 PM
It was my impression that it's crucially important to making experimental data scientifically useful, regardless of the field.


Not according to Kai, who said

Quotemy research doesn't really deal with statistical analysis thank fuck, the only stats needed will be cladistics based.

Which would be, uh, that context thing I was talking about.

That would be, uh, that impression thing I was talking about.
Title: Re: Experimental statistics...
Post by: Cain on January 10, 2009, 03:27:14 PM
Quote from: Net on January 10, 2009, 08:53:42 AM
Quote from: Cain on January 09, 2009, 10:32:45 AM
Quote from: Net on January 08, 2009, 11:00:56 PM
It was my impression that it's crucially important to making experimental data scientifically useful, regardless of the field.


Not according to Kai, who said

Quotemy research doesn't really deal with statistical analysis thank fuck, the only stats needed will be cladistics based.

Which would be, uh, that context thing I was talking about.

That would be, uh, that impression thing I was talking about.

Look, if you're trying to pick a fight, go buy a punching bag.

Cain,
not taking your bait.
Title: Re: Experimental statistics...
Post by: rong on January 10, 2009, 03:51:55 PM
100% of statisticis is bullshit
Title: Re: Experimental statistics...
Post by: Vene on January 10, 2009, 04:43:54 PM
Quote from: rong on January 10, 2009, 03:51:55 PM
100% of statisticis is bullshit
Meh, it's more of its own language.  And those who don't understand it can be very easily manipulated.  Too bad that does make up a lot of statistics.
Title: Re: Experimental statistics...
Post by: Kai on January 10, 2009, 05:33:16 PM
Quote from: Vene on January 10, 2009, 04:43:54 PM
Quote from: rong on January 10, 2009, 03:51:55 PM
100% of statisticis is bullshit
Meh, it's more of its own language.  And those who don't understand it can be very easily manipulated.  Too bad that does make up a lot of statistics.

Its useful when you have population parameters that need to be described, either by finding the parameter itself (near impossible) or by taking a sample and finding a statistic of the parameter.
Title: Re: Experimental statistics...
Post by: Vene on January 10, 2009, 05:36:37 PM
Quote from: Kai on January 10, 2009, 05:33:16 PM
Quote from: Vene on January 10, 2009, 04:43:54 PM
Quote from: rong on January 10, 2009, 03:51:55 PM
100% of statisticis is bullshit
Meh, it's more of its own language.  And those who don't understand it can be very easily manipulated.  Too bad that does make up a lot of statistics.

Its useful when you have population parameters that need to be described, either by finding the parameter itself (near impossible) or by taking a sample and finding a statistic of the parameter.
Yep, which is largely what I used statistics for in genetics.
Title: Re: Experimental statistics...
Post by: rong on January 10, 2009, 05:43:53 PM
i was half joking.  i.e. 76% of statistics are made up.

seriously, i think statistics is a religion.  a religion i have no faith in.  i mean, statistically, the average person has one breast and one testicle.  i'm mainly just jaded because people that think they are smarter than me errantly use statistics to make my job suck more.
Title: Re: Experimental statistics...
Post by: Kai on January 10, 2009, 05:51:04 PM
Quote from: rong on January 10, 2009, 05:43:53 PM
i was half joking.  i.e. 76% of statistics are made up.

seriously, i think statistics is a religion.  a religion i have no faith in.  i mean, statistically, the average person has one breast and one testicle.  i'm mainly just jaded because people that think they are smarter than me errantly use statistics to make my job suck more.

Nahh, statistics are usually not made up. What are made up are the graphs, tables, writings and hidden biases that are inserted into statistics and are used to exibit someones view or justify someone's cause that purposefully mislead whoever is looking at them. The reason its easy to do is because people don't know statistics. If you know statistics its easy to see that sort of bias.

And statistically, number of breasts or testicles should be taken in a random stratified sample, separating male bodied and female bodied people and surveying separately and you wouldn't HAVE that problem. The average (or mean if you prefer that) number of breasts on a human doesn't make sense when breasts are almost exclusively found on female bodied people, therefore, its experimental design and not statistics is the error.

Statistics is just like computers, garbage in, garbage out.
Title: Re: Experimental statistics...
Post by: rong on January 10, 2009, 05:58:12 PM
Quote from: Kai on January 10, 2009, 05:51:04 PM
And statistically, number of breasts or testicles should be taken in a random stratified sample, separating male bodied and female bodied people and surveying separately and you wouldn't HAVE that problem. The average (or mean if you prefer that) number of breasts on a human doesn't make sense when breasts are almost exclusively found on female bodied people, therefore, its experimental design and not statistics is the error.


yeah, but see how many words you have to use just to start using statistics to describe something so obvious and mundane in a remotely "useful" way?

significance is and always will be an opinion.  statistical or not.
Title: Re: Experimental statistics...
Post by: Kai on January 10, 2009, 06:18:09 PM
Quote from: rong on January 10, 2009, 05:58:12 PM
Quote from: Kai on January 10, 2009, 05:51:04 PM
And statistically, number of breasts or testicles should be taken in a random stratified sample, separating male bodied and female bodied people and surveying separately and you wouldn't HAVE that problem. The average (or mean if you prefer that) number of breasts on a human doesn't make sense when breasts are almost exclusively found on female bodied people, therefore, its experimental design and not statistics is the error.


yeah, but see how many words you have to use just to start using statistics to describe something so obvious and mundane in a remotely "useful" way?

significance is and always will be an opinion.  statistical or not.

Then you use that many words. You use however many words you need to use to communicate what you are trying to communicate. Anything else is laziness.
Title: Re: Experimental statistics...
Post by: Vene on January 10, 2009, 06:19:09 PM
Quote from: rong on January 10, 2009, 05:58:12 PM
Quote from: Kai on January 10, 2009, 05:51:04 PM
And statistically, number of breasts or testicles should be taken in a random stratified sample, separating male bodied and female bodied people and surveying separately and you wouldn't HAVE that problem. The average (or mean if you prefer that) number of breasts on a human doesn't make sense when breasts are almost exclusively found on female bodied people, therefore, its experimental design and not statistics is the error.


yeah, but see how many words you have to use just to start using statistics to describe something so obvious and mundane in a remotely "useful" way?

significance is and always will be an opinion.  statistical or not.
You mean it's complex and contains subtleties, oh no!
Title: Re: Experimental statistics...
Post by: rong on January 10, 2009, 06:25:26 PM
you got me.  the real reason i don't like teh statistics is cuz i'm too dumb to understand 'em.

guess all those regressions are linear after all. . .
Title: Re: Experimental statistics...
Post by: Kai on January 10, 2009, 06:29:46 PM
Quote from: rong on January 10, 2009, 06:25:26 PM
you got me.  the real reason i don't like teh statistics is cuz i'm too dumb to understand 'em.

guess all those regressions are linear after all. . .

Oh, don't pull that shit.
Title: Re: Experimental statistics...
Post by: Mesozoic Mister Nigel on January 10, 2009, 07:30:19 PM
Quote from: rong on January 10, 2009, 05:43:53 PM
i was half joking.  i.e. 76% of statistics are made up.

seriously, i think statistics is a religion.  a religion i have no faith in.  i mean, statistically, the average person has one breast and one testicle.  i'm mainly just jaded because people that think they are smarter than me errantly use statistics to make my job suck more.

Statistics is a tool.
Title: Re: Experimental statistics...
Post by: ñͤͣ̄ͦ̌̑͗͊͛͂͗ ̸̨̨̣̺̼̣̜͙͈͕̮̊̈́̈͂͛̽͊ͭ̓͆ͅé ̰̓̓́ͯ́́͞ on January 11, 2009, 02:53:14 AM
Quote from: Cain on January 10, 2009, 03:27:14 PM
Quote from: Net on January 10, 2009, 08:53:42 AM
Quote from: Cain on January 09, 2009, 10:32:45 AM
Quote from: Net on January 08, 2009, 11:00:56 PM
It was my impression that it's crucially important to making experimental data scientifically useful, regardless of the field.


Not according to Kai, who said

Quotemy research doesn't really deal with statistical analysis thank fuck, the only stats needed will be cladistics based.

Which would be, uh, that context thing I was talking about.

That would be, uh, that impression thing I was talking about.

Look, if you're trying to pick a fight, go buy a punching bag.

Cain,
not taking your bait.

It looks more like that is what you were trying to do.
Title: Re: Experimental statistics...
Post by: ñͤͣ̄ͦ̌̑͗͊͛͂͗ ̸̨̨̣̺̼̣̜͙͈͕̮̊̈́̈͂͛̽͊ͭ̓͆ͅé ̰̓̓́ͯ́́͞ on January 15, 2009, 08:44:04 AM
I'm sorry I made the above assumption about your intentions Cain.
Title: Re: Experimental statistics...
Post by: willem on January 25, 2009, 09:38:54 AM
[wiseass intermezzo: Numbers don't lie to people. People lie with numbers.]

Are you sure that the difference between the two observed populations is not significant? What certainty level are we working with here?