Scribd
Upload
710 views

T Distibution

The Student's t-distribution is a probability distribution used to estimate population parameters when sample sizes are small (less than 30) and the population variance is unknown. It was developed by William Gosset under the pseudonym "Student" and is called the Student's t-distribution. The t-distribution is similar to the normal distribution but has heavier tails, meaning it is more prone to extreme values. As the degrees of freedom increase, the t-distribution approaches the normal distribution.

Uploaded by

Fritz Astillo
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
710 views

T Distibution

The Student's t-distribution is a probability distribution used to estimate population parameters when sample sizes are small (less than 30) and the population variance is unknown. It was developed by William Gosset under the pseudonym "Student" and is called the Student's t-distribution. The t-distribution is similar to the normal distribution but has heavier tails, meaning it is more prone to extreme values. As the degrees of freedom increase, the t-distribution approaches the normal distribution.

Uploaded by

Fritz Astillo
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 25

Definitions.

The Student’s t-distribution is a probability distribution that


is used to estimate population parameters when the sample
size is small (𝑖. 𝑒. 𝑠𝑎𝑚𝑝𝑙𝑒 𝑠𝑖𝑧𝑒 < 30) and/or when the
population variance is unknown. It was developed by
William Sealy Gosset in 1908. He used the pseudonym or pen
name “Student” when he published his paper which describes
the distribution. That is why it is called “Student’s t-
distribution”. He worked at a brewery and was interested in
the problems of small samples, for example, the chemical
properties of barley. In the problem he analyzed, the sample
size might be as low as three.
Suppose you are about to draw a random sample of n observations
from a normally distributed population, to construct a confidence
interval for the population mean, 𝜇.
𝑥−𝜇
𝑧 =
𝜎
Previously,
𝑥ҧ − 𝜇
𝑧=
𝜎/ 𝑛
Instead of using the “population” standard deviation, 𝜎; you are
going to use your “sample” standard deviation s, to estimate it.
When you are sampling from a normally distributed population,
the quantity
𝑥ҧ − 𝜇
𝑡=
𝑠/ 𝑛
has the t-distribution with n-1 degrees of freedom.
The exact shape of the t-distribution depends on the degrees of freedom. The figure above tells you
that as the degrees of freedom increase, the t-distribution tends toward the standard normal
distribution. At 30 degrees of freedom, the blue curve might look very close to the normal curve. But
if you look very closely, you will see that the t-distribution still has slightly heavier tails and slightly
lower peak. But if you let those degrees of freedom continue to increase, the t-distribution is going to
get closer and closer to the standard normal distribution.
Properties of t-distribution
The t-distribution has the following properties:
1. The t-distribution is symmetrical about 0. That means if you draw
a segment from the peak of the curve down to the 0 mark on the
horizontal axis, the curve is divided into two equal parts or areas.
The t- scores on the horizontal axis will be divided also with half of
the t-scores being positive and half negative.
Properties of t-distribution
2. The t-distribution is bell-shaped like the normal distribution but
has heavier tails. That means it is more prone to producing values
that fall far from the mean. The tails are asymptotic to the horizontal
axis. (Each tail approaches the horizontal axis but never touches
it.)
Properties of t-distribution
3. The mean, median, and mode of the t-distribution are all equal
to zero.
𝑣
4. The variance is always greater than 1. It is equal to where 𝑣 is
𝑣−2
the number of degrees of freedom. As the number of degrees of
freedom increases and approaches infinity, the variance
approaches 1. Using the formula, if the number of degrees of
10 10
freedom is 10, the variance is = = 1.25.
10−2 8
Properties of t-distribution
5. As the degrees of freedom increase, the t-distribution curve
looks more and more like the normal distribution. With infinite
degrees of freedom, t-distribution is the same as the normal
distribution.
Properties of t-distribution
6. The standard deviation of the t-distribution varies with the
sample size. It is always greater than 1. Unlike the normal
distribution, which has a standard deviation of 1.
7. The total area under a t-distribution curve is 1 or 100%. One can
say that the area under the t-distribution curve represents the
probability or the percentage associated with specific sets of t-
values.
Properties of t-distribution
8. The exact shape of the t-distribution depends on the degrees of
freedom. Remember that the lesser the degree of freedom, the
lower is its peak and the thicker is its tails. As the degree of
freedom increases, the tails become flatter, and the peak becomes
higher. That means, given the area or the probability, the t-value
depends on the number of degrees of freedom. For example, with
the given area of 0.05 on the right tail of t-distribution, the t-value is
2.015 with 5 degrees of freedom. But with 20 degrees of freedom,
the t-value is equal to 1.725.
Here is an example of a t-table. It is a right-tailed t-table because
the given areas in this table are areas on the right tail of the t-
distribution. Some t-tables are slightly different in format. Look at
the t-table below. In the first column in the left-most part, you have
the degrees of freedom. It ranges from 1 down to ∞. While the first
row in the upper part of the t-table represents the area under the
right tail of the t-distribution. Some of the given areas are from 0.25
down to 0.0005. The rest of the entries in the body of the table are
the values of the variable t (t-values).
Identifying Percentiles Using the t-Table
A percentile is a value on a t-distribution that is less than the probability in the given percentage. For
example, the 90th percentile of the t-distribution is that t-value whose left tail probability is 90% and
whose right-tail probability is 10%. Since the area under the t-distribution curve also represents the
probability, the 90th percentile of the t-distribution is the t-value whose area on its left tail is 0.90 and
whose area on its right tail is 0.10.
Example 1
Find the 95th percentile of a t-distribution with 6 degrees of
freedom.
Example 2
What is the area to the right of 2.4 under a t-distribution with 7
degrees of freedom?

So, using the table you found that the area to the right of 2.4 under the t-distribution
with 7 degrees of freedom lies somewhere between 0.02 and 0.025.
If you need to get the exact value, you need to use software that easily calculates the area
under the t-distribution curve with the given t-value and number of degrees of freedom.
Using such software, you could find that the area to five decimal places is 0.02373.
Example 2
What if you needed to use the t-table to find the area to the left of
2.4?
▪ Since the area under the entire curve is 1, the area to the left of 2.4 is equal to 1 minus the area to
the right of 2.4. So, based on the table the area to the left of 2.4 under the t distribution with 7
degrees of freedom must lie somewhere between 0.98 and 0.975 (1 – 0.02 = 0.98 and 1 – 0.025 =
0.975).
▪ But since you already knew that the area to the right of 2.4 is 0.02373, you could find the exact area
to the left of 2.4 to five decimal places as 1 minus 0.02373 or 0.97627.
Definitions.
The Confidence level of an interval estimate of a parameter is the
probability that the interval estimate contains a parameter, it
describes what percentage of intervals from many different
samples contains the unknown population parameter.
The confidence level has its corresponding coefficient which is
called confidence coefficients or critical values. These
coefficients are used to find the margin of error, for instance, the
table below shows the corresponding coefficient confidence level.
Definitions.
Confidence interval or interval estimate is a range of values that
is used to estimate a parameter. This estimate may or may not
contain the true parameter value. For instance, we write it in this
form
Lower limit < μ< Upper limit
or
(Lower limit, Upper Limit)
The Lower limit is obtained by using the formula 𝐿𝐿 = 𝑥ҧ − 𝐸, while
the Upper limit is obtained by using the formula 𝑈𝐿 = 𝑥ҧ + 𝐸, where
𝐸 is the margin of error and 𝑥ҧ is the sample mean.
Definitions.
As mentioned earlier, the confidence coefficient is used on finding
the margin of error, which is the range of values above and below
the sample statistic. For instance, Margin of error is obtained using
the formula:
𝜎
𝐸 = 𝑧𝛼/2 •
𝑛
where, 𝑛 = 𝑠𝑎𝑚𝑝𝑙𝑒 𝑠𝑖𝑧𝑒
𝑧𝛼 = 𝑐𝑜𝑛𝑓𝑖𝑑𝑒𝑛𝑐𝑒 𝑐𝑜𝑒𝑓𝑓𝑖𝑐𝑖𝑒𝑛𝑡
2
𝜎 = 𝑝𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛 𝑠𝑡𝑎𝑛𝑑𝑎𝑟𝑑 𝑑𝑒𝑣𝑖𝑎𝑡𝑖𝑜𝑛
𝐸 = 𝑚𝑎𝑟𝑔𝑖𝑛 𝑜𝑓 𝑒𝑟𝑟𝑜𝑟
The length of the confidence interval is given by L=2E or L=UL-
LL.
Example 1
A random sample of 46 scores from the examination of ABM learners is taken and it
gives a sample mean of 78 with 0.82 margin of error having 90% level of
confidence.
What is the upper limit? What is the lower limit?
Upper Limit
𝑈𝐿 = 𝑥ҧ + 𝐸 = 78 + 0.82 = 78.82
Lower Limit
𝐿𝐿 = 𝑥ҧ − 𝐸 = 78 − 0.82 = 77.18
What is the confidence interval in the given statement?
77.18 < 𝜇 < 78.82
(77.18,78.82)
How will you conclude?
We can be 90% confident that the true mean lies between 77.18 and 78.82.
What is the length of the confidence interval?
Length of the confidence interval = 2𝐸 = 2 0.82 = 1.64
Example 2
Isabel owns a shoe store. She used 160 pairs of shoes as her samples for the
different designs. The population standard deviation of the price of the shoes is ₱75.
Suppose that Isabel wants a 95% level of confidence to determine the mean price of
all her shoes she is selling. Compute for the margin of error of her estimate and the
length of the confidence interval.
Sample Size Determination
To make an accurate estimate of the parameter, it is vital to
determine the appropriate size of the sample. The factors that
affect the sample size are the maximum error of estimate, the
population standard deviation, and the confidence level.
You must also consider the table below in computing for an
appropriate sample size.
Example 3
Compute for the appropriate sample size if the margin of error is 6,
with a sample standard deviation of 25 and a 90% level of
confidence.
Given:
For a 90% confidence level, 𝑧𝛼/2 = 1.645
𝐸 = 6 and 𝜎 = 25
𝑧𝛼/2 ∙𝜎 2
Formula: 𝑛 =
𝐸
1.645∙25 2
Substitute: 𝑛 =
6
Solve: 𝑛 = 46.98
Evaluate: 𝑛 = 47
Example 4
A HUMSS student is conducting a research on the awareness of SHS learners on the
names of the Philippine Presidents. Given a margin of error of 4 with a standard
deviation of 9 having a 99% confidence level. What should be the minimum sample
size that the student-researchers need to estimate the population mean?
Solution:
Confidence level is 99% then
𝑧𝛼 = 2.58 𝐸=4 𝜎=9
2

𝑧𝛼/2 ∙𝜎 2
Formula: 𝑛 =
𝐸
(2.58)(9) 2
Substitute: 𝑛 =
4
Solve: 𝑛 ≈ 33.70
Evaluate n=34
Conclusion: Therefore, to be 99% confident that the estimate differs from the true mean by 4, the
minimum sample size needed is at least 34.
Note: When determining sample size, we always round off the resulting value to the next whole
number.
Example 5
An Environmentalist is making a survey on the number of households that applies
the principles of Waste Management in a certain community. Given the standard
deviation of 2.6, how large a sample must be selected if we want to be 95%
confident of finding whether the true mean differs from the sample mean by 1.5
hour?
Solution:
Confidence level is 99% then
𝑧𝛼 = 1.96 𝐸 = 1.5 𝜎 = 2.6
2

𝑧𝛼/2 ∙𝜎 2
Formula: 𝑛 =
𝐸
(1.96)(2.6) 2
Substitute: 𝑛 =
1.5
Solve: 𝑛 ≈ 11.54
Evaluate 𝑛 = 12
Conclusion: Therefore, to be 95% confident that the estimate differs from the true mean by 1.5, the
minimum sample size needed is at least 12.

You might also like