Fixed n.

n! a ! b !n! !... c
Coded data

Table s

Binomial distribution B(n, p)

Geometric distribution Geo(p) Assumptions: independence, constant probability, 2 outcomes.

P( =

x 1

Think intuitively Permutations

P( X=
Combination s Order unimportant

V a[ rx =X x 2 [E X ] ] n n x) x= C p q
Special distributions

Order important

Raw data

s.d =Data in tables n

Mean and standard deviation

n! CrSum of is 1. = probabilities E[ Xr]!(n r )! x =

Permutations 2 and combinations

Discrete random variables Tree diagrams

x x=
Median and quartiles

Variance is the square of s.d.

Numerical statistics

fx x=
Probability

Conditional probability

Stem-and-leaf diagrams

Statistics 1 Mind map

Scatter graphs

where d is the difference in ranks.

Measures tendency for x and y to increase or decrease together.

Box plots

Spearmans Rank

s. d =

x
n

Graphical statistics

2
Histograms

-1 r 1 r = 1: perfect straight line (+ve gradient) r = -1: perfect straight line (-ve gradient)

Correlation

P(A and B) P( A | B) = the strength Measures of a linear P( B) relationship.

Product moment correlation coefficient

Cumulative frequency graphs

Linear regression

Plot points at end of each interrval Used to estimate the median and quartiles for grouped data

Variables can be random or controlled. Use if we want to work out x when given y. x on y: x = a + by where and

Two possible regression lines Use if we want to find y OR if x is a controlled variable. y on x: y = a + bx where and

If the last interval is open-ended, you choose its width to be twice the width of the previous interval.

Predictions will be accurate provided the points lie close to a straight line and provided we are making predictions within the range of data collected.

Both regression lines pass through the mean point

The PMCC is unaffected by linear transformations of the variables.

A Duncombe

frequency freq density = class w idth

r ==

Sx y Sx x