wdprice3 BinaryBuffonary 45912 Posts user info edit post |
I have results that say the following:
A = B B = C A != C
lolwut? maybe I should have paid attention when I took the class. 9/28/2010 6:09:46 PM |
BubbleBobble :3 114232 Posts user info edit post |
k 9/28/2010 6:11:35 PM |
wdprice3 BinaryBuffonary 45912 Posts user info edit post |
k doesn't come into play for these analyses. that's a different data set. 9/28/2010 6:13:13 PM |
BubbleBobble :3 114232 Posts user info edit post |
Quote : | "FOR
THESE
ANALYSES" |
Quote : | "DIFFERENT
DATA
SET" |
9/28/2010 6:15:16 PM |
Joie begonias is my boo 22491 Posts user info edit post |
i thought stats was easy. i kinda liked it.
but it was the only math class i "got". i just completely understood it.
i hoever have no idea what you are talking about lol
[Edited on September 28, 2010 at 6:18 PM. Reason : V oops] 9/28/2010 6:15:19 PM |
Mr. Joshua Swimfanfan 43948 Posts user info edit post |
9/28/2010 6:17:59 PM |
EMCE balls deep 89740 Posts user info edit post |
I mean, are your df the same? Are you using an identical significance value? Universal sample sizes?
come on wdprice3, help US help YOU. 9/28/2010 6:18:02 PM |
GeniuSxBoY Suspended 16786 Posts user info edit post |
Quote : | "A = B B = C A != C" |
Sounds like a wrong answer9/28/2010 6:20:12 PM |
khcadwal All American 35165 Posts user info edit post |
Quote : | "i thought stats was easy. i kinda liked it.
but it was the only math class i "got". i just completely understood it.
i hoever have no idea what you are talking about lol" |
ditto. i got dat 190% A+ in stat. it was one of several i got during my college career. A+s always made me feel warm and fuzzy inside. especially because most of the classes i got them in i rarely attended 9/28/2010 6:21:37 PM |
BigMan157 no u 103353 Posts user info edit post |
more like I hate random processes 9/28/2010 6:22:36 PM |
jstpack All American 2184 Posts user info edit post |
98.3% of all U.S. citizens hate statistics 9/28/2010 6:22:57 PM |
GeniuSxBoY Suspended 16786 Posts user info edit post |
likely probability 9/28/2010 6:23:22 PM |
Nerdchick All American 37009 Posts user info edit post |
the more you learn about statistics, the less you will trust them 9/28/2010 6:25:51 PM |
BigMan157 no u 103353 Posts user info edit post |
it is witchcraft and devilty 9/28/2010 6:26:40 PM |
wdprice3 BinaryBuffonary 45912 Posts user info edit post |
^^exactly
Quote : | "i thought stats was easy." |
as did I. but class != actual use of stats
Quote : | "I mean, are your df the same? Are you using an identical significance value? Universal sample sizes?" |
oh hell no son. this is real world data with shit all over the place. nothing about these data sets allows for basic assumptions.... AFAIK. well I'm using 0.05 for the significance level for all tests.
[Edited on September 28, 2010 at 6:30 PM. Reason : .]9/28/2010 6:29:46 PM |
0EPII1 All American 42535 Posts user info edit post |
stats is fucking easy, at least at the undergrad level.
Quote : | "maybe I you should have paid attention when I you took the class." |
[Edited on September 28, 2010 at 6:31 PM. Reason : ^ ok, fair enough]9/28/2010 6:30:22 PM |
wdprice3 BinaryBuffonary 45912 Posts user info edit post |
^ Quote : | "class != actual use of stats" |
when it takes a well known stats guy two weeks to come up with a proper model and testing methods, stats is not fucking easy.
[Edited on September 28, 2010 at 6:32 PM. Reason : ^ok, gotcha in the edit]9/28/2010 6:31:27 PM |
0EPII1 All American 42535 Posts user info edit post |
^
Quote : | "^ ok, fair enough" |
9/28/2010 6:32:12 PM |
wdprice3 BinaryBuffonary 45912 Posts user info edit post |
^
Quote : | " ^ok, gotcha in the edit" |
9/28/2010 6:33:08 PM |
jokar2694 All American 801 Posts user info edit post |
Quote : | "98.3% of all U.S. citizens hate statistics" |
9/28/2010 6:38:50 PM |
seedless All American 27142 Posts user info edit post |
99% of statistics is 100% bullshit. 9/28/2010 6:40:05 PM |
LeonIsPro All American 5021 Posts user info edit post |
9/28/2010 6:41:20 PM |
BigMan157 no u 103353 Posts user info edit post |
9/28/2010 8:43:44 PM |
datman All American 4812 Posts user info edit post |
i had one class of that in undergraduate and i was -------- that close to shooting my teacher in the face 9/29/2010 1:04:20 AM |
NC86 All American 9134 Posts user info edit post |
stats is cake for undergrad.
grad level statistics is harder but not grad level hard. 9/29/2010 1:07:17 AM |
AstralAdvent All American 9999 Posts user info edit post |
I have a ST371 test tomorrow.
I went to class once (2 days ago) to turn in homework, and left as soon as he started talking about what a sample size was.
cool huh?
I'm AstralAdvent and i approved this message. 9/29/2010 1:56:45 AM |
shmorri2 All American 10003 Posts user info edit post |
A results B and B results C but A does not result C ? 9/29/2010 2:18:58 AM |
wdprice3 BinaryBuffonary 45912 Posts user info edit post |
^correct 11/12/2010 2:21:23 PM |
jbtilley All American 12796 Posts user info edit post |
When I took statistics it was nothing more than a class to show proofs for things related to calculus. It sucked. 11/12/2010 2:32:29 PM |
wdprice3 BinaryBuffonary 45912 Posts user info edit post |
le sigh. this isn't a class. this is real shit. 11/12/2010 2:33:23 PM |
Mr. Joshua Swimfanfan 43948 Posts user info edit post |
i'm in a stat class this fall
i swear it's special ed or something 11/12/2010 2:43:49 PM |
DeltaBeta All American 9417 Posts user info edit post |
237 out of 299 people hate that shit. 11/12/2010 2:53:30 PM |
TerdFerguson All American 6584 Posts user info edit post |
i recommend you graff that shizzzz 11/12/2010 2:59:33 PM |
humandrive All American 18286 Posts user info edit post |
5/4 people have a problem with fractions 11/12/2010 9:11:34 PM |
aea All Amurican 5269 Posts user info edit post |
Quote : | "A = B B = C A != C " |
perfectly legitimate answer in statistics. seeing as statistics are just lies. damn lies.11/12/2010 9:16:41 PM |
Spontaneous All American 27372 Posts user info edit post |
I see all the obvious jokes have been made. Well done. 11/12/2010 9:21:36 PM |
AlaskanGrown I'm Randy 4693 Posts user info edit post |
Transitive property does not work in the real worlds? I never took a stats class, not sure how I avoided them but alas I was successful. The stats/probability section on the FE was easy as hell dough. 11/13/2010 11:11:39 AM |
Tarun almost 11687 Posts user info edit post |
Quote : | "should have paid attention" |
11/13/2010 11:14:21 AM |
McDanger All American 18835 Posts user info edit post |
whatcha workin on man
stats are easy at the undergraduate level but, as with most mathematics, the whole story is not easy 11/13/2010 1:55:44 PM |
wdprice3 BinaryBuffonary 45912 Posts user info edit post |
it was my research; not a class or anything. bacteria levels. twas a bitch 12/16/2010 8:09:19 PM |
merbig Suspended 13178 Posts user info edit post |
^ I think you got the wrong answer. 12/16/2010 8:54:55 PM |
JBaz All American 16764 Posts user info edit post |
zombie bacteria? 12/16/2010 9:10:02 PM |
neolithic All American 706 Posts user info edit post |
Maybe I'm getting trolled, but could you be a little more specific? Are those composite hypotheses? Independent ones? Contrasts/Comparisons? What is your model? What are your assumptions? What is your design?
It's hard to know what you are talking about with any detail. 12/17/2010 12:35:29 AM |
0EPII1 All American 42535 Posts user info edit post |
Quote : | "When I took statistics it was nothing more than a class to show proofs for things related to calculus. It sucked." |
what kind of stats was that? 12/17/2010 12:59:24 AM |
jbtilley All American 12796 Posts user info edit post |
^ST 371. We covered probabilities and the like in the first 2 or 3 classes. The rest of the semester was spent doing formal proofs of various calculus principles.
The professor was out quite often because he was an expert witness on a murder trial. In the trial he played the part of C3PO, calculating and communicating the probability of the DNA evidence at the crime scene not matching the suspect's DNA.
That class sucked. After the first test all remaining tests were about formal proofs. So other people taking stats with a different professor got to say the odds of picking a green jellybean were 20%. We had to show a proof for what would happen if people were constantly adding and removing green jellybeans in an environment where the rate of white jellybeans doubled every 10 minutes.... squared. A poor example, but it's been too long for me to remember. IIRC most proofs were related to calculus principles, not really related to statistics at all.
It's a class I want to forget have forgotten.
I guess it all depends on your professor. There was another class like that MA 305 "Introductory Linear Algebra and Matrices". Most people I knew got stuck doing proofs. I took it in the summer and it was really easy. Plug in the number, chug, get the answer. No proofs whatsoever.
[Edited on December 17, 2010 at 10:55 AM. Reason : -] 12/17/2010 10:41:34 AM |
Specter All American 6575 Posts user info edit post |
use the convolution integral. 12/17/2010 3:10:51 PM |
neolithic All American 706 Posts user info edit post |
Quote : | "^ST 371. We covered probabilities and the like in the first 2 or 3 classes. The rest of the semester was spent doing formal proofs of various calculus principles." |
I took ST 371 as an engineering undergrad too, but now I am a grad student in statistics. While it is true that a lot of results in statistics are a consequence of calculus (namely taylor series manipulations), I don't remember there being too much theorem proving in ST 371. Actually, in most of my 1st year theory courses, there wasn't all that much theorem proving either. Do you remember who your professor was? That could make a huge difference.
As to your example, you have to realize that those are probabilistic statements. So really what you are saying is that A is plausibly equal to B, B is plausibly equal to C, but A is not plausibly equal to C at this level test. It's not hard to imagine that the 95% CI for A and B overlap (so they are plausibly equal), B and Cs 95% CIs overlap (and so they are plausibly equal), but A and C's 95% CIs do not, thus they are not plausibly equal at the .05 level.
Transitivity does not hold for probabilistic hypotheses.12/18/2010 5:20:31 PM |
Fermat All American 47007 Posts user info edit post |
sounds more topology or set theory 12/18/2010 8:14:47 PM |
neolithic All American 706 Posts user info edit post |
Well, what he is talking about is an extension of set theory. Each hypothesis is a mapped to a rejection region, which is a set in R^k, where k is the number of hypotheses being tested. So this actually an extension of set theory, but the sets themselves are random. 12/19/2010 11:16:55 PM |
FAI756843 All American 908 Posts user info edit post |
i cant believe i understood what neolithic said
damn, i've become a nerd.
12/19/2010 11:19:13 PM |