In City 1, a sample of 80 people in a profession have mean income 7.6 lakh, stdev 1.2 lakh.

In City 2, a sample of 60 people in the same profession have mean income 7 lakh, stdev 1 lakh.
Does City 1 pay more on average? Confidence 95%. Assume population variances are equal.

If the last line assuming equal variance was not there, we would have to do F-test for Variance first
to see if Variances are equal or unequal. But now we can move directly to pooled-variance test.

Samples from two different populations, testing the mean, variances equal; we use the pooled-variance test.

City 1: mean = xbar1 = 7.6 std dev = S1 = 1.2 size = n1= 80

City 2: mean = xbar2 = 7 std dev = S2 = 1 size = n2= 60

H0: City 1 <= City2 ----> City 1 - City 2 <= 0

H1: City 1 > City 2 ----> City 1 - City 2 > 0

Right tailed test, alpha = 1 - confidence = 0.05

Hypothesized difference = u1 - u2 = 0

df = n1 + n2 - 2 = 138

Sp^2 = ((n1 - 1)*(S1^2) + (n2 - 1)*(S2^2))/(n1-1+n2-1) = 1.251884

S.E. = ((Sp^2)*((1/n1)+(1/n2)))^0.5 = 0.191084

tstat = ((x1bar - x2bar) - (u1 - u2))/S.E. = 3.139972

|t(0.05,df)| = abs(t.inv(0.05,138)) = 1.65597

Right-tailed test, since tstat = 3.14 > |t(0.05,138)| = 1.656, reject H0.
So, City 1 does pay more than City 2 on average.
In City 1, a sample of 80 people in a profession have mean income 7.6 lakh, stdev 1.2 lakh.
In City 2, a sample of 60 people in the same profession have mean income 7 lakh, stdev 1 lakh.
Does City 1 pay more on average? Confidence 95%. Assume population variances are equal.

n1= 80 n2= 60
df1= 79 df2= 59
mean1= 7.6 mean2= 7
SD1= 1.2 SD2= 1
var1= 1.44 var2= 1
alpha= 0.05

F TEST two tailed test

H0= variances are equal
H1= variances are unequal

F(stat)= 1.44
F(alpha,df1 1.506748

df1*var1 + df2*var2 /
df1+df2
Rejecting H0 since F(stat) > F(alpha,df1,df2)

Using Separate Variance Test / Pooled Variance test

H0: c1-c2<=0
H1: c1-c2>0
right tailed test

SE= 1.251884 172.76 138

T(stat)= 3.139972 0.6 0.191084
df= 138
T(alpha,df) 1.65597
, stdev 1.2 lakh.
7 lakh, stdev 1 lakh.
iances are equal.

ar1 + df2*var2 /
A sample of 40 men have average height 171cm, stdev 16.2cm.
A sample of 50 women have average height 164cm, stdev 10.1cm.
Are men on average 10 cm taller than women? Alpha = 0.1. Assume unequal variances.

Samples from two populations, testing means, variances unequal; use separate-variance test.

Men mean = xbar1 = 171 std dev = S1 = 16.2 size = n1= 40

Women mean = xbar2 = 164 std dev = S2 = 10.1 size = n2= 50

H0: men - women = 10

H1: men - women != 10

Two-tailed test, alpha = 0.1

Hypothesized difference = u1 - u2 = 10

Let a = (S1^2)/n1 = 6.561 Let b = (S2^2)/n2 = 2.0402

Defining 'a' and 'b' is not necessary, we are just using it as a way to make the calculations simpler
df = ((a+b)^2) / ((a^2)/(n1-1) + (b^2)/(n2-1)) =

S.E. = (a+b)^0.5 = 2.93278

tstat = ((x1bar - x2bar) - (u1 - u2))/S.E. = -1.02292

+\- t(alpha/2,df) = +\- t.inv(0.05,df) = -1.669804 to 1.6698

Two-tailed test, since tstat = -1.023 is between +\- 1.67, we accept H0.
Thus, men are 10cm taller than women.
A sample of 40 men have average height 171cm, stdev 16.2cm.
A sample of 50 women have average height 164cm, stdev 10.1cm.
Are men on average 10 cm taller than women? Alpha = 0.1. Assume unequal variances.

n1= 40 n2= 50
df1= 39 df2= 49
mean1= 171 mean2= 164
SD1= 16.2 SD2= 10.1
var1= 262.44 var2= 102.01

F TEST two tailed test -3

H0= variances are equal 8.6012
H1= variances are unequal

F(stat)= higher variance / lower variance 2.57268895206352

F(alpha,df1,df2)= F.INV.RT(alpha,df1,df2) 1.47107451146256

Rejecting H0 since F(stat) > F(alpha,df1,df2)

T TEST
H0: mean1-mean2<=0
H1: mean1-mean2>0

T(stat)= -1.02292014517331
df= 62.2361066418964
T(alpha,df)= T.INV(alpha,df) -1.29535576176057

Rejecting H0 since t(stat) >=< t(alpha,df)

ual variances.

2.93278

73.98064
On Road 1, a sample of 213 cars travel at an average speed of 77.2 kmph, stdev 15.4.
On Road 2, a sample of 246 cars travel at an average speed of 74.4 kmph, stdev 10.8.
Is the variance of car speed the same on both roads? Alpha = 0.03.

Samples from two populations, testing the variance, we use the F-test for Variances. (L5 ppt slide 5)

Road 1: mean = xbar1 = 77.2 stdev = S1 = 15.4 size= n1 = 213

Road 2: mean = xbar2 = 74.4 stdev = S2 = 10.8 size= n2 = 246

H0: Population variances are equal

H1: Population variances are not equal

alpha = 0.03

Since S1 > S2, Fstat = Larger variance / smaller variance = (S1^2)/(S2^2) = 2.033265

Since S1 > S2, df1 = n1 - 1 = 212 If S2 had been greater than S1, then
df2 = n2 - 1 = 245 Fstat = (S2^2)/(S1^2) and df1 = n2 - 1; df2 = n1 - 1.

F(0.03,df1,df2) = f.inv.rt(0.03,212,245) = 1.282639

Since Fstat = 2.03 > 1.28, reject H0. Thus, variances are not equal.

Is the average car speed of Road 1 higher than Road 2? Alpha = 0.03.

Sample from two populations, testing for mean, variances unequal (proven above), so we
use the separate-variance test.

H0: Road 1 - Road 2 <= 0

H1: Road 1 - Road 2 > 0

Right-tailed test, alpha = 0.03

Hypothesized difference = u1 - u2 = 0

Let a = (S1^2)/n1 = 1.113427

Let b = (S2^2)/n2 = 0.474146

df = ((a+b)^2) / ((a^2)/(n1-1) + (b^2)/(n2-1)) = 372.544

S.E. = (a+b)^0.5 = 1.25999

tstat = ((x1bar - x2bar) - (u1 - u2))/S.E. = 2.222241

|t(0.03,df)| = abs(t.inv(0.03,df)) = 1.886546

Right-tailed test, since tstat = 2.22 > |t(0.03,df)| = 1.89, we reject H0. So, mean speed of Road 1 > Road 2.
On Road 1, a sample of 213 cars travel at an average speed of 77.2 kmph, stdev 15.4.
On Road 2, a sample of 246 cars travel at an average speed of 74.4 kmph, stdev 10.8.
Is the variance of car speed the same on both roads? Alpha = 0.03.

n1= 213 n2= 246

mean1= 77.2 mean2= 74.4
SD1= 15.4 SD2= 10.8
var1= 237.16 var2= 116.64

F test two tailed test

H0= variances are equal
H1= variances are unequal

F(stat)= 2.033265
F(alpha,df1,df2)= 1.282639
a= 1.113427
rejecting H0, variances are unequal b= 0.474146

num= 2.52039
T test
den= 0.006765
H0= r1-r2=0
H1 r1-r2=!0

df= 372.544
T(stat)= 2.222241
T(alpha,df)= 1.886546

reject H0
not equal
7.2 kmph, stdev 15.4. On Road 1, a sample of 213 cars travel at an average speed of 77.2
4.4 kmph, stdev 10.8. On Road 2, a sample of 246 cars travel at an average speed of 74.4
Is the variance of car speed the same on both roads? Alpha = 0.03.

n1= 213 n2=

df1= 212 df2=
mean1= 77.2 mean2=
SD1= 15.4 SD2=
var1= 237.16 var2=

F TEST two tailed test

H0= variances are equal
H1= variances are unequal

F(stat)= higher variance / lower variance

F(alpha,df1,df2)= F.INV.RT(alpha,df1,df2)

Rejecting H0 since F(stat) > F(alpha,df1,df2)

T TEST
H0:
H1:

T(stat)= ref
df= ref
T(alpha,df)= T.INV(alpha,df)

Rejecting H0 since t(stat) >=< t(alpha,df)

an average speed of 77.2 kmph, stdev 15.4. On Road 1, a sample of 213 cars travel at an average speed of 77.2
an average speed of 74.4 kmph, stdev 10.8. On Road 2, a sample of 246 cars travel at an average speed of 74.4
oth roads? Alpha = 0.03. Is the variance of car speed the same on both roads? Alpha = 0.03.

246 n1= 213 n2= 246

245 df1= 212 df2= 245
74.4 mean1= 77.2 mean2= 74.4
10.8 SD1= 15.4 SD2= 10.8
116.64 var1= 237.16 var2= 116.64
alpha= 0.03

F TEST two tailed test

H0= variances are equal
H1= variances are unequal
er variance
F(stat)= 2.033265
F(alpha,df1 1.282639

Rejecting H0 since F(stat) > F(alpha,df1,df2)

Using Separate Variance Test / Pooled Variance test

H0: 1-2=0
H1: 1-2!=0
right / left / two tailed test

SE= ref
T(stat)= 2.222241 2.8 1.25999
df= 372.544 2.52039 0.006765
T(alpha,df) 2.178448

rejecting / accepting H0 since t(stat) >=< t(alpha,df)

an average speed of 77.2 kmph, stdev 15.4.
an average speed of 74.4 kmph, stdev 10.8.
oth roads? Alpha = 0.03.

NO SE REQUIRED

Variance test

< t(alpha,df)
In a sample of 243 people from City 1, 67 say they regularly drink cola.
In a sample of 306 people from City 2, 52 say they regularly drink cola.
Is the proportion of cola drinkers in City 1 more than 5% higher than that of City 2? Alpha = 0.01.

Samples from two populations, testing the proportion, Z-test for difference in proportion L5 slide 7

City 1: Size= n1 = 243 Cola drinkers = x1 = 67 p1 = x1/n1 = 0.27572

City 2: Size= n2 = 306 Cola drinkers = x2 = 52 p2 = x2/n2 = 0.169935

H0: p(City 1) - p(City 2) <= 0.05

H1: p(City 1) - p(City 2) > 0.05

Right-tailed test of proportion, alpha = 0.01

Hypothesized difference = (pi1 - pi2) = 0.05

pbar = (x1 + x2)/(n1 + n2) = 0.216758

SEp = (pbar*(1-pbar)*((1/n1)+(1/n2)))^0.5 = 0.035404 (denominator of Zstat)

Zstat = ((p1 - p2) - (pi1 - pi2))/SEp = 1.575665

|Z(0.01)| = abs(norm.s.inv(0.01)) = 2.326348

Right-tailed test, Zstat = 1.576 < |Z(0.01)| = 2.326, so we accept H0.

Thus, Cola drinker proportion of City 1 - City 2 <= 5%.
In a sample of 243 people from City 1, 67 say they regularly drink cola.
In a sample of 306 people from City 2, 52 say they regularly drink cola.
Is the proportion of cola drinkers in City 1 more than 5% higher than that of City 2? Alpha =

x1= 67 x2= 52
n1= 243 n2= 306
p1= 0.27572 p2= 0.169935

confidence level=
alpha= 0.01

H0: p1-p2<=0.05
H1: p1-p2>0.05
right tailed test

p(bar)= 0.216758 denominator

Z(stat) 1.575665 0.055786 0.035404
Z(alpha)= 2.326348
than that of City 2? Alpha = 0.01.