Large Sample Test
Large Sample Test
Hypothesis:
Statistical Hypothesis
A statistical hypothesis is an assumption about the parameter of the
population or the nature of the population. e.g.
i. The population mean μ = 25
ii. The average weights of students of college A and college B are
same
iii. 20% of the students of a college are non vegetarians.
iv. The given population is a Binominal population are some of the
hypothesis.
Null Hypothesis:
A statistical hypothesis which is taken for the possible acceptance is
called a null hypothesis and it is denoted by Ho. The neutral attitude of
the decision maker, before the sample observations are taken is the
keynote of the null hypothesis. e.g.
Mean of the population is 60
H0: μ = 60.
Means of both populations are equal
H0: μ1 = μ2
The coin is unbiased
1
H0: P =
2
Level of Significance:
In any test procedure both the types of errors should be kept minimum.
But as they are inter-related it is not possible to minimize both the
errors simultaneously. Hence in practice, the probability of type-l error
is fixed and the probability of type-ll error is minimized. The fixed value
of type-I error is called level of significance and it is denoted by a. Thus,
level of significance is the probability of rejecting a hypothesis which
ought to be accepted. The most commonly used level of significance
are 5% and 1%. When a decision is taken at 5% level of significances, it
is meant that in 5 cases out of 100, it is likely to reject a hypothesis
which ought to be accepted. In other words, our decision to reject H0 is
95% correct.
Two – tailed and One – tailed Test:
In a standard normal curve the critical region may be given ways: (i) on
both the tails of the curve (ii) on one tail of the curve i.e. either on the
right tail of the curve or on the left tail of the curve. If the given null
hypothesis is tested using the critical region represented by both the
tails of the normal curve, it is called two tailed test and if the null
hypothesis is tested using the critical region represented by only one
tail of the normal curve, it is called one-tailed test. Whenever it is
required to test whether the sample statistic is significantly different
from the population parameter. i.e. whether the difference may be
regarded as due to chance or due to sampling fluctuations, the two tail
test is used. If we want to test that the mean of the population is
greater than specified value or the mean of one population is greater
than the mean another population or the proportion of defectives in
the population is less than a specified proportion, one tail test is used.
The following figures show the critical regions corresponding to 5% and
1% levels of significance for two-tailed and one tailed test.
Decision Making:
From the given data the value of test statistic Z is computed. If the
computed value of Z falls in the acceptance region i.e. for two tailed
test the value of Z is less than 1.96, the null hypothesis may be
accepted. If the computed value of Z falls in the critical region i.e. for
two tailed test Z is greater than 1.96 the null hypothesis may be
rejected.
Large Sample Test:
Step 1: Stating clearly null and alternative hypothesis.
Step 2: Finding out the difference between observed value and the
value taken in null hypothesis.
Step 3: Calculating S.E. of statistic
Step 4: Computing the value of Z
𝐷𝑖𝑓𝑓𝑒𝑟𝑒𝑛𝑐𝑒
Z=
𝑆.𝐸.
1% 5% 10%
Two-tailed test 2.58 1.96 1.645
One-tailed test 2.33 1.645 1.282
𝜎1 2 𝜎22
3. S.E. of S1 – S2 = √ +
2𝑛1 2𝑛2
𝐷𝑖𝑓𝑓𝑒𝑟𝑒𝑛𝑐𝑒
4. Z =
𝑆.𝐸.
𝐷𝑖𝑓𝑓𝑒𝑟𝑒𝑛𝑐𝑒
4. Z =
𝑆.𝐸.
𝐷𝑖𝑓𝑓𝑒𝑟𝑒𝑛𝑐𝑒
4. Z =
𝑆.𝐸.