Welcome to Scribd, the world's digital library. Read, publish, and share books and documents. See more
Standard view
Full view
of .
0 of .
Results for:
P. 1
statistical formulas

statistical formulas

Ratings: (0)|Views: 2|Likes:

Categories:Types, Research

Availability:

See more
See less

08/09/2014

pdf

text

original

Statistics formula sheet
Summarising data
Sample mean:
x
=1
n
n
i
=1
x
i
.
Sample variance:
s
2
x
=1
n
1
n
i
=1
(
x
i
x
)
2
=1
n
1
n
i
=1
x
2
i
nx
2
.
Sample covariance:
g
=1
n
1
n
i
=1
(
x
i
x
)(
y
i
y
) =1
n
1
n
i
=1
x
i
y
i
nxy
.
Sample correlation:
r
=
gs
x
s
y
.
Probability
(
A
B
) =
(
A
) +
(
B
)
(
A
B
)
.
Multiplication law:
(
A
B
) =
(
A
)
(
B
|
A
) =
(
B
)
(
A
|
B
)
.
Partition law: For a partition
B
1
,B
2
,...,B
k
(
A
) =
k
i
=1
(
A
B
i
) =
k
i
=1
(
A
|
B
i
)
(
B
i
)
.
Bayes’ formula:
(
B
i
|
A
) =
(
A
|
B
i
)
(
B
i
)
(
A
)=
(
A
|
B
i
)
(
B
i
)
ki
=1
(
A
|
B
i
)
(
B
i
)
.
Discrete distributions
Mean value:
(
X
) =
µ
=
x
i
S
x
i
p
(
x
i
)
.
Variance:Var(
X
) =
x
i
S
(
x
i
µ
)
2
p
(
x
i
) =
x
i
S
x
2
i
p
(
x
i
)
µ
2
.
The binomial distribution:
p
(
x
) =
nx
θ
x
(1
θ
)
n
x
for
x
= 0
,
1
,...,n.
This has mean
and variance
(1
θ
).The Poisson distribution:
p
(
x
) =
λ
x
exp(
λ
)
x
!for
x
= 0
,
1
,
2
,....
This has mean
λ
and variance
λ
.
Continuous distributions
Distribution function:
(
y
) =
(
X
y
) =

y
−∞
(
x
)d
x.
Density function:
(
x
) =dd
x
(
x
)
.
Evaluating probabilities:
(
a < X
b
) =

ba
(
x
)d
x
=
(
b
)
(
a
)
.
Expected value:
(
X
) =
µ
=

−∞
xf
(
x
)d
x.
Variance:Var(
X
) =

−∞
(
x
µ
)
2
(
x
)d
x
=

−∞
x
2
(
x
)d
x
µ
2
.
Hazard function:
h
(
t
) =
(
t
)1
(
t
)
.
Normal density with mean
µ
and variance
σ
2
:
(
x
) =1
2
πσ
2
exp
12
x
µσ
2
for
x
[
−∞
,
]
.
Weibull density:
(
t
) =
λκt
κ
1
exp(
λt
κ
) for
t
0
.
Exponential density:
(
t
) =
λ
exp(
λt
) for
t
0
.
This has mean
λ
1
and variance
λ
2
.
Test for population mean
Data:
Single sample of measurements
x
1
,...,x
n
.
Hypothesis:
:
µ
=
µ
0
.
Method:
Calculate
x
,
s
2
, and
t
=
|
x
µ
0
|
n/s
.
Obtain critical value from
t
-tables,
d
=
n
1.

Reject
at the 100
p
% level of signiﬁcance if
|
t
|
> c
,where
c
is the tabulated value corresponding to col-umn
p
.
Paired sample
t
-test
Data:
Single sample of
n
measurements
x
1
,...,x
n
whichare the pairwise diﬀerences between the two original setsof measurements.
Hypothesis:
:
µ
= 0.
Method:
Calculate
x
,
s
2
and
t
=
x
n/s
.
Obtain critical value from
t
-tables,
d
=
n
1.
Reject
at the 100
p
% level of signiﬁcance if
|
t
|
> c
,where
c
is the tabulated value corresponding to col-umn
p
.
Two sample
t
-test
Data:
Two separate samples of measurements
x
1
,...,x
n
and
y
1
,...,y
m
.
Hypothesis:
:
µ
x
=
µ
y
.
Method:
Calculate
x
,
s
2
x
,
y
, and
s
2
y
.
Calculate
s
2
=
(
n
1)
s
2
x
+ (
m
1)
s
2
y
/
(
n
+
m
2)
.
Calculate
t
=
x
y

s
2
1
n
+1
m
.
Obtain critical value from
t
-tables,
d
=
n
+
m
2.
Reject
at the 100
p
% level of signiﬁcance if
|
t
|
> c
,where
c
is the tabulated value corresponding to col-umn
p
.
CI for population mean
Data:
Sample of measurements
x
1
,...,x
n
.
Method:
Calculate
x
,
s
2
x
.
Look in
t
-tables,
d
=
n
1, column
p
. Let thetabulated value be
c
say.
100(1
p
)% conﬁdence interval for
µ
is
x
±
cs
x
/
n
.
CI for diﬀerence in population means
Data:
Separate samples
x
1
,...,x
n
and
y
1
,...,y
m
.
Method:
Calculate
x
,
s
2
x
,
y
,
s
2
y
.
Calculate
s
2
=
(
n
1)
s
2
x
+ (
m
1)
s
2
y
/
(
n
+
m
2)
.
Look in
t
-tables,
d
=
n
+
m
2, column
p
. Let thetabulated value be
c
say.
100(1
p
)% conﬁdence interval for the diﬀerence in
population
means i.e.
µ
x
µ
y
, is(
x
y
)
±
c

s
2
1
n
+1
m
.
Regression and correlation
The linear regression model:
y
i
=
α
+
βx
i
+
z
i
.
Least squares estimates of
α
and
β
:ˆ
β
=
ni
=1
x
i
y
i
nxy
(
n
1)
s
2
x
,
and ˆ
α
=
y
ˆ
βx.
Conﬁdence interval for
β
Calculateˆ
β
as given previously.
Calculate
s
2
ε
=
s
2
y
ˆ
β
2
s
2
x
.
Calculate
SE
(ˆ
β
) =

s
2
ε
(
n
2)
s
2
x
.
Look in
t
-tables,
d
=
n
2, column
p
. Let thetabulated value be
c
.
100(1
p
)% conﬁdence interval for
β
isˆ
β
±
cSE
(ˆ
β
).
Test for
ρ
= 0
Hypothesis:
:
ρ
= 0.
Calculate
t
=
r
n
21
r
2
1
/
2
.
Obtain critical value from
t
-tables,
d
=
n
2.
Reject
at 100
p
% level of signiﬁcance if
|
t
|
> c
,where
c
is the tabulated value corresponding to col-umn
p
.
Approximate CI for proportion
θ
p
±
1
.
96

p
(1
p
)
n
1where
p
is the observed proportion in the sample.