SAMPLING DISTRIBUTION

OF SAMPLE MEAN
 SAMPLING DISTRIBUTION OF SAMPLE MEAN

Is the distribution of all possible observations of sample means for

samples of a given size from a population. If population has normal
distribution or sample size is at least 30, distribution of sample means
tend to follow normal distribution (central limit theorem).

SAMPLING ERROR
Is the error resulting from using a sample to estimate the population
characteristics such as population mean. Sample provides data for only
portion of entire population, we can not expect sample to yield perfectly
information about population.

SAMPLING ERROR AND SAMPLE SIZE

The larger the sample size, the smaller the sampling error tend to be in
estimating population mean by sample mean
 SAMPLING DISTRIBUTION OF SAMPLE MEAN

In practice, populations with which we work are large and population

mean is unknown. We use some relationship to approximate the
population mean by sampling distribution of sample mean.

Mean of sample means,

Standard deviation of sample mean, ,for infinite population.

 CONFIDENCE INTERVAL (Unknown µ and known σ)

• We use a sample mean, 𝑥ҧ to estimate the population mean 𝜇, because of

sampling error one cannot expect 𝑥ҧ to equal 𝜇 exactly.

• So information about accuracy of estimate is important, which lead to

discussion of confidence interval for population mean.

• For normally distributed variable with confidence level 0.95, that is 95%.
Confidence interval for population mean is given by
 HYPOTHESIS TEST
✓ Used to make decisions or judgement about value of parameter such as population
mean. For example, we have to decide whether the mean weight of all bags packed by
a particular company differs from advertised weight of 454 grams.

✓ hypothesis involves two hypotheses namely

I. Null hypothesis, H0
This is the hypothesis to be tested

II. Alternative hypothesis, Ha

This is the hypothesis to be considered as alternative to Null hypothesis.

So, hypothesis test is to decide whether the null hypothesis should be rejected
in favor of alternative hypothesis.
 HYPOTHESIS TEST

✓ In other words, Hypothesis testing is a procedure to evaluate the strength of evidence

against a null hypothesis. Given the null hypothesis is true, we calculate the probability
of obtaining the observed evidence or more extreme, which is called p-value.

✓ If the p-value is small enough (small enough is decided by significance value, ∝ ), we

reject the null hypothesis.
 HYPOTHESIS TEST

The following are steps in hypothesis test

I. Formulation of hypotheses
That is Null hypothesis, H0 and Alternative hypothesis, Ha

Null hypothesis, H0: 𝜇=𝜇0

Alternative hypothesis, Ha depends on primary concern

✓ If concern is whether the population mean 𝜇 is different from specified value
𝜇0,then
Ha: 𝜇≠𝜇0, we call it two-tailed test
 HYPOTHESIS TEST

• ✓If concern is whether the population mean 𝜇 is less than specified value
𝜇0,then

Ha: 𝜇<𝜇0, we call it left tailed test

• ✓If concern is whether the population mean is greater than a specified

value, then

Ha: 𝜇>𝜇0, we call it right tailed test

 HYPOTHESIS TEST

II. Obtain significance level, ∝

✓ Is the probability such that when p-value is equal or less to it, p-value
is considered as small enough, e.g. 5% = 0.05

III. Select and compute the appropriate test statistics.

✓ Selection of test statistic depends on probability distribution of the variable
of interest.
 HYPOTHESIS TEST

✓ For normally distributed variable with unknown standard deviation, σ and sample
size (n < 30), the following test statistic is used:

t=

✓ For normally distributed variable with known standard deviation, σ and sample
size (n < 30), the following test statistic is used:

t=
 HYPOTHESIS TEST

IV. Conclusion
✓ For normally distributed variable t, We reject Null hypothesis (H0) if the value of
t is within the range

[Z∝, ±∞] or if its p-value is less than significance value ∝

 HYPOTHESIS TEST
(Normal distribution with Known σ)
 HYPOTHESIS TEST
(Normal distribution with unknown σ)