Scribd
Upload
38 views3 pages

Some Basic Formulae For Use in ST205: The Expectation

This document provides formulae for calculating expectations, variances, and covariances of random variables that will be useful for an exam on statistics. It explains that the expectation of a sum is the sum of the individual expectations, and the expectation of a constant times a random variable is that constant times the expectation. It also describes that the variance of a sum is not generally equal to the sum of the variances, and provides the formula that includes the covariances between variables. It gives additional formulae for variances, covariances, and how multiplying a variable by a constant affects its variance. Everything in the note should be well understood for the exam.

Uploaded by

Cartie
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
38 views3 pages

Some Basic Formulae For Use in ST205: The Expectation

This document provides formulae for calculating expectations, variances, and covariances of random variables that will be useful for an exam on statistics. It explains that the expectation of a sum is the sum of the individual expectations, and the expectation of a constant times a random variable is that constant times the expectation. It also describes that the variance of a sum is not generally equal to the sum of the variances, and provides the formula that includes the covariances between variables. It gives additional formulae for variances, covariances, and how multiplying a variable by a constant affects its variance. Everything in the note should be well understood for the exam.

Uploaded by

Cartie
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
You are on page 1/ 3

Some basic formulae for use in ST205

This note is meant to be helpful in understanding the derivations given in the


lecture notes. Everything in this note should be well understood for the exam.
Below, y1, . . . , yn are arbitrary random variables. Remember that the yi’s can
basically be anything: in the lecture notes, yi usually represents the observed Y
value for the ith sampling unit. But we may also consider the means for the ith
sample, usually denoted in this course as yi. The rules below also hold for such
sample means, or any other statistics (usually estimators in this course).

The expectation

Firstly, the expectation of a sum is the sum of the expectations:

E(y1 + y2) = Ey1 + Ey2

Note that by repeatedly applying this formula we can deduce

Secondly, the expectation of a constant times a random variable equals the


constant times the expectation:

E(cy1) = cEy1

Thus, expectations are relatively easy to work with, the rules for variances are
somewhat more involved.

The variance
Firstly, the variance of a sum is generally not equal to the sum of the variances:

var(y1 + y2) = var(y1) + var(y2) + 2cov(y1,y2)

The variance of a difference is as follows

var(y1 − y2) = var(y1) + var(y2) − 2cov(y1,y2)

A relation between the variance and the covariance is:

var(y1) = cov(y1,y1)

That is, the variance of a variable is the covariance of the variable with itself.
For the covariance, we have, in a way, simpler calculation rules than for the
variance: cov(y1,y2 + y3) = cov(y1,y2) + cov(y1,y3)
1

and of course cov(y2 + y3,y1) = cov(y1,y2) + cov(y1,y3)


This formula explains the above rule for the variance of a sum, since
var(y1 + y2) = cov(y1 + y2,y1 + y2)
= cov(y1 + y2,y1) + cov(y1 + y2,y2)
= cov(y1,y1) + cov(y1,y2) + cov(y2,y1) + cov(y2,y2)
= var(y1) + cov(y1,y2) + cov(y2,y1) + var(y2)
= var(y1) + var(y2) + 2cov(y1,y2)
So studying this derivation we see that the formula for the variance of a sum
follows a sound logic.
By repeatedly applying the above formulae for the variance and covariance
we can deduce

This formula can be written, perhaps slightly less elegantly, but useful for some
purposes, as

Verify what happens if n = 2!


Note that the variance of a sum equals the sum of the variances only if the
covariances between all pairs of different variables equal zero. This occurs, for
example, in stratified sampling for which

where Tˆk is the estimated total in stratum k. The reason is that Tˆk is
independent of Tˆl for all k 6= l (so their covariance equals zero).
Finally, we see what happens to the variance if we multiply a random
variable by a constant c:

var(c y1) = c2 var(y1)

For the covariance, with constants a and b, we have

cov(a y1, b y2) = a b cov(y1, y2)


Verify that the formula for var(c y1) follows from this!

You might also like