Z-test and T-Test

Objectives
At the end of the lesson, the learner

1. identifies the parameter to be tested given a real-life

problem;
2. formulates the appropriate null and alternative
hypotheses on a population mean;
3. identifies the appropriate form of the test-statistic when
(i) the population variance is assumed to be known, (ii)
the population variance is assumed to be unknown and
(iii) the Central Limit Theorem is to be used;
4. computes for the test-statistic value (population mean);
5. draws conclusion about the population mean based on
the test-statistic value and the rejection region; and
6. solves problems involving test of hypothesis on the
population mean.
REVIEW
Differentiate Descriptive statistics and Inferential
statistics?

Descriptive statistics is analyzing data that helps describe,

show or summarize data in a meaningful way such that
patterns might emerge from the data. It does not allow
conclusions beyond the data have analyzed or reach
conclusions regarding any hypotheses. This includes
Measures of central tendency (mean, median & mode) and
measures of variability (variance, standard deviation, age,
etc.).

Inferential statistics involves techniques that allow us to use

samples accurately to represent a population. It recognizes the
fact that sampling naturally incurs sampling error and thus a
sample is not expected to perfectly represent the population.

Definitions:
 

The areas under the normal curve are given in
terms of -values or scores. The formula for
calculating is with given measure , population mean
and population standard deviation . The formula will
be used if sample mean , population mean ,
population standard deviation and sample size n are
given.
The t-value is given by , where is the sample
standard deviation.
A   researcher used a developed problem solving test to
randomly select 50 Grade 6 pupils. In this sample, = 80 and
= 10. The mean and the standard deviation of the
population used in the standardization of the test were 75
and 15 respectively. Using 95% confidence level, does the
sample mean differ significantly from the population mean?

1. What are given in the problem?

= 80 = 10 = 75 = 15 n = 50
 
2. Is the sample small or large?
Large
3. Does the problem require a one-tailed or a two-tailed
distribution?
Two-tailed distribution
  Aresearcher used a developed problem solving test to
randomly select 50 Grade 6 pupils. In this sample, = 80
and = 10. The mean and the standard deviation of the
population used in the standardization of the test were 75
and 15 respectively. Using 95% confidence level, does the
sample mean differ significantly from the population
mean?

4. What are the appropriate hypotheses for the problem?

Null hypothesis: There is no significant difference
between the sample mean and the population mean.
 
Alternative hypothesis: There is a significant
difference between the sample mean and the
population mean.
 Definition:

A test statistic is a value used to determine the probability

needed in decision-making.
 
Two ways in testing hypothesis:

Traditional or classical method – the test statistics is the value

determined by a computational formula that is compared with a
confidence coefficient.
P-value method – the test statistics is the value determined by
the probability of committing a Type I error which is compared
to the alpha level. It becomes popular with easy access to
computer software and high-powered statistical calculators.
 
A one sample test is a test conducted on one sample
purportedly coming from a population with mean .
  If the sample is large, the test statistic is the . The z statistic or
z-test measures the number of standard deviations between the
observed value of the sample mean and the null hypothesized value of
the population mean. It has two cases.
 
1. The sample is large (n 30). Apply the Central Limit Theorem and use
the normal curve as a model.
2. When the CLT is applied, the sample standard deviation may be used
as an estimate of the population standard deviation when the value of
is unknown.
 
If the sample is small, the CLT cannot be applied, then t statistic or
t-test will be used. Generally, a t-test is used when the population
standard deviation is unknown. Nonetheless, a t-distribution approaches
z-distribution when the sample size becomes larger.
 
EXAMPLE 1:
A researcher used a developed problem solving test to randomly select 50
Grade 6 pupils. In this sample, = 80 and = 10. The mean and the standard
deviation of the population used in the standardization of the test were 75
and 15 respectively. Using 95% confidence level, does the sample mean
differ significantly from the population mean?

Based on the information from the problem in motivation, the

appropriate test statistic is z. The following information are also implied in
the problem:
 

 
Using the traditional method:

 
The computed z is greater than the critical value, 2.36 > 1.960. The
result suggests that the null hypothesis stating that there is no significant
difference between the sample mean and the population mean is rejected.
This implies that the difference between the means is significant and the
sample does not belong to the population.
 

Using the p-value method:

P(-2.36 < z < 2.36) = 0.4909 + 0.4909 = 0.9818

p-value = 1 – P(-2.36 < z < 2.36) = 1 – 0.9818 = 0.0182
 
The p-value is less than the alpha level, 0.0182 < 0.05.
The result suggests that the null hypothesis stating that there is
no significant difference between the sample mean and the
population mean is rejected. This implies that the difference
between the means is significant and the sample does not
belong to the population.
 
 EXAMPLE 2:
Drinking water has become an important concern among people.
The quality of drinking water must be monitored as often as possible
during the day for possible contamination. Another variable of concern is
the pH level, which measures the alkalinity or the acidity of the water. A
pH below 7.0 is acidic while a pH above 7.0 is alkaline. A pH of 7.0 is
neutral. A water-treatment plant has a target pH of 8.0. Based on 16
random water samples, the mean pH and their standard deviation were
7.6 and 0.4 respectively. Does the sample mean provide enough
evidence that it differs significantly from the target mean at two-tailed
distribution under = 0.05?

Since the sample size is small, t-test will be used.

 
= 7.6 = 8.0 = 0.4 n = 16 = 0.05 df = 15
t critical values = 2.131
 
Null hypothesis: There is no significant difference between the mean pH
of water samples and the target pH.
 
Alternative hypothesis: There is a significant difference between the
mean pH of water samples and the target pH.
 

Using the traditional method:

The computed t is less than the critical value, - 4 < - 2.131. The
result suggests that the null hypothesis stating that there is no
significant difference between the mean pH of water samples and the
target pH is rejected. This implies that there is sufficient evidence to
reject the null hypothesis. In addition, the result shows that the sample
does not belong to the population.
 

Using the p-value method:

 
Based on the t-table, it is not possible to identify the exact two
tailed alpha error for t = - 4 at df = 15. However, it is safe to say that
the p-value corresponding to (t = - 4) < 0.01. Since 0.01 < 0.5, then p-
value (t = - 4) < 0.05.
 
The p-value of t = -4 is less than the alpha level, p-value (t = -
4) < 0.05. The result suggests that the null hypothesis stating that
there is no significant difference between the mean pH of water
samples and the target pH is rejected. This implies that there is
sufficient evidence to reject the null hypothesis. In addition, the result
shows that the sample does not belong to the population.
 
 

EXAMPLE 3:

A sample of five measurements, randomly selected from an

approximately normally distributed population, resulted in the summary
statistics = 4.6 and = 1.5. Test the null hypothesis that the mean of the
population is 6 against the alternative hypothesis < 6 at = 0.05.
 
= 4.6 = 6.0 = 1.5 n=5 = 0.05 df = 4
t critical values = 2.132
 
Null hypothesis: There is no significant difference between the mean of the
samples and of the population.
 
Alternative hypothesis: The mean of the samples is less than the mean of
the population.
 
 

Using the traditional method:

 
The computed t is greater than the critical value, - 2.087 > -
2.132. The result suggests that the null hypothesis stating that
there is no significant difference between the mean of the
samples and of the population is accepted. This implies that
there is no sufficient evidence to reject the null hypothesis. In
addition, the result shows that the sample belongs to the
population with mean of 6.
 

Using the p-value method:

 
The alpha error for t = - 2.087 at df = 4 cannot be determined from the table.
However, it can be concluded that p-value (t = - 2.087) > 0.05.
 
The p-value of t = - 2.087 is greater than the alpha level, p-
value (t = - 2.087) > 0.05. The result suggests that the null hypothesis
stating that there is no significant difference between the mean of the
samples and of the population is accepted. This implies that there is no
sufficient evidence to reject the null hypothesis. In addition, the result
shows that the sample belongs to the population with mean of 6.
 EXAMPLE 4:
The owner of a factory that sells a particular bottled fruit juice
claims that the average capacity of their product is 250 ml. To test the
claim, a consumer group gets a sample of 100 such bottles, calculates
the capacity of each bottle, and then finds the mean capacity to be 248
ml. The standard deviation is 5 ml. Is the claim true at one-tailed = 0.05?

= 248 ml = 5 ml = 250 ml n = 100 = 0.05

Null hypothesis: There is no significant difference between the mean

capacities of the sample bottled fruit juice and of the production
claim.
 
Alternative hypothesis: The mean capacity of the sample bottled fruit
juice is less than that of the production claim.
 
 EXAMPLE 4:
The owner of a factory that sells a particular bottled fruit juice
claims that the average capacity of their product is 250 ml. To test the
claim, a consumer group gets a sample of 100 such bottles, calculates the
capacity of each bottle, and then finds the mean capacity to be 248 ml. The
standard deviation is 5 ml. Is the claim true at one-tailed = 0.05?

Using z-test:
 
z critical value = -1.645

 
The computed z is less than the critical value, - 4 < - 1.645. The
result suggests that the null hypothesis stating that there is no significant
difference between the mean capacities of the sample bottled fruit juice
and of the production claim is rejected. This implies that the difference
between the samples and the production claim is significant. In addition,
the result shows that the claim is not true.
 Using t-test:
 
df = 99 t critical value = - 1.645
 
t
 
The computed t is less than the critical value, - 4 < - 1.645.
The result suggests that the null hypothesis stating that there is no
significant difference between the mean capacities of the sample
bottled fruit juice and of the production claim is rejected. This implies
that the difference between the samples and the production claim is
significant. In addition, the result shows that the claim is not true.
EXAMPLE 5:

Study the tables below and supply what is asked in the item.

The drug 6-mP (6-mercaptopurine) is used to treat leukemia. Remission time, measured
in weeks, refers to the period where no leukemic cells can be found in the blood or bone
marrow and the bone marrow is working normally again.

Table 1. The Mean Remission Time of the Sample Group Using 6-mP

N Mean Std. Deviation Std. Error Mean

Remission Times 21 17.0952 9.99952 2.18207
 Find the following:

1. population mean
2. sample mean
3. sample standard deviation
4. sample size
5. null hypothesis
6. alternative hypotheses (directional and non-directional)
7. t-value or computed t
8. confidence level
9. alpha error
10.two-tailed p-value
11.one-tailed p-value
12.two-tailed decision for the null hypothesis
13.one-tailed decision for the null hypothesis
14.two-tailed interpretation of results
The SPSS (Statistical Package for the Social
Sciences) is a comprehensive system for analyzing data.
SPSS can take data from almost any type of file and use
them to generate tabulated reports, charts, and plots of
distributions and trends, descriptive statistics, and
complex statistical analysis. It is presently known as
PASW (Predictive Analytics SoftWare). z-test cannot be
done in SPSS. The results of SPSS are compared to the
alpha error.
Table 1. The Mean Remission Time of the Sample Group Using 6-mP

N Mean Std. Deviation Std. Error Mean

Remission Times 21 17.0952 9.99952 2.18207

Table 2. One-Sample t-test of the Remission Times of the Sample Group and Population
Using 6-mP

Test Value = 12.5

99% Confidence Interval of the
Difference
T df Sig. (2-tailed) Mean Difference Lower Upper
Remission Times 2.106 20 .048 4.59524 -1.6135 10.8040

1. population mean = 12.5 weeks

2. sample mean = 17.0952 weeks
3. sample standard deviation = 9.99952 weeks
4. sample size = 21
Table 1. The Mean Remission Time of the Sample Group Using 6-mP

N Mean Std. Deviation Std. Error Mean

Remission Times 21 17.0952 9.99952 2.18207

Table 2. One-Sample t-test of the Remission Times of the Sample Group and Population
Using 6-mP

Test Value = 12.5

99% Confidence Interval of the
Difference
T df Sig. (2-tailed) Mean Difference Lower Upper
Remission Times 2.106 20 .048 4.59524 -1.6135 10.8040

5. null hypothesis: There is no significant difference between the mean remission time
of the sample group and the mean remission time of all individuals use 6-mP.
6. alternative hypotheses (directional and non-directional)
Non-directional: There is no significant difference between the mean
remission time of the sample group and the mean remission time of all
individuals use 6-mP.
Right-tailed: The mean remission time of the sample group is significantly
higher than the mean remission time of all individuals use 6-mP.
Left-tailed: The mean remission time of the sample group is significantly
lower than the mean remission time of all individuals use 6-mP.
7. t-value or computed t = 2.106
8. confidence level = 99%
9. alpha error = 0.01
10. two-tailed probability = 0.048
11. one-tailed probability = 0.024 half of the two-tailed probability
12. two-tailed decision for the null hypothesis
Accept null hypothesis
p-value and alpha error: 0.048 > 0.01. To validate, compare the computed
t-value with the t critical value: p-value corresponding to t=2.106 > 2.845
13. one-tailed decision for the null hypothesis
Accept null hypothesis
p-value and alpha error: 0.024 > 0.01. To validate, compare the computed t-
value with the t critical value: p-value corresponding to t= 2.106 > 2.528.
14. two-tailed interpretation of results
The results show that the p-value of the mean remission time of the
sample group is greater than the alpha error, 0.048 > 0.01. The results suggest
that the null hypothesis stating that there is no significant difference between the
mean remission time of the sample group and the mean remission time of all
individuals use 6-mP is accepted. This implies that there is no sufficient
evidence to reject the null hypothesis. Thus, the mean remission time of the
sample group is similar with that of all the individuals who take 6-mP.