CHAPTER 8: STATISTICAL INFERENCES

Inference is the process of making interpretations or conclusions from sample data for the totality
of the population.
 It is only the sample data that is ready for inference.
 In statistics there are two ways though which inference can be made.
 Statistical estimation
 Statistical hypothesis testing.
Statistical Estimation
This is one way of making inference about the population parameter where the investigator does
not have any prior notion about values or characteristics of the population parameter.
There are two ways of estimation.
1) Point Estimation
It is a procedure that results in a single value as an estimate for a parameter.
2) Interval estimation
It is the procedure that results in the interval of values as an estimate for a parameter, which is
interval that contains the likely values of a parameter. It deals with identifying the upper and
lower limits of a parameter. The limits by themselves are random variable.
Confidence interval for the parameter is:
Estimate ± critical value × Standard error of the estimator
Properties of best estimator
The following are some qualities of an estimator.
 It should be unbiased.
 It should be consistent.
 It should be relatively efficient.
To explain these properties let 𝜃̂ be an estimator of θ.
1. Unbiased Estimator: An estimator whose expected value is the value of the parameter
being estimated. i.e., E(𝜃̂ ) = θ.
2. Consistent Estimator: An estimator which gets closer to the value of the parameter as
the sample size increases. i.e., 𝜃̂ gets closer to θ as the sample size increases.
3. Relatively Efficient Estimator: The estimator for a parameter with the smallest variance.
This actually compares two or more estimators for one parameter.
Definitions
The confidence level of an interval estimate of a parameter is the probability that the
interval estimate will contain the parameter.
A confidence interval is a specific interval estimate of a parameter determined by using
data obtained from a sample and by using the specific confidence level of the estimate.
Degrees of Freedom: The number of data values which are allowed to vary once a statistic has
been determined.

1
Estimator: A sample statistic which is used to estimate a population parameter. It must be
unbiased, consistent, and relatively efficient.
Estimate: Is the different possible values which an estimator can assumes.
Point and Interval estimation of the population mean: μ
 Point Estimation
Another term for statistic is point estimate, since we are estimating the parameter value. A point
estimator is the mathematical way we compute the point estimate. For instance, sum 𝑥𝑖 over n is
∑𝑛 𝑥
𝑖
the point estimator used to compute the estimate of the population means, 𝜇 .That is 𝑥̅ = 𝑖=1 is
𝑛
a point estimator of the population mean.
 Confidence interval estimation of the population mean
Although X possesses nearly all the qualities of a good estimator, because of sampling error, we
know that it's not likely that our sample statistic will be equal to the population parameter, but
instead will fall into an interval of values. We will have to be satisfied knowing that the statistic
is "close to" the parameter. That leads to the obvious question, what is "close"? We can phrase
the latter question differently: How confident can we be that the value of the statistic falls within
a certain "distance" of the parameter? Or, what is the probability that the parameter's value is
within a certain range of the statistic's value? This range is the confidence interval.
The confidence level is the probability that the value of the parameter falls within the range
specified by the confidence interval surrounding the statistic.
There are different cases to be considered to construct confidence intervals.
Case 1: If the sample size is large with known variance
Recall the Central Limit Theorem, which applies to the sampling distribution of the mean of a
sample. Consider samples of size n drawn from a population, whose mean is 𝜇 and standard
deviation is 𝜎 with replacement and order important. The population can have any frequency
𝜎
distribution. The sampling distribution of 𝑥̅ will have a mean 𝜇𝑥̅ = 𝜇 and a standard 𝜎𝑥̅ = 𝑛, and
√
approaches a normal distribution as n gets large. This allows us to use the normal distribution
curve for computing confidence intervals.
𝑥̅ −𝜇
⇒𝑧 = 𝜎 is normal with mean 0 and standard deviation 1.
⁄ 𝑛
√
⇒𝜇 = 𝑥̅ ± 𝑧 𝜎⁄ = 𝑥̅ ± 𝜀, where 𝜀 is a measure of error.
√𝑛
⇒𝜀 = 𝑧 𝜎⁄
√𝑛
 For the interval estimator to be good the error should be small. How it be small?
 By making n large
 Small variability
 Taking Z small
To obtain the value of Z, we have to attach this to a theory of chance. That is, there is an area of
size 1-a such

𝑃(−𝑍𝛼 ⁄2 < 𝑍 < 𝑍𝛼 ⁄2 ) = 1 − 𝛼, Where α is risk probability and 1- α confidence level.

̅ −𝜇
𝑋
⇒𝑃 (−𝑍𝛼 ⁄2 < 𝜎⁄ < 𝑍𝛼 ⁄2 ) = 1 − 𝛼
√𝑛

⇒𝑃(−𝑍𝛼 ⁄2 𝜎⁄√𝑛 < 𝑋̅ – 𝜇 < 𝑍𝛼 ⁄2 𝜎⁄√𝑛) = 1 − 𝛼

2
⇒𝑃(−𝑍𝛼 ⁄2 𝜎⁄√𝑛 − 𝑋̅ < −𝜇 < −𝑋̅ + 𝑍𝛼 ⁄2 𝜎⁄√𝑛) = 1 − 𝛼
⇒𝑃(𝑋̅ − 𝑍𝛼 ⁄2 𝜎⁄√𝑛 < 𝜇 < 𝑋̅ + 𝑍𝛼 ⁄2 𝜎⁄√𝑛) = 1 − 𝛼
The last statement clearly shows that, there is a (1- 𝛼) 100% confidence interval for population
mean (μ) to lie in the interval
(𝑋̅ − 𝑍𝛼 ⁄2 𝜎⁄√𝑛 , 𝑋̅ + 𝑍𝛼 ⁄2 𝜎⁄√𝑛)
This interval is known as a (1- 𝛼) 100% confidence interval for population mean (μ).

Here are the Z values corresponding to the most commonly used confidence levels.

(1- 𝛼) 100% 𝛼 𝛼 ⁄2 𝑍𝛼 ⁄2
90 0.10 0.05 1.645
95 0.05 0.025 1.96
99 0.01 0.005 2.58
Case 2: If sample size is small and the population variance 𝜎 2 is not known.
̅̅̅
𝑡 = 𝑆𝑋⁄−𝜇
√𝑛
has t distribution with n-1 degrees of freedom.
𝑆
(𝑋̅ – 𝑡𝛼 ⁄2(𝑛−1) 𝑛 , 𝑋̅ + 𝑡𝛼 ⁄2(𝑛−1) √𝑛
𝑆
) is a 100 (1- 𝛼)% confidence interval for 𝜇
√
The unit of measurement of the confidence interval is the standard error. This is just the standard
deviation of the sampling distribution of the statistic.
Examples:
1. From a normal sample of size 25 a mean of 32 was found .Given that the standard deviation
is 4.2. Find

a) A 95% confidence interval for the population mean.

b) A 99% confidence interval for the population mean.

Solution:
a) Given: n = 25 𝑋̅ = 32, S = 4.2, 1-α = 0.95 ⟹ α = 0.05, α2 = 0.025
⟹ 𝑡α⁄2,24 = 2.064 𝑓𝑟𝑜𝑚 𝑡𝑎𝑏𝑙𝑒.

𝑆
⟹ The required interval will be (𝑋̅ – 𝑡𝛼 ⁄2(𝑛−1) , 𝑋̅ + 𝑡𝛼 ⁄2(𝑛−1) √𝑛
𝑆
)
√𝑛
4.2
= 32 ± 2.064 ×
√25
= 32 ± 1.73
= (30.27, 33.73)
b) Given: n = 25, 𝑋̅ = 32, S = 4.2, 1-α = 0.99 ⟹ α = 0.01, α2 = 0.005

3
 ⟹ 𝑡α⁄2,24 = 2.797 𝑓𝑟𝑜𝑚 𝑡𝑎𝑏𝑙𝑒.

𝑆
⟹ The required interval will be (𝑋̅ – 𝑡𝛼 ⁄2(𝑛−1) , 𝑋̅ + 𝑡𝛼 ⁄2(𝑛−1) √𝑛
𝑆
)
√𝑛
4.2
= 32 ± 2.797×
√25
= 32 ± 2.35
= (29.65, 34.35)
2. A drug company is testing a new drug which is supposed to reduce blood pressure. From the
six people who are used as subjects, it is found that the average drop in blood pressure is 2.28
points, with a standard deviation of 0.95 points. What is the 95% confidence interval for the
mean change in pressure?
Solution:
𝑥̅ = 2.28 𝑠 = 0.95 1 − 𝛼 = 0.95 ⇒ 𝛼 = 0.05, 𝛼⁄2 = 0.025
⇒ 𝑡𝛼⁄2 (5) = 2.571 from table.
𝑆 0.95
⇒ The required interval will be (𝑋̅ ± 𝑡𝛼 ⁄2(𝑛−1) )= (2.28 ± 2.571 ∗ ) = (2.28 ±
√𝑛 √6
1.008) = (1.28, 3.28)

Statistical Hypothesis testing

Hypothesis Testing:-This is also one way of making inference about population parameter,
where the investigator has prior notion about the value of the parameter.
Definitions:
 Statistical hypothesis: is an assertion or statement about the population whose plausibility is
to be evaluated on the basis of the sample data.
 Test statistic: is a statistics whose value serves to determine whether to reject or accept the
hypothesis to be tested. It is a random variable.
 Statistic test: is a test or procedure used to evaluate a statistical hypothesis and its value
depends on sample data.
There are two types of hypothesis:
1. Null hypothesis:
- It is the hypothesis to be tested.
- It is the hypothesis of equality or the hypothesis of no difference.
- Usually denoted by H0.
2. Alternative hypothesis:
- It is the hypothesis available when the null hypothesis has to be rejected.
- It is the hypothesis of difference.
- Usually denoted by H1 or Ha.
Types and size of errors:
Testing hypothesis is based on sample data which may involve sampling and non sampling
errors. The following table gives a summary of possible results of any hypothesis test:

Decision Ho is true Ho is false

4
 Reject Ho Type I error Correct decision

Accept Ho Correct decision Type II error

 Type I error: Rejecting the null hypothesis when it is true.

 Type II error: Failing to reject the null hypothesis when it is false.
NOTE:
1. There are errors that are prevalent in any two choice decision making problems.
2. There is always a possibility of committing one or the other errors.
3. Type I error (  ) and type II error (β ) have inverse relationship and therefore, cannot be
minimized at the same time.
 In practice we set α at some value and design a test that minimize β . This is because a type I
error is often considered to be more serious, and therefore more important to avoid, than a
type II error.
General steps in hypothesis testing:
1. The first step in hypothesis testing is to specify the null hypothesis (H0) and the
alternative hypothesis (H1).
2. The next step is to select a significance level, α
3. Identify the sampling distribution of the estimator.
4. The fourth step is to calculate a statistic analogous to the parameter specified by the null
hypothesis.
5. Identify the critical region.
6. Making decision.
7. Summarization of the result.
Hypothesis testing about the population mean,

Suppose the assumed or hypothesized value of μ is denoted by μ o, then one can formulate two
sided (1) and one sided (2 and 3) hypothesis as follows:

1. Ho: μ =μo versus H1: μ  μo (two sided test)

2. Ho: μ = μo versus H1: μ < μo (one sided test)
3. Ho: μ= μo versus H1: μ > μo (one sided test)
To identify the sampling distributions, there are some cases:
Case-I: Population variance (σ2) is known and parent population is normal.
𝑋̅−𝜇
The relevant test statistic is 𝑍 = 𝜎 ~ N (0, 1).
⁄ 𝑛
√

- After specifying  we have the following regions (critical and acceptance) on the standard
normal distribution corresponding to the above three hypothesis.
Summary table for decision rule:
Alternative hypothesis Accept H0 if Reject H0 if Inconclusive if

5
  0  Z / 2  Z C  Z / 2 Z C  Z  / 2 or Z C  Z  / 2 ZC  Z / 2 or ZC  Z / 2

  0 ZC  Z Z C  Z Z C  Z
  0 Z C  Z Z C  Z Z C  Z

Case-II: Population variance (σ2) is unknown and parent population is normal and small sample
size.
𝑋̅−𝜇0
The relevant test statistic is 𝑡 = 𝑆⁄ ~ t with n-1 degrees of freedom.
√𝑛

After specifying  we have the following regions on the student t distribution corresponding to
the above three hypothesis.
Alternative hypothesis Accept H0 if Reject H0 if Inconclusive if
  0  t / 2  tC  t / 2 tC  t / 2 or tC  t / 2 tC  t / 2 or tC  t / 2

  0 tC  t tC  t tC  t
  0 tC  t tC  t tC  t
𝑥̅ −𝜇0
Where 𝑡𝑐 = 𝑠
⁄ 𝑛
√

Case-III: When sampling from a non normal population and when the sample size is large the
distribution of X depends on Central Limit Theorem (with known and unknown variance).
𝑋̅−𝜇0
a) The test statistic is: 𝑍𝑐 = 𝜎 ~ N (0, 1) with known variance
⁄ 𝑛
√
𝑋̅−𝜇0
b) The test statistic is: 𝑍𝑐 = 𝑆⁄ ~ N (0, 1) with unknown variance.
√𝑛

Note: Zc refers to Zcalculated

Example1: The average number of mango fruits per tree in a particular region was known from a considerable
experience as 520 with a standard deviation 4.0. A sample of 20 trees gives an average number of fruits 450
per tree. Test whether the average number of fruits per tree selected in the sample is in agreement with the
average production in that region?
Solution:

* H 0 :   520 versus H A :   520 ,   0.05

* Z  Z0.025  1.96 (critical value)

2

6
 |𝑥̅ −𝜇0 | |450−520|
∗ 𝑍𝑐 = 𝜎 = 4.0⁄ = 78.26 (𝑐𝑎𝑙𝑐𝑢𝑙𝑎𝑡𝑒𝑑 𝑣𝑎𝑙𝑢𝑒)
⁄ 𝑛
√ √20

* Since |𝑍𝐶 | > 𝑍𝛼⁄2 , the null hypothesis is rejected. Therefore, it can be concluded that there is
significant difference between sample mean and population mean with respect to average
performance at 5% level of significance.
Example2: Based upon a random sample of size 100 with an average of 3.4 minutes and a
standard deviation of 2.8 minutes, is the claim that the average telephone call is 4 minutes true
with a confidence of 95%?

Solution: Given: n  100, x  3.4 min, s  2.8 min,   0.05

H 0 :  4
To test:
H A :  4

Since  is unknown this should be a t-distribution; however, since n  100 is large the z-satistic
is used.

X  0 3.4  4
Zc    2.14
S/ n 2.8 / 10

From the standard normal table we have,  Z / 2   Z 0.025  1.96

Since the calculated value is less than the tabulated value (-2.14<-1.96), the null hypothesis
will be rejected. Therefore average telephone call is significantly different from 4 minutes at
  0.05.

Example 3:The average daily milk production of a particular variety of buffaloes was given as 12
kgs. The distribution of daily milk yield in a dairy farm was given as follows:

Daily milk yield (kgs) 6-8 8-10 10-12 12-14 14-16 16-18
No. of buffaloes 9 20 35 42 17 7
Test whether the performance of dairy farm was in agreement with the record. Use   0.05
(Exercise)
Example 4: A sample of 16 students gave an average mark of 53.8 with a standard deviation of
5.2. Can we conclude that the population mean of marks is 50 at   0.05 ?

Solution: H 0 :   50 H A :   50

  0.05 and hence  / 2  0.025

7
 t / 2,n1  t 0.025, 15  2.131 .

x  0 53.8  50 3.8
tC     2.92.
s/ n 5.2 / 16 1.3

Since tc  2.92  2.131, H 0 is rejected. Therefore the population mean mark is significantly

different from 50 at   0.05.

Chi-Square Test of association of attributes

- The chi-square procedure test is used to test the hypothesis of independency of two attributes
.For instance we may be interested

 Whether the presence or absence of hypertension is independent of smoking habit or

not.
 Whether the size of the family is independent of the level of education attained by the
mothers.
 Whether there is association between father and son regarding boldness.
 Whether there is association between stability of marriage and period of acquaintance
ship prior to marriage.

- The  2 statistic is given by:

 (Oij  eij ) 2 
 ~  ( r 1)( c 1)
r c 2
 2
cal   
i 1 j 1  eij 

Where Oij  the number of units that belong to category i of A and j of B.
eij  Expected frequency that belong to category i of A and j of B.

The eij is given by:

Ri * C j
eij 
n

Where Ri  the i th row total.

C j  the j th column total.
n  total number of oservations

8
 Remark:

r c r c
n    Oij    eij
i 1 j 1 i 1 j 1

- The null and alternative hypothesis may be stated as:

H 0 : There is no association between A and B.
H1 : not H 0 ( There is association between A and B ).

Decision Rule:

- Reject H0 for independency at  level of significance if the calculated value of  2 exceeds

the tabulated value with degree of freedom equal to (r  1)(c  1) .
r c  
 Reject H 0 if  2 cal     (Oij  eij )    2( r 1)( c 1) at 
2

i 1 j 1  eij 

Example1: Random samples of 200 men, all retired were classified according to education and
number of children is as shown below.

Level of Number of children

education 0-1 2-3 Over3
elementary 14 37 32
Secondary 31 59 27
and above
Test the hypothesis that the size of the family is independent of the level of
education attained by fathers. (Use 5% level of significance)

Solution:

𝐇𝟎 : There is no association between family size and level of education

𝐇𝐚 : There is association between family size and level of education

- First calculate the row and column totals

𝑅1 = 83, 𝑅2 = 117, 𝐶1 = 45, 𝐶2 = 96, 𝐶3 = 59

- Then calculate the expected frequencies( eij’s)

𝑅𝑖 ∗𝐶𝑗
𝑒𝑖𝑗 = 𝑛

9
 𝑅1 ∗ 𝐶1 83 ∗ 45
𝑒11 = = = 18.7
𝑛 200
𝑅1 ∗ 𝐶2 83 ∗ 96
𝑒12 = = = 39.8
𝑛 200
𝑅1 ∗ 𝐶3 83 ∗ 59
𝑒13 = = = 24.5
𝑛 200
𝑅2 ∗ 𝐶1 117 ∗ 45
𝑒21 = = = 26.3
𝑛 200
𝑅2 ∗ 𝐶2 117 ∗ 96
𝑒22 = = = 56.2
𝑛 200
𝑅2 ∗ 𝐶3 117 ∗ 59
𝑒23 = = = 34.5
𝑛 200

- Obtain the calculated value of the chi-square.

2 3 2
(𝑂𝑖𝑗 − 𝑒𝑖𝑗 )
𝑋 2 𝑐𝑎𝑙 = ∑∑[ ]
𝑒𝑖𝑗
𝑖=1 𝑗=1

(14 − 18.7)2 (37 − 39.8)2 (32 − 24.5)2 (31 − 26.3)2

= + + +
18.7 39.8 24.5 26.3
(59 − 56.2)2 (27 − 34.5)2
+ + = 6.29
56.2 34.5
- Obtain the tabulated value of chi-square
  0.05
Degrees of freedom  (r  1)(c  1)  1 * 2  2
 02.05 (2)  5.99 from table.

- The decision is to reject H0 since

 2 cal   2 ( 2)
0.05

Conclusion: At 5% level of significance we have evidence to say there is association between

family size and level of education, based on this sample data.

10