Okay.

So now welcome to the 8th lecture of empirical research methods in medical engineering.

The topic is hypothesis testing and significance tests.

So first we're going to talk about hypothesis testing using classical significance test

in general overview overall how this works.

Then we're going to talk about effect and effect sizes.

And then we're going to discuss the students T test which is a, you can also be used as

a significance test in more detail.

Now we start with the hypothesis testing using classical significance tests.

The procedure to start a test statistic would be the formulation of a null hypothesis

and a so-called alternative hypothesis.

In a moment I will explain what that means.

Then we're choosing a proper test.

Then we calculate the critical region to the significance level of alpha that we just

talked about.

Then we have the calculation of the test value.

For example the T test from the samples and then we make a decision whether it is part

of the critical region then the null hypothesis is not discarded or if the T test is an element

of the critical region the h null or the null hypothesis is discarded in favor of the

alternative hypothesis.

So for each alternative hypothesis the theory which h1 the theory based existence often

is the direction or sometimes the multitude of a population effect and null hypothesis

is set up which negates the relevant effect h1 and h0 so the alternative hypothesis and

the null hypothesis together form a pair of hypothesis that includes all possible manifestations

of the effect under consideration.

So the example that we will have on the next slide be the you could have the alternative

hypothesis that there is a positive relationship between stress at work among employed people

and the null hypothesis for this pair would be that there is no connection or even a negative

connection between stress at work and the absenteeism among employed persons.

So they are different hypothesis types and we are going to talk about the definitions

and the examples.

The interesting was the example that we just had so we have alternative hypothesis verbably

null hypothesis.

The incentive hypothesis postulates the existence often the direction and sometimes even the

magnitude of a specific effect in the population.

Mostly the research hypothesis is an alternative hypothesis so that is your hypothesis what you

want to find out and you formulate it in this way and the null hypothesis negates the

effect postulated by the alternative hypothesis and claims that there is no or the opposite

effect in this population.

Hi.

So the example would be there is a positive relationship between stress at work.

It is a positive relationship between stress at work among employed people and the null

hypothesis to that would be there is no connection or even a negative connection between stress

at work and absenteeism among employed persons.

This is actually the both of these are actually examples of verbal hypothesis.

So hypothesis are first formulated verbally in terms of content.

The verbal or content verbal hypothesis are to be converted into statistical hypothesis

which contain the relevant population parameters.

Mostly they have the parameters and they have the concrete numbers in them.

So the hypothesis can be rejected or not.