CONFIDENTIAL Appendix 1(1) CS/STA408

CONFIDENCE INTERVALS

Parameter & description Two-sided (1 - )100% confidence interval

Mean,  , variance,  2 unknown,  s 

small samples
x  t 2, df   ; df = n − 1
 n

Difference in means, 1 − 2 ,
(x1 − x2 )  t 2, df s p 1 1
+ ;
variances 12 = 22 and unknown n1 n2

( n1 − 1) s12 + ( n2 − 1) s22
df = n1 + n2 − 2 , sp =
n1 + n2 − 2

Difference in means, 1 − 2 , 2 2

variances 12  22 and unknown (x1 − x2 )  t 2, df s1 s2

+ ;
n1 n2

df =
(s 1
2
n1 + s22 n2 )
2

(s
1
2
n1 ) (
2
s 2 n
+ 2 2
)
2

n1 − 1 n2 − 1

Mean difference for  s 

paired samples,  d d  t 2, df  d  ; df = n − 1 where n is no. of pairs
 n

Variance,  2  (n − 1) s 2 (n − 1) s 2 
 , 2 ; df = n − 1
 2 2, df  
 1− 2 , df 

Ratio of the variances 12 22 s2 1 s12 

 1 , F 2;  2 ,1  ; 1 = n1 − 1 ,  2 = n2 − 1
 s F 2 ;  ,
2
s2 2 
 2 1 2 

© Hak Cipta Universiti Teknologi MARA CONFIDENTIAL

CONFIDENTIAL Appendix 1(2) CS/STA408

HYPOTHESIS TESTING

Null Hypothesis Test statistic

x − 0
H0 :  =  0 t calc = ; df = n − 1
s
 2 unknown n
(x − x2 ) − (1 − 2 )
t calc = 1 ;
H0 : 1 −  2 = 0 1 1
sp +
12 = 22 and unknown n1 n2

( n1 − 1) s12 + ( n2 − 1) s22
df = n1 + n2 − 2 , sp =
n1 + n2 − 2

t calc =
(x1 − x2 ) − (1 − 2 ) ;
H0 : 1 −  2 = 0
s12 s22
+
variances 12  22 and unknown n1 n2

df =
(s1
2
n1 + s22 n2 )
2

(s 1
2
n1 ) (
2
s 2 n
+ 2 2
)
2

n1 − 1 n2 − 1
d − d
H0 :  d = 0 t calc = ; df = n − 1 where n is no. of pairs
sd
n

calc
2
=
(n − 1) s 2 ; df = n − 1
H0 :  = 0
2 2
0 2

s12
Fcalc = ; 1 = n1 − 1 ,  2 = n2 − 1
H0 : 12 = 2 2
s22

CORRELATION AND REGRESSION

SSxy
r =
SSxxSSyy

 xy −  n ,
Product Moment Correlation x y
where SSxy =
Coefficient

SSxx = x −
( x )
2
2

, SSyy
( y )
= y − M2
2

n un
Least-squares regression line of Y = a + bX k
Y against X SSxy a
where b = and a = y − b x
SSxx s
u
r
© Hak Cipta Universiti Teknologi MARA a CONFIDENTIAL
t