ANOVA

Loading...

來自 University of Cape Town 的課程

Understanding Clinical Research: Behind the Statistics

729 個評分

University of Cape Town

Understanding Clinical Research: Behind the Statistics

729 個評分

If you’ve ever skipped over`the results section of a medical paper because terms like “confidence interval” or “p-value” go over your head, then you’re in the right place. You may be a clinical practitioner reading research articles to keep up-to-date with developments in your field or a medical student wondering how to approach your own research. Greater confidence in understanding statistical analysis and the results can benefit both working professionals and those undertaking research themselves. If you are simply interested in properly understanding the published literature or if you are embarking on conducting your own research, this course is your first step. It offers an easy entry into interpreting common statistical concepts without getting into nitty-gritty mathematical formulae. To be able to interpret and understand these concepts is the best way to start your journey into the world of clinical literature. That’s where this course comes in - so let’s get started! The course is free to enroll and take. You will be offered the option of purchasing a certificate of completion which you become eligible for, if you successfully complete the course requirements. This can be an excellent way of staying motivated! Financial Aid is also available.

從本節課中

Which test should you use?

The most common statistical test that you might come across in the literature is the t-test. There are, in actual fact, a few t-tests, but the one most are familiar with, is of course, Student’s t-test and its ubiquitous p-value. Not everyone, though, knows that the name Student was actually a pseudonym, used by William Gosset (1876 - 1937). Parametric tests have very strict assumptions that must be met before their use is justified. In this lesson we take a closer look at these tests, but perhaps more importantly, their strict assumptions. Once you know these, you will be able to identify when these tests are used inappropriately.

Introduction to parametric tests2:07

Student's t-test15:08

Linear Regression4:00

(Optional) Student's t-test in action12:36

與講師見面

Juan H Klopper
Dr
Department of Surgery

Now we've spoken about student's t-test and everything really follows from that.

If you really understand how that works, how the central limit theorem works,

how we are just converting everything to so

many units of standard error, and what the area under the curve represents,

you really understand most of statistical analysis.

Now, student's t-test, and all the t-tests, really for comparing the means

of two groups, but sometimes we want to compare the means of more than two groups.

And that is when we have Anova, the most common one is called a one-way Anova.

The one-way refers to the fact that we are only dealing with one factor.

So imagine, we're just looking at comparing the means of white cell counts

between four or five groups of participants.

So it's just the white cell count, that's the only thing we're looking at,

that is just one factor.

And also has some levels to it that just really refers to how many groups we ask.

If it's a four group, that's really a four level, one-way Anova if we look at that.

Now t-test, remember we are comparing the means of two groups,

but really over ANOVA three or more groups we are comparing to each other.

Now all the assumptions we spoke about when we spoke about student's t-test

has to be met.

The ANOVA is still a parametric test, so we have to have those same assumptions.

Most notably, the data points must come from, or any variable, must come from

a variable that has a normal distribution in the patient or in the population.

The actual data point values, they must be from numerical type and

they must be a continuous data type.

We also have to have roughly equal variances between the values

in each group.

If we work out the variances, standard deviation for each individual group,

they've gotta be roughly the same.

And also these participants really have to be independent.

We can't be dealing with one is a called it triplets for instance, as an example.

Now what the Anova does,

it only determines that there's a difference between the groups.

It doesn't highlight one group over any of the other groups,

two groups over any of the other groups.

It can not do that.

It's only going to tell you that there's some fundamental difference between

the groups but there's no individual group analysis that goes on.

Now, it uses what we call the F-distribution.

We haven't spoken about the F-distribution before,

but it's a distribution, just as any other.

Doesn't look normally distributed, though.

It has this positive skew, so it's got this tail going out towards the right.

And it is shaped by the degrees of freedom, just like the t-test,

the T-distribution I should say, and it has a mean of around one.

We're going to do exactly the same thing though,

there's going to be some critical values.

So many standard errors away from the mean would represent an area under the curve of

open 0.05.

And the differences between all of those groups is gonna be converted to an

F-value and that F-value is gonna fall either side of this critical value.

Everything would be the same.

Here you see a beautiful example of random variables taken from,

I think it was about 10,000 random variables, taken from an F-distribution.

And you can beautifully see the mean is around one, but there is this tail that

goes off to the right, there's quite a bit of skewness to the F-distribution.

So what the Anova test really does, it compares the intergroup,

the within group variability to the between group variability.

And if it finds that the between group variability is much bigger than

the intergroup variability, the variability of the data values within each

group, then it really tends towards an f-value that'll be right

far away from the critical value that will have statistical significance in some way.

It's kind of the rule of thumb.

Now, there are really many Anovas.

You can well imagine from what I've discussed now, how many groups there are,

how many values but different groups there are, how they interact with each other,

how many factors we are dealing with is a lot of different types of the Anova.

In essence though, we are comparing the means of more than two groups.

探索我們的目錄

免費加入並獲得個性化推薦、更新和優惠。

Coursera 致力於普及全世界最好的教育，它與全球一流大學和機構合作提供在線課程。

© 2018 Coursera Inc. 保留所有權利。

通過 App Store 下載

通過 Google Play 獲取