skewed maths
How did you get to that conclussion as you still have to test your null hypothesis?
(1) In statistical correlation, you need 2 sets of values which we will call
(x), (y)
In laymans language, this should translate the number of children born during (Moi, Kibaki) (Moi, Uhuru) (Kibaki, Uhuru).
(2) The sum Σ= (x)+(y) [SIZE=1]Example Σ=20 when x=10 and y=10[/SIZE]
Then you will need Variance (S) for both (x) & (y)
for (x) => S[SIZE=1]xx[/SIZE]
for (y) => S[SIZE=1]yy[/SIZE]
[SIZE=4]Variance is the square of the standard deviation, the second central moment of a distribution, and the CoVariance of the random variable with itself.
ie (Moi) Sqr of babies born in year n= 1, 2,3,…24
(Kibaki) sqr ----"----------year 1, 2, 3, 4,…10 etc[/SIZE]
SIZE=4 ρ= sample in a population [/SIZE]
Now we should compare the 2 variances and compute to find
CoVariance S[SIZE=1]xy[/SIZE]
In probability theory and statistics, covariance is a measure of how
much two random variables change together.
A need to further establish the respective mean values of both these X and Y-values is necessary at this juncture.
Validity of the correlation
Now, at this juncture all the necessary values are present to go ahead and examine if there exists
a linear correlation. A linear correlation coefficient, measures the strength and the direction of a linear relationship between two variables. It is also necessary to define a correlation coefficient
The -1 < r < +1 are the signs used for positive and negative linear.
S[SIZE=1]xy[/SIZE]
r = ------------
SQR(S[SIZE=1]xx[/SIZE] * S[SIZE=1]yy[/SIZE])
If r =1 then the correlation is very strong!!
(it’s easier to get Safaricom bundles in Jupiter than getting r=1)
The t-value for null hypothesis
r - [SIZE=4] ρ
t= ------------[/SIZE]
SQR[SIZE=4](1- r^[/SIZE][SIZE=1]2[/SIZE][SIZE=4] / n-2) r to the power of 2[/SIZE]
Therein, one can answer if there is an apparent correlation or the existence of such is simply a coincidence!
Examination if the correlation coefficient obtained is different from 0 gives.
For the null-hypothesis
H[SIZE=1]0[/SIZE][SIZE=4] : ρ=[/SIZE]0
Much much later, thid null hyoothesid can further be tested by examining if the trend line/regression line
Yr=b[SIZE=1]0[/SIZE] + b[SIZE=1]1[/SIZE]x
b[SIZE=1]1[/SIZE]= the slope
S[SIZE=1]xy[/SIZE]
b[SIZE=1]1 [/SIZE]= -------
S[SIZE=1]xx[/SIZE]
tuendelee baadaye ukielewa hizi babaa