cREVIEW FOR FINAL EXAM

Checklist of points to cover in the final:

1. Binomial distribution
- Formula to find the probability in the binomial distribution
X~ Bin(n,p), n is number of trials, p= P(Success), 1-p=P(failure)
Pn(X=x) = C(n,x)*px (1-p)n-x
E(X) =np; V(X)= np(1-p); sd(X)=sqrt V(X)
2. Normal distribution: X~N(µ,σ2) à E(X)=µ; V(X)= σ2
2 types of problems:
- Given the value, find the probability (P(X >a)
- Given probability, find the value, P(X <x0) =α
Standardization: X à Z: Z= (X-µ)/σ
3. Confidence Interval for p: given confidence level: 1-α

ME=E=z α/ 2
√ p(1− p)
n
Sample size determination:
2
(z α / 2) p (1− p)
n= 2
z0.025 =1.96, z0.05 =1.645; z0.01= 2.33
ME

4. CI for mean: µ

Sample size:
2 2
(z α / 2) σ
n= 2
ME

5. Test to compare the mean, µ with µ0: σ2 unknown

Choose 1 of 3 possible hypothesis: two tail, Right tail or left tail
x−μ0
Calculate t=
s/√n
When to reject Ho:

Two tail: Reject Ho if t > t(n-1)α/2 or t <- t(n-1)α/2

Right tail: Reject Ho if t > t(n-1)α
Left tail: Reject Ho if t < - t(n-1)α
 p-value: in case of using t –test.
Example: pvalue in Right tail test: P(T > t) with degree of freedom n-1
n=25, t= 1.84, pvalue = P(T >1.84)= 0.025 < pvalue < 0.05 (table)

In case we use two tail test, t=1.84, n-1=24 à 0.05 < pvalue <0.1

6. Test for proportion, p

7. Compare two proportions: p1 and p2

To test, we also have three types of hypothesis:

Two tail: Ho: p1=p2; H1: p1≠p2

Right tail Ho: p1 ≤p2; H1: p1 >p2
Left tail Ho: p1 ≥ p2; H1: p1 <p2

Test statistic:
( p1 −p 2)
z=
s p −p
1 2

In which:

pbar1 and pbar2: sample proportions of two sample, m1/n1 and m2 /n2

Rule of rejection:
Two tail: Ho: p1=p2; H1: p1≠p2 à Reject Ho if z > zα/2 or z < -zα/2
Right tail Ho: p1 ≤p2; H1: p1 >p2 à Reject Ho if z > zα
Left tail Ho: p1 ≥ p2; H1: p1 <p2 à Reject Ho if z < -zα

Example: To compare the proportion of students who hold the IELTS certificate
In two University A and B, we take two sample of 200 and 250 students from each
Uni. There are 90 and 120 students who have held the certificate in each Uni respective
Test with significant level α= 10%.

Hypothesis: To compare p1 and p2, we use the two tail:

Ho: p1=p2; H1: p1≠p2

90
p1 = =0.45 ;
200
120
p2 = =0.48
250

sp − p =
1 2
√ n1
+
n2 √
p 1 (1− p1 ) p 2 (1− p2 )
=
0.45(1−0. 45) 0.48(1−0.48)
200
+
250
=0.047
 ( p1 −p 2) 0.45−0.48
z= = =−0.64
s p −p
1 2
0.047

Reject Ho if z > z0.05= 1.645 or z < -1.645

So, we do not reject Ho –
Conclusion: the proportion of two Uni are the same

8. Regression:

Given model, with estimated coefficients, b0, b1, b2

- Explain the meaning of coefficients
- Calculate Coefficient of determination, R2 = SSR/TSS= RSS/TSS
- Using R-adjusted square (Rbar2) to choose between two models
Model with higher Rbar2 is the better.

2 2 n−1
Rb a r =1−(1−R )
n−p−1
p= the number of independent variables (p=2)
- Predict the value of Y given the value of X

EXAMPLE:

1. Binomial distribution:
The proportion of defective products from machine is 10%. Take randomly
20 products from this machine (independent).
a. Find the probability that there is no more 2 defective products taken?
b. Each defective product costs 100000$ for reproducing. Find the average
And standard deviation of cost of 20 product taken due to defective.

Solution:

3 conditions: 20 independent trials, defect and defect, P(defect)=0.1, P(nondefect)=0.9

Let X be the number of defect within 20 productsà X ~Bin(20, 0.1)

a. P(X ≤ 2) = P(X=0)+P(X=1)+P(X=2)
P(X=0)= C(20,0)*0.10*0.920
b. Find the average defects and sd of defect
E(X)=np=2; sd(X)= np(1-p)=2*0.9=1.8

Cost = Y = 100000*X
Average cost = E(Y) =E(100000X) = 100000*E(X)=
Sd(cost) = sd(100000X) =100000*sd(X)

2. Normal distribution:
Score of Maths of NEU Students has the normal distribution with mean
 of 76 point and standard deviation of 8
a. Take a student randomly, find the probability that his math score
Is more than 88?
b. The probability of 25% that a math score less than how much?

Solution:

X~N(76, 82)

a. P(X > 88) =P((X-76)/8) > (88-76)/8) = P(Z > 1.5) = 1- P(Z <1.5)= 1-0.9332
= 0.0668
b. Let the score be xo: 0.25= P(X <x0)

P(Z < (xo-76)/8) =0.25

Tableà (x0-76)/8 = -0.675 à x0 =76 -0.678*8=

3. - Estimate the CI 90% of students who hold the English certificate in NEU,
Take a sample of 120 students in which there are 46 of them holding
Certificate.
- Find the sample size needed such that the confidence level used be
98% and margin of error be 3%?
- Take another sample of 150 students in FTU in which there are 70 of them
Holding the certificate. Compare the proportions of students holding
Certificate of two universities.
4. Test for mean, µ
The time for coming home from Uni of NEU students is a variable with
Normal distribution. There is the hypothesis that the average time is more
than 34 minutes. To test for this hypothesis, we measure time of 28 students
Found that their average time (sample mean) be 37.5m and sd= 10.2m.
Test with significant level alpha =0.05.

Solution: Ho: µ =< 34; H1: µ >34

Calculate t = (37.5-34)/ (10.2/sqrt(28))=1.81

Reject Ho if t > t270.05= 1.703

So, we reject Ho, the hypothesis is True

- Find the p-value of the test?

5. Regression: refer to the points above.