DHS Distribution Summary Updated Oct 2013
DHS Distribution Summary Updated Oct 2013
DHS Distribution Summary Updated Oct 2013
2013 / YR 6
Distribution BINOMIAL POISSON NORMAL UNKNOWN
DECIDE ON THE APPROPRIATE DISTRIBUTION
1) Discrete 1) Discrete
1) Continuous
2) “Number of success out of n 2) “Number of occurrences in fixed Distribution
2) Stated in question, e.g
Clues trials” length/period not stated in
- follows a normal distribution
3) Exists an upper bound n for 3) No upper bound for number of question
- is normally distributed
number of success occurrences
DEFINE & STATE DISTRIBUTION
Define Random
Variable Let X be the number of success out Let X be the number of occurrences in Let X be the attribute of a randomly
of n trials. fixed length/period. chosen object/person.
State
X ~ B(n, p) X ~ Po( )
Distribution X ~ N( , 2 )
where n is the number of trials & where is the mean occurences in
where is the mean of X and
p is the probability of success of fixed length/period.
1 trial. 2 is the variance of X .
E( X ) = np E( X ) =
Properties E( X ) = Var( X ) =
Var( X ) = np (1 − p ) Var( X ) = 2
Distribution BINOMIAL POISSON NORMAL UNKNOWN
FIND THE REQUIRED DISTRIBUTION OR USE A SUITABLE APPROXIMATION
For ANY distribution, the following results holds:
Key words: Suppose E( X ) = & Var( X ) = 2 . If sample size n is large (≥ 50), by CLT,
2 Use CLT if
Mean, 2 X ~ N , regardless of sample size n sample size n
X ~ N , approximately n
Average n is large
X + ... + X n
X= 1
n
Sample Sum
Suppose E( X ) = & Var( X ) = 2 . If sample size n is large (≥ 50), by CLT, Use CLT if
Key words:
Sum, Total ( )
S ~ N n , n 2 approximately
(
S ~ N n , n 2 ) regardless of sample size n sample size n
is large
S = X1 + ... + X n
EVALUATE
Strategies
1) Standardise if either or 2 is
Note: Use c.c when applicable Note: Use c.c when applicable unknown.
2) Draw the normal curve when using
invnorm.