# Test for Differences Between Means: Paired Data

When testing for the difference between two population means, independent samples may introduce some caveats. The real difference can be masked by variability within each sample caused by factors other than those of interest to the researcher. Random variation within a sample can be so large that it blurs the real difference – that which has been caused by a phenomenon or factor in which an analyst is interested. For this reason, a paired data test may be preferred.

As a way of controlling the external variability, data analysts often collect a pair of responses from each subject. They then compute and work with the differences within pairs.

### Example 1: Paired Data Test

A group of financial analysts wishes to determine whether the payment of dividends has a significant effect on the share price. Sample data on share prices can be collected from a set of companies prior to the payment of dividends, say, 7 days before dividend announcements. To ensure the samples are independent, the companies should be in the same line of business. This would ensure that all of them are susceptible to similar specific risks and other business-related conditions. Seven days after all the companies have paid the dividends, the analysts can recollect data from each company. The analysts would end up with paired data from each company. Such data can then be used to establish if any significant differences exist between pre-dividend and post-dividend share prices.

The overriding assumption during paired data tests is that the observed differences constitute a sample from a normal distribution.

Basic Situation

H0: μD (= μ1 – μ2) = σ

The test statistic is

$$\frac { \bar { D } -\sigma }{ \frac { { S }_{ D } }{ \sqrt { n } } } \sim { t }_{ n-1 } \text{ under } { H }_{ 0 }$$

In case n is large, the normal distribution assumption is not necessary.

### Example 2: Paired Data Test

Doctors in a certain general hospital wish to determine if a special diet can reduce body weight. To do this, they select ten overweight patients with an average body weight of 77 kg (B’). After being subjected to the special diet, the average weight reduces to 75 kg (A’).

Test whether the special diet reduces body weight.

Additional information: $$\sum (B_i – A_i)^2 = 70 \quad \quad i = 1, 2, 3, …, 9, 10.$$

Solution

As always, begin by stating the hypothesis:

H0: μA = μB

H1: μA < μB

Where A stands for “after” and B is “before.”

The test statistic is:

$$\frac { \bar { D } -\sigma }{ \frac { { S }_{ D } }{ \sqrt { n } } } \sim { t }_{ n-1 }\quad under\quad { H }_{ 0 }$$

Where Di is the difference within each pair (Ai – Bi).

\begin{align*} D’ & = A’ – B’ = 75 – 77 = -2 \\ S_D^2 & =\cfrac {1}{(n – 1)} \left( \sum D_i – D’ \right)^2 =\cfrac {1}{(n – 1)} \left\{ \sum D_i^2 – {nD’}^2 \right\} \\ & =\cfrac {1}{9} \left\{ 70 – 10(-2^2) \right\} = 3.333 = 1.826^2 \end{align*}

Therefore,

$$\text{Test statistic} =\cfrac {(75 – 77)}{\left( \frac {1.826}{\sqrt{10}} \right)} = -3.591$$

The test statistic (-3.591) is lower than the lower 5% point of the t-distribution at 9 degrees of freedom (-1.833).

This gives us sufficient evidence to reject the H0 at 5% significance. Therefore, it would be very reasonable to conclude that the special diet reduces body weight.

## Question

Repeat the example above using a significance level of just 0.05% and state the decision rule.

A. Reject the H0 and conclude that the special diet reduces body weight.

B. Do not reject the H0 and conclude that the special diet has an insignificant effect on body weight.

C. Reject the H0 and conclude that the special diet has an insignificant effect on body weight.

Solution

The observed test statistic still remains -3.591. The lower 0.05% point of the t9 distribution is -4.781.

It is clear that the test statistic is way above the t9 distribution, i.e., it lies outside the critical or rejection region. Therefore, we would not reject the H0.

Note to candidates: Reducing the significance level basically reduces the probability of rejecting a true null hypothesis (type I error).

Shop CFA® Exam Prep

Offered by AnalystPrep

Featured Shop FRM® Exam Prep Learn with Us

Subscribe to our newsletter and keep up with the latest and greatest tips for success

Sergio Torrico
2021-07-23
Excelente para el FRM 2 Escribo esta revisión en español para los hispanohablantes, soy de Bolivia, y utilicé AnalystPrep para dudas y consultas sobre mi preparación para el FRM nivel 2 (lo tomé una sola vez y aprobé muy bien), siempre tuve un soporte claro, directo y rápido, el material sale rápido cuando hay cambios en el temario de GARP, y los ejercicios y exámenes son muy útiles para practicar.
diana
2021-07-17
So helpful. I have been using the videos to prepare for the CFA Level II exam. The videos signpost the reading contents, explain the concepts and provide additional context for specific concepts. The fun light-hearted analogies are also a welcome break to some very dry content. I usually watch the videos before going into more in-depth reading and they are a good way to avoid being overwhelmed by the sheer volume of content when you look at the readings.
Kriti Dhawan
2021-07-16
A great curriculum provider. James sir explains the concept so well that rather than memorising it, you tend to intuitively understand and absorb them. Thank you ! Grateful I saw this at the right time for my CFA prep.
nikhil kumar
2021-06-28
Very well explained and gives a great insight about topics in a very short time. Glad to have found Professor Forjan's lectures.
Marwan
2021-06-22
Great support throughout the course by the team, did not feel neglected
Benjamin anonymous
2021-05-10
I loved using AnalystPrep for FRM. QBank is huge, videos are great. Would recommend to a friend
Daniel Glyn
2021-03-24
I have finished my FRM1 thanks to AnalystPrep. And now using AnalystPrep for my FRM2 preparation. Professor Forjan is brilliant. He gives such good explanations and analogies. And more than anything makes learning fun. A big thank you to Analystprep and Professor Forjan. 5 stars all the way!
michael walshe
2021-03-18
Professor James' videos are excellent for understanding the underlying theories behind financial engineering / financial analysis. The AnalystPrep videos were better than any of the others that I searched through on YouTube for providing a clear explanation of some concepts, such as Portfolio theory, CAPM, and Arbitrage Pricing theory. Watching these cleared up many of the unclarities I had in my head. Highly recommended.