[R.app GUI 1.43 (5989) x86_64-apple-darwin9.8.0]

>library(outliers)

> signs
> shapiro.test(signs)

Shapiro-Wilk normality test

data: signs
W = 0.9509, p-value = 0.6502


Normally distributed.



> mean(signs)
[1] 166.0833
> sd(signs)
[1] 21.13359

> dnorm(209, mean=166, sd=21.13)
[1] 0.002380873
> dnorm(139, mean=166, sd=21.13)
[1] 0.008345574


On the tails of their own distribution.

Are they really outliers?



> grubbs.test(signs, type=11)

Grubbs test for two opposite outliers

data: signs
G = 3.3123, U = 0.4707, p-value = 0.6774
alternative hypothesis: 139 and 209 are outliers


Nope.

They don't appear to be extraordinary.

DO NOT WANT Poster Girl

Have you ever tried throwing your library at traffic accident statistics? Traffic accident stats are overdispersed. That leads to a high rate of claims that changes made things better/worse when the truth is, humans are just really weird.

That's because Aries was the god of war. If you know that, and everyone tells you "You are an Aries." eventually you'll get combative.

KkrrRRZZZZ: That's because Aries was the god of war. If you know that, and everyone tells you "You are an Aries." eventually you'll get combative.

I just bought a Ram GOD OF WAR truck, good truck.

It gets 100 dead Trojans to the gallon of booze.

I'm disappointed with the Scorpios. We're supposed to be the scum of the zodiac.

Mugato: I'm disappointed with the Scorpios. We're supposed to be the scum of the zodiac.

Geminis are supposed to be "social butterflies" and have many friends and enjoy parties...OH yeah, I'll be social...alone, in front of a computer, typing to other computer bound introverts....

Macular Degenerate: As a graduate student... I will bet a paycheck

You reminded me of an episode of Cheers where two people make a bet of a week's pay, and the loser turned out to make $1 per year.

DO NOT WANT Poster Girl: [R.app GUI 1.43 (5989) x86_64-apple-darwin9.8.0]

>library(outliers)

> signs
> shapiro.test(signs)

Shapiro-Wilk normality test

data: signs
W = 0.9509, p-value = 0.6502


Normally distributed.



> mean(signs)
[1] 166.0833
> sd(signs)
[1] 21.13359

> dnorm(209, mean=166, sd=21.13)
[1] 0.002380873
> dnorm(139, mean=166, sd=21.13)
[1] 0.008345574


On the tails of their own distribution.

Are they really outliers?



> grubbs.test(signs, type=11)

Grubbs test for two opposite outliers

data: signs
G = 3.3123, U = 0.4707, p-value = 0.6774
alternative hypothesis: 139 and 209 are outliers


Nope.

They don't appear to be extraordinary.

I was gonna come in and do that stats if someone else didn't get to it first.
Junk farking science again.

Thank you for your hard work.

-R

The age you start school depends on your birthday. You could start 1st grade at 6 years and 9 months or 5 years and 9 months if your birthday is on January 1 or December 31. Many things could follow from that. Are you the biggest guy in your class or the smallest?

Even without fancy statistical analysis, experience tells me that these studies never match--the relationship of astrological signs to whatever is being measures--car accidents, tickets, crime, etc.--varies from one set of data to another. In other words, variation is normal and essentially random. There might be causes of bias in the original birth statistics--for example seasonal variations in the weather, dates which doctors and parents avoid for superstitious reasons or because of holidays, and so forth--but in the end, you'd need to do some sophisticated weighting to eliminate all these causes of variation.

In short, no two studies ever match--the police in another jurisdiction could do the same to their data and different signs would come out on top, on bottom or in the middle.

I did a "what star sign do you most ressemble?" test and found four birth signs that I ressembled more than my own, four birth signs I ressembled less and three star signs that were middling. In other words, randomness governs your relationship to your birth sign. I may have cheated a tiny bit on a few questions, but all in all, you shouldn't find such a high level of randomness if the birth sign really meant anything statistically significant.

That said, this particular study shows a good correlation to the "laid back" star signs and low arrest levels. Others would not. I'm not saying that your birth date doesn't have an effect on your life, personality and behaviour, just that this type of study is the wrong way to prove it, and in any case, it wouldn't have anything to do with the Zodiac if it does work. The Zodiac is an arbitrary interpretation made by connecting the dots (stars) in the sky and then creating a myth about the "image" that you create. These images are totally made up and very low resolution--a couple of dots is a fish or a crab? Not in my colouring and activity book.

As I always say about UFO photographs, fuzzy dots prove nothing!

steveGswine: DO NOT WANT Poster Girl: [R.app GUI 1.43 (5989) x86_64-apple-darwin9.8.0]

I really, really have to start playing with R.

Point me in a direction?

There are a lot of online tutorials, and specfic books written for learning stats and R - the big blue "R BOOK" is OK if you want a cookbook of things you can do with R code, but there's nothing really as good as having a problem and using R to figure it out.

Most important recommendation I can make to anyone who's used other programming languages and R is that there are only procedural similarities in the program (like loops and booleans etc), but the data structures underneath are very different than standard programming/scripting languages. For instance, there are no single characters or numbers in R, only vectors of length 1.

In R, the data type called "dataframe" is (like MS Excel), a collection of vectors (columns) of mixed type but identical length. There's another type, the matrix, which is not a data frame -- matrices are arrays of a single data type.

So -- find tutorials that start you out on the data types and don't skip the early chapters/examples. You can't understand R without understanding the underlying data models. After that it's really easy stuff.

weiner dog: I'm one of the Pisces/Aries cusp people, but more on the Aries side, thus I have very little in common with my western astrological assignment. I decided to go by the Chinese chart (dog) since it's more me-ish.

/css

Any astrology is still retarded.
Think about it. Everyone born between date x and date y are automatically "vicious executive types bent on world domination". Right.

The Chinese one ascribes entire years' worth of children particular traits.