Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                

BUSINESS STATISTICS - Chp6

Download as pdf or txt
Download as pdf or txt
You are on page 1of 14

Chapter 6

HYPOTHESIS TESTING

Learning Objectives:

After reading information on Chapter 6, you should be able to:

1. Appreciate hypothesis testing


2. Make statistical hypothesis
3. Determine the level of significance
4. Conduct Test Static acceptance region
5. Make statistical decision
6. Interpret and draw conclusion
7. Summarized the steps in hypothesis testing

6.1 Statistical Hypothesis

Statistical hypothesis is an assertion concerning one or more populations.


The correctness or falseness of a statistical hypothesis is unknown with absolute
certainty unless we look into the entire population. In most situations, this is not
possible. Instead, we take a random sample from the population of interest and
use the data contained in this sample to provide evidence that either accepts or
rejects the hypothesis.
There are two types of statistical hypothesis. This includes the null
hypothesis and the alternative hypothesis. The null hypothesis (H0) is a
hypothesis of no difference. A null hypothesis should always be stated for the
purpose of being rejected or not being rejected.
Examples:
1. The average scores of students in abstract reasoning are 85. (H0: 𝜇= 85)
2. The average time of antibiotic to take effect is 30 minutes. (H0: 𝜇= 35)
3. There is no difference in mean scores between male and female students
in their mathematical skills. (H0 :𝜇m – 𝜇f = 0)

The rejection of the null hypothesis leads to the acceptance of an alternative


hypothesis (H1,). The alternative hypothesis is sometimes called the research
hypothesis, i.e., the hypothesis that the researcher believes in. After properly
stating the null hypothesis, the alternative hypothesis allows for the possibility of
several values, be it not equal to the stated null hypothesis, less than the
hypothesized values, or greater than the hypothesized values.
Alternative hypotheses are classified into non directional alternative
hypothesis or directional alternative hypothesis. A non-directional alternative
hypothesis is an assertion that there-is significant difference between two
descriptive measures. This hypothesis makes use of the “not equal to (≠)” symbol.
It is called two-sided hypothesis.

Examples:
1. The average scores of students in abstract reasoning are not equal to 85. (H1:
𝜇m – 𝜇 ≠85)
2. 2. The average time of antibiotic to take effect is not equal to 30 minutes.
(H1: 𝜇m – 𝜇 ≠30)
3. There is difference in mean scores between male and female students in their
mathematical skills. (H1: 𝜇m – 𝜇f≠0)

A directional alternative hypothesis is an assertion that one descriptive


measure is less than or greater than another descriptive measure. This hypothesis
makes use of the “less than (<)” symbol or “greater than (>)” symbol. It is
sometimes called one-sided hypothesis. :
Examples:
1. The average scores of students in abstract reasoning are greater than 85. (H1:
𝜇> 85)
2. The average time of antibiotic to take effect is less than 30 minutes. (H1: 𝜇<
30)
3. The mean scores of male students are higher than female students in their
mathematical skills. (H1:𝜇m – 𝜇f > 0)

The following are guidelines in determining which hypothesis should be


stated as null hypothesis or alternative hypothesis:

1. Read and understand the problem carefully. After which, decide on what you
want to test.
2. 2. If simple direction such as more than, less as, superior to, inferior to, and
so on is suggested, then alternative hypothesis will be stated ‘using the
inequality symbols (< or >).
3. If compound direction such as at least, equal to or greater than, at most, no
more than, and so on is suggested, then this entire compound direction (< or
>) is expressed as null hypothesis, but using only the equality sign and
alternative hypothesis is given the opposite direction.
4. If no direction is suggested, then alternative hypothesis is stated using the
not equal ≠symbol.

6.2 Level of Significance

Level of significance is the probability that the null hypothesis will be


rejected in error when it is true, 𝛼. Significance levels show how likely a
result is due to chance.
Most researchers have used any of the three levels of significance, i.e., '
0.10, 0.05 and 0.01. The choice of significance levels is subjective. The
lower the significance level, the more the data must diverge from the null
hypothesis to be significant. Therefore, the 0.01 level is more conservative
than the 0.05 level.
Whenever a null hypothesis is “rejected” or “fail to reject” on the basis -
of a sample, there are four possibilities as shown in Table 6.1

Table 6.1: Decisions in Testing a Statistical Hypothesis

Decisions/Fact H0 is true H0 is false


Fail to reject H0 Correct decision Type II error
Reject H0 Type I error Correct decision

From the above table, of the four possibilities, there are two correct
decisions and two possible errors that may be committed in hypothesis
testing. The first correct decision in hypothesis testing is that of failing to
reject the null hypothesis when it is really true (possibility 1). A second
correct decision in hypothesis testing is that of rejecting the null
hypothesis when it is really false (possibility 4).

On the other hand, there are two possible types of error that may be
committed in hypothesis testing. These include Type I error and Type II
error. The first type of error committed is that of rejecting the null
hypothesis when it is really true (possibility 2) on the basis of the sample
results. This error is called a Type I error. The probability of committing
Type l error is denoted by𝛼. Different 𝛼-levels have different advantages
and disadvantages. A very small𝛼 −level (say, 0.01) is less likely to be more
extreme than the critical value and so is more significant than high 𝛼-level
values (say 0.05).

The second error committed is that of failing to the null hypothesis


when in fact it is really false (possibility 3) on the basis of the sample
results. This type of error is called Type II error. The probability of
committing Type II error is designated as 𝛽. In general, the smaller the value
of 𝛼, the lesser is the probability of rejecting a true null hypothesis and the
higher is the probability of failing to reject a false null hypothesis, and so
have less power.

The selection of an a-level inevitably involves a compromise between


significance and power and consequently between Type I error and Type II
error,

6.3 Test Static Acceptance Region

 Test Statistic
The selection of the test statistic depends on the types of variables; level
of measurement, whether the samples are independent or dependent, and the
assumption on the distribution about the data (refer to Table 6.2).
Table 6.2: Some Commonly Used Test Statistics in Research
Type of Test Statistic Variables Sample Sizes Remarks
Sample
One- z-test Quantitative Large Mean
Sample z-test Qualitative Large Proportion
t-test Quantitative Small Mean
sign test Continuous Small Median
Independent z-test Quantitative Large Two Means
Samples z-test Qualitative Large Two Proportions
t-test Quantitative Small Two Means
chi-square test Qualitative Large Association
chi-square test Qualitative Large Two proportions
fisher test Qualitative Small Association
Mann-Whitney Continuous Small Two Medians

Dependent Paired t-test Quantitative Small Two Means


Sample McNemar’s test Qualitative Regardless Two Proportions
Signed-Ranked Continuous Small Two Medians
k- F-test Quantitative Large Among Means
Independent Chi-square test Qualitative Large Association
Samples Chi-square test Qualitative Large Among
Kruskal Wallis Continuous Small Proportions
Among Medians

 Critical and Acceptance Regions


The critical region is the region in which the null hypothesis is rejected if
the test statistic falls within. By specifying a level of significance, the critical
value can be determined, which defines a region of rejection and a region of
acceptance. This critical value serves as the basis whether the hypothesis is
rejected or failed to reject.
The size of the area of the region of critical region is given by the level of
significance,𝛼. The location of the critical region is determined by the nature of
the alternative hypothesis, whether one-sided or two-sided alter- native
hypothesis.
If the alternative hypothesis is non-directional (i.e., H1: 𝜃 ≠ ∅0), both tails of
the distribution will contain equal areas serving as the critical region. A test of
statistical hypothesis, where the alternative hypothesis is two-sided (non-
directional) is called a two-sided test.
For a two-tailed test, Figure 22(a) and Figure 22(b) showed that the critical z
values are ±1.96 and ± 2.576 for the 0.05 and 0.01 levels of significance,
respectively. Thus, we reject H0, at 0.05 level of significance if the z computed
statistic is greater than 1.96 or less than -1.96; and we reject A, at 0.01 level of
significance if the z computed value is more than 2.576 or less than -2.576.

If the alternative hypothesis is directional (i.e., H1S:𝜃>𝜃0), the critical


region or rejection region lies entirely in the right tail of distribution. A test
of statistical hypothesis corresponding to this is called one-sided test. Figure
23(a) and Figure 23(b) showed the critical value for a one-tailed test which
requires that the region of rejection must lie entirely in the right tails of the
distribution which is z = 1.645 when = 0.05 and z = 2.326 when @= 0.01.
This means we reject 7, at 0.05 level of significance if the z-computed value
falls beyond 1.645, and we reject H0 at 0.01 level of significance if the z-
computed value is greater than 2.326 by using one tailed test (right-tail). The
region of rejection in each case is the area to the right of the critical value,
while the region of acceptance is the area to the left of the critical value.

If the alternative hypothesis is directional (.e, H1:𝜃>𝜃 0), the critical


region or rejection region lies entirely in the left tail of distribution. A test of
statistical hypothesis corresponding to this is also called one-sided test.
Figure 24(a) and Figure 24(b) showed that the critical value requires that the
region of rejection must be placed entirely in the left tails of the distribution,
z = —1.645 for 𝛼 = 0.05 and z =-2.326 for 𝛼= 0.01. Thus, any computed z
values fall beyond -1.645 and -2.326 for the 0.05 and 0.01 levels of
significance, respectively; the null hypothesis should be rejected, since the
value falls on the region of rejection. The region of rejection in each case is
the area to the left of the critical value, while the region of acceptance is the
area to the right of the rejection region.

The critical regions in statistical hypothesis testing can be summarized as follows:

Type of Test
Level of Significance One-tailed Two-tailed
left tail right

a = 0.05 z ≤ −1.645 z ≤ 1.645 z ≤ −1.96 𝑜𝑟 z ≤ 1.96


a = 0.01 z ≤ −2.326 z ≤ 2.326 z ≤ −2.576 𝑜𝑟 z ≤ 2.576
 Computation of the Test Statistic
It is vital that before computation of the test statistic, the level of
significance have been set and the corresponding critical region have been located
properly, in order to avoid manipulation of these entities after the test statistic have
been computed to obtain desired outcome. Computation differs depending on the
test statistic used as enumerated previously.

6.4 Statistical Decision

Statistical decision in hypothesis testing is either to reject or not to reject the


null hypothesis. If the computed value of the test statistic falls within the critical
region, the null hypothesis is rejected. On the other hand, if the computed value
lies within the acceptance region, then do not reject the null hypothesis.
Another method is to look for the p-value of the test statistic. Smaller p-value
(<0.05) means reject the null hypothesis.

6.5 Interpretation/Draw Conclusion

If the statistical decision is to reject the null hypothesis, then the conclusion
arrived at is simply by restating the alternative hypothesis. On the other hand, if
the decision is not to reject the null hypothesis, then the interpretation is that we
have to say that there is no sufficient evidence to conclude whatever is stated in
the alternative hypothesis. A statistically significant difference simply means
there is statistical evidence that there is a difference; it does not mean the
difference is necessarily large, important, or significant in the usual sense of the
word.
6.6 Summary of Steps in Hypothesis Testing

 Summary of Steps in Hypothesis Testing


To prepare for hypothesis testing, a set of steps should be followed in order
to arrive at an objective and meaningful result. These include the following:
1. State the null and alternative hypothesis.
2. Choose a level of significance of size 𝛼.
3. Select the appropriate test statistics.
4. Establish the critical region.
5. Compute the value of the test statistic from the sample data.
6. Provide statistical decision.
7. Make an interpretation and, if possible, draw a conclusion.
SELF TEST

MULTIPLE CHOICE: Encircle the letter of your best answer.


1. Statistically significant means:
a. Do not reject HO
b. Accept HO
c. Reject HO
d. Either A or B
e. None of the above
2. Suppose that the obtained absolute value of test statistic falls within the
critical region, then the statistical decision is to:
a. Do not reject HO
b. Accept HO
c. Reject HO
d. Either A or B
e. None of the above
3. The hypothesis of no difference is called.
a. Research hypothesis
b. Null Hypothesis
c. Alternative hypothesis
d. Both A or C
e. None of the above
4. The proportion of males is different from the proportion of females in their
attitudes toward abortion:
a. Research hypothesis
b. Null hypothesis
c. Alternative hypothesis
d. Both A or C
e. None of the above
5. The proportion of heart attack is higher among smokers than non-smokers:
a. Non-directional alternative hypothesis
b. Directional alternative hypothesis
c. Null hypothesis
d. HO
e. None of the above
6. The test for the hypothesis that the prevalence of hypertension among
diabetics is higher than the prevalence among non-diabetics is an example of:
a. Two-tailed test
b. Two sided test
c. One-tailed test
d. Either A or B
e. None of the above
7. What is the critical value for the z-test statistic for two-tailed at the 0.05
level of significances?
a. 1.960
b. 3.841
c. 5.991
d. 6.035
e. 0.990
8. If the computed test statistics is 2.05 and the tabulated value at 𝛼 = 0.05 is
1.71, then the statistical decision is:
a. Reject H1
b. Reject HO
c. Fail to reject H1
d. Fail to reject HO
e. None of the above because the sample size used was not equal
9. The probability of rejecting the null hypothesis when it is true is called:
a. Critical region
b. Test statistic
c. Level of significance
d. Statistical decision
e. None of the above
10. For the hypothesis is which states that “the average daily allowance of
nursing students is less than P 150”, the critical region lies on:
a. Both tails of the normal distribution
b. Right tail of the normal distribution
c. Left tail of the normal distribution
d. Middle portion of the normal distribution
e. None of the above

You might also like