Biostatistics & Research Methodology
2021/22
Block 6 – Lecture 1
Student’s T
Distribution
Aktham Osama Abdulazeez, MBChB
Objectives
o Shed some light on aspects of hypothesis testing and its role in
research.
o Define T distribution and outline its differences from Z distribution.
o Outlines the statistical scenarios where use of the T distribution is
prudent.
o Understand the concept of degrees of freedom.
o Highlight the difference in analysis of individual & paired data using
different variations of the T test.
Introduction
o Probability in normally distributed populations can be
calculated using two parameters: population mean (𝜇) and
population standard deviation (σ)
o When the population variance (and hence standard
deviation) of a population is unknown, we can use the sample
variance (S²) as the best point estimate for the population
variance σ².
o In such cases, The distribution will not follow the standard
normal distribution (Z distribution), but will follow the
Student’s T distribution instead.
Student’s T Distribution
Characteristics of Student’s T Distribution
o It has a mean of zero.
o It is symmetric around the mean.
o It ranges between ‐ ∞ and +∞.
o Compared to the standard normal distribution, the curve is
less peaked with higher tails (a fat curve).
o The quantity n‐1, which is called the degrees of freedom (df),
is used in computing the sample variance.
o The t‐distribution approaches the standard normal
distribution as the degrees of freedom approach infinity.
Degrees of Freedom (For Pros)
“Maximum number of logically independent
values, which are values that have the freedom to
vary, in the data sample.”
Relevant in multiple statistical tests (T test, F test, Chi-square test, etc..)
Degrees of Freedom (For Newbies)
Degrees of Freedom in the Context of a t Test
Dataset 1
Value 1
Imagine a hypothetical Value 2
population with a data set of Value 3
10 entities. Value 4
Value 5
Value 6
Value 7
Value 8
Value 9
Value 10
Sum
Mean
Calculating a t Score
𝑥ҧ − 𝜇 𝑥ҧ − 𝜇
𝑍= 𝜎 𝑡=
ൗ 𝑛 𝑆ൗ
𝑛
𝑥ҧ = sample mean
𝜇 = population mean
𝜎 = population standard deviation 𝑆 = sample standard deviation
n = sample size
Exercise
A law school claims it’s graduates earn an average of $300 per
hour. A sample of 15 graduates is selected and found to have a
mean salary of $280 with a sample standard deviation of $50.
Assuming the school’s claim is true, what is the probability that
the mean salary of graduates will be no more than $280?
Confidence Intervals for a Single Sample
o Confidence Interval for the mean of a normal distribution
with unknown population variance, and a small sample size:
C.I. = Estimator ± R.C. x SE
o The reliability coefficient will be the t value corresponding to
the particular confidence level and the degree of freedom.
R.C. = t1‐α/2 , df=n‐1
T
Distribution
Table
Calculation of SE Depending on
Estimator (Review)
Exercise
Specimen obtained from 10 individuals reveal a mean and
standard deviation of calcium content of the teeth of 35.7, and
0.7 respectively. Find the 95% CI for the mean calcium content.
Confidence Intervals for 2 Sample Means
o Confidence interval for difference between two population
means when the population variances are unknown and
assumed to be unequal.
C.I. = (X1 ‐ X2) ± t1‐α/2, df=n1+n2‐2 x S.E.
Exercise
A breast CA research team
collected the following data
on tumor size.
Construct a 95% CI for the
difference between
populations means, assuming
unequal population
variances.
Exercise
Confidence Intervals for 2 Sample Means
o Confidence interval for difference between two population
means when the population variances are unknown but
assumed to be equal.
o We should first find the Pooled Variance S2p.
C.I. = (X1 ‐ X2) ± t1‐α/2, df=n1+n2‐2 x Sp√1/n1 +1/n2
Exercise
In a cancer institute a research was done on a drug to prolong life in
patients with throat Ca. 200 patients are randomly divided into two
equal groups, one kept on the drug and the other on a placebo. The
following was found:
Assuming normality, find the 95% CI for the difference in means, if the
variances are assumed to be equal.
Exercise
Paired Samples T Test
o Many studies are designed to produce observations in pairs
.i.e.: BP, RBS….before and after giving certain treatments
(Before & After readings).
o A measurement done by two instruments, individuals, times,
etc….
o Every individual here has a pair of readings.
o In such situations, you need to find the CI for the mean
difference.
Paired Samples T Test
o Find the difference 𝑑ҧ
o Find the mean of difference 𝑑
o Find the standard deviation of the difference Sd
o The df= n‐1 (since we have only one sample)
o Apply the formula:
CI = d ± t1‐α/2, df = n‐1 x Sd /√n
Exercise
In a pediatrics clinic, a study
was carried out to see the
effectiveness of a certain
antipyretic.
Twelve 4‐years old girls
suffering from influenza had
their temperatures measured
immediately before and 1-hour
after administering the drug.
For the following results, find
the 95% C.I. for the mean
difference.
Exercise
Exercise
Wrapping Up
Continuous
Data
Population
Variance known?
Yes No
Z test T Test*
Single Paired
2 Samples
Sample Data
Thank You
Questions? Ask in the group or note it down for our next
live session!
Hope to see you next time ☺
