You are on page 1of 7

List of population parameter and corresponding statistic, standard error and marginal error.

Population Sample statistics Standard error statistic Marginal error Confidence interval Table to be
Standard Distribution Use when
parameter ( estimator) use

Population x−μ X is Normal or σ


Mean(µ) σ z∗σ z∗σ z= is known or
x ± x± σ Z -table
√n √n √n √n n ≥30

Population x−μ
Mean(µ) s s s t= n < 30 and or σ
x ±tn-1¿ x ± tn-1¿ s t-table
√n √n √n is unknown
√n
Population p−P n ^p ,n(1-

√ √ √
^p (1− ^p ) ^p (1− ^p) ^p (1− ^p ) z=


proportion (P) ^p ¿ ≥10
^p or p ± z∗ ^p ± z∗ PQ Z -table
n n n
n
Difference of two Both Normal
population mean ( x 1−x 2 )−( μ1 −μ 1 ) distribution or

√ √ √
2
σ1 σ2
2
σ1 σ2
2 2
σ1 σ2
2 2 z=


(µ1- µ2) ( x 1−x 2 ) + ± z∗ + ( x 1−x 2 ) ± z∗ n + n σ 21 σ 22 n1 , n2>= 30 or Z -table
n1 n2 n1 n2 1 2 +
n1 n2 σ1 , σ2 are
known

Difference of two ( x1 −x 2) − ( μ1−μ 1)

√ √ √
n1 , n2< 30 and
population mean
( n1−1 ) s21 + ( n2−1) s22 ( n1−1) s21 + ( n2−1) s22 ( n1 −1 ) s21 + ( t= 2
n2−1 ) s 2 2
( x 1−x 2 )

2
(µ1- µ2) n1+ n2−2
± t n +n −2∗
1 2
n1+ n2−2
( x 1−x 2 ) ±t n +n −2∗
1 2
n + n −2
( n 1−1 ) s1 + ( n2−1 ) s 2 σ1 , σ2 are t-table
1 2 unknown
n1 +n 2−2
Difference of two (P ¿ ¿1−P 2)
n ^p ,n(1-

√ √ √
^p1 (1−^p 1) ^p2 (1−^p 2) ^p1 (1−^p 1) ^p2 (1− ^p 2) z=(p
^p1 (1−^p 1) ^p2 (1−^p 2¿)¿ 1− p2 )− ¿¿


proportion
( ^p1− ^p2 ) + ± z∗ + ( ^p1− ^p2 ) ± z∗ + ^p 1 (1− ^p1 )^p ¿^p≥10
(1− ^
p ) Z -table
(P1- P2) n1 n2 n1 n2 n1 n2 + 2 2
n1 For each groupn2

Difference of two
proportion ^
PQ^ (1+1) When samples
n 1 n2 are drawn from
(P ¿ ¿ 1−P2) same

√ √
(P1- P2) 1 1 z=(p ¿ ¿ 1− p2 )− ¿¿
^ (1+1)

( ^p1− ^p2 ) where ± z∗ ^
PQ^ ( + ) ( ^p1− ^p2 ) ± z∗^P Q ^Q^ ( 1 population
1 Z -table
n1 n2 n 1 n2 P + )
^p 1 n1 + ^p2 2 n1 n2 Or
^
P=
n 1 +n2 P1= P2

Pair difference (di sd Data are taken


= yi − xi) and d = ±tn-1¿ sd
√n d before and
mean difference sd d ±tn-1¿ t= after in same
d √n sd t-table
√n n= Number of pair unit and paired
√n difference is
d =meanof pair difference normal

Selection of Hypothesis Test

σ n α and
Mean test population sample Hypothesis Test statistics critical Decision
sd. size value

Single mean Known No H0: µ = test x−μ Use z table for


mater of value z= α=5 %
Or one σ
sample
sample size
H1: µ ≠ test value
√n (For ≠) z
=±1.96
µ> test
value (For >) z
=+1.645
µ< test
value
(For <) z =-
1.645
H0: µ = test Use z table,
value α=5 %
H1: µ ≠ test value (For ≠) z
x−μ
µ> test z= =±1.96
value
σ^
Unknown n>= 30 (For >) z
µ< test
√n =+1.645
value
σ =σ^ =s (For <) z =-
1.645

H0: µ = test
value
Use t-table for
H1: µ ≠ test
x−μ
value t= (n-1) degree
Unknown n < 30 s freedom ,
µ> test
value √n given α and
H1 ≠, >,<
µ< test
value

Two sample σ1 & σ2 Known No H0: µ1 - µ2= 0 ( x 1−x 2 )−( μ1 −μ 2 ) Use z table for
and mater of z= α=5 %


(Difference of H1: µ1 - µ2≠ 0
sample σ 21 σ 22
two σ1 ≠ σ 2 size or + (For ≠) z
population sample
µ1 - µ2> 0 n1 n2 =±1.96
mean) size>30 µ1 - µ2< 0 (For >) z
=+1.645

(For <) z =-
1.645
Use z table for
α=5 %
Known No H0: µ1 - µ2= 0
mater of
σ1 = σ 2 ( x 1−x 2 )−( μ1 −μ 2 ) (For ≠) z
sample H1: µ1 - µ2 ≠ 0 z= =±1.96
Or from
same
population
size or

sample
size>30
µ1 - µ 2 > 0

µ1 - µ 2 < 0
σ/
√ 1 1
+
n1 n2
(For >) z
=+1.645

(For <) z =-
1.645

Use t table for


α=5 % and
H0: µ1 - µ2= 0 degree
Unknown ( x 1−x 2) −( μ1−μ 2) freedom (n1 +
H1: µ1 - µ2≠ 0 t=


n1 and n2 < 2 2 n1 - 2)
Different 30 s 1 s2
assumed
µ1 - µ2> 0 + (For ≠) t =±
n 1 n2
µ1 - µ2< 0 (For >) t =+

(For <) t =-

( x 1−x 2) −( μ1−μ 2) Use t table for


t= α=5 % and

Unknown

Equal
n1 and n2
H0: µ1 - µ2 = 0

H1: µ1 - µ2 ≠ 0
s
√ 1 1
+
n1 n2
degree
freedom (n1 +
n1 - 2)
< 30 µ1 - µ 2 > 0
assumed (For ≠) t =±


µ1 - µ 2 < 0 2
( s¿¿ 2−1)
(For >) t =+
s= n1(s ¿¿ 1−1)2 +n 2 ¿¿
n1+ n2−2
(For <) t =-

Single Binomial n large H0: P = test p−P Use z table for


z=


proportion distribution n>=30 value α=5 %
follow, PQ
H1: P ≠ test n (For ≠) z
N large, value =±1.96

np>=5 and P> test (For >) z


nq>=5
=+1.645
value
(For <) z =-
P< test 1.645
value

Use z table for


Two α=5 %
population
with n1 and H0: P1 - P2 = 0 (For ≠) z
Binomial n2 > 30 (P ¿ ¿1−P 2)
=±1.96
Test of two H1: P1 - P2 ≠ 0 z=(p ¿ ¿ 1− p2 )− ¿¿


distribution
population np>=5 ^p 1 (1− ^p1 ) (For
^p 2>) z ^
( 1− p2 )
proportion condition P1 - P 2 > 0 +=+1.645
and n1 n2
P1 & P2 nq>=5 P1 - P 2 < 0 (For <) z =-
different 1.645
assumed

Test of two Two n1 and n2 H0: P1 - P2 = 0 ( p ¿ ¿ 1− p2 ) Use z table for


z= ¿

√( (
population population > 30 α=5 %

))
H1: P1 - P2 ≠ 0
proportion with
np>=5 ^Q
P ^ 1+1 (For ≠) z
Binomial P1 - P 2 > 0 n1 n2
distribution and =±1.96
condition nq>=5 P1 - P 2 < 0 (For >) z
P1 & P2 equal =+1.645
different ^ p1 n1 + p 2 n 2
P= (For <) z =-
assumed, or n1 + n2 1.645
two samples
are drawn
from same
population ^
Q=1− ^
P
p1= x1/n1

p2=x2/n2
Useχ2table for
α=5 % and
degree
H0: V1 and V2
Two freedom (r-1)
are
categorical (c-1)
At least independent
variable V1 and v1
two r is number of
two H1: V1 and V2
Chi square category category in
categorical are dependent O is observed
test of in each row
variables
independenc variable
E is expected
r is number of
e
category in
column

H0: µ1 = µ2 = µ3
Use F table for
More than No H1: µ1 ≠ µ2≠ µ3
Population is α=5 % and
two variable mater
Normally at least one is degree
sample
ANOVA test distributed different freedom ν1
size
and ν2

You might also like