TEALE FENNING

rpwalavalkar · Teale Fenning Administrator Joined: 20 Jul 2006 Posts: 973

hi nick,

have tried and tried, but can't quite grasp the concept of 'analysis of variance' , came across a mcq and got all stems wrong. Sad

.

could you explain this please?

raj

Nick Raine-Fenning · Posted: Sun Jul 23, 2006 11:49 am Post subject:

It's quite simple.

ANOVA (Analysis of Variance) is a test used to see if there are any significant differences between the means of 3 or more variables.

It is essentially a t-test but for more than 2 groups.

The 'variance' part comes from the fact the statistical analysis considers the distribution of the data / values around the mean.

In medicine we often use t-tests as we only have two datasets - treatment and placebo or pregnant and not pregnant for example.

These tests assume the data are normally distibruted.

Does that help?

Why don't you post the question and we cna have a look at it?
_________________
"Teale Fenning Medical Education" delivering evidence-based, exam-orientated learning since 1997

rpwalavalkar · Teale Fenning Administrator Joined: 20 Jul 2006 Posts: 973

hi,

the mcq was

analysis of varience

1. is a method of comparing 3 or more groups

2. assumes similar varience in each group

3. uses the t distribution for significance testing

4. requires the calculation of a varience factor F

5. requires equal numbers in each group

Nick Raine-Fenning · Posted: Sun Jul 23, 2006 2:13 pm Post subject:

Goodness me you ask some difficult questions Wink

Analysis of variance

1. is a method of comparing 3 or more groups

True - see comments above

2. assumes similar variance in each group

True

The measurements / values should have a normal distribution and the same variance / standard deviation

3. uses the t distribution for significance testing

False

This relates to the t-test when comparing the means of two populations.

The F test (see below) is used.
This is an extension of the t test to more than two groups.

4. requires the calculation of a varience factor F

True

The varience factor F is a little more complex.

It is the ratio of the variances seen between and within groups. The larger the value of F the more signiifcant the result and thereofee the chance that the means of the groups differ.

There is a table of F values that the F value calculated in your population can be compared to. This includes the degrees of freedom available.

5. requires equal numbers in each group

False

It needs enough values in each group to allow assessment that the data are normally disributed but the exact number in each group often differs.
_________________
"Teale Fenning Medical Education" delivering evidence-based, exam-orientated learning since 1997

Nick Raine-Fenning · Posted: Sun Jul 23, 2006 2:13 pm Post subject:

Well - how did I do?

Did I get them right?
_________________
"Teale Fenning Medical Education" delivering evidence-based, exam-orientated learning since 1997

rpwalavalkar · Teale Fenning Administrator Joined: 20 Jul 2006 Posts: 973

ALL CORRECT !!!! Smile

Question -- can i fracture my arm and take you to the exam as my writer????? Wink

raj

Nick Raine-Fenning · Posted: Sun Jul 23, 2006 2:45 pm Post subject:

Of course ... I'll put a skirt on and let my hair grow Rolling Eyes

_________________
"Teale Fenning Medical Education" delivering evidence-based, exam-orientated learning since 1997

rpwalavalkar · Teale Fenning Administrator Joined: 20 Jul 2006 Posts: 973

Ha Ha Ha

sasiselvam · Joined: 16 Aug 2006 Posts: 13 Location: india

Hi NIck,

Thanks for the crystal clear explanation for ANOVA.Did a few related MCQ and got the ans correct.
May not remember me. Had attended the Essay technique course.

Sasi

Nick Raine-Fenning · Posted: Sun Aug 20, 2006 3:52 pm Post subject:

Of course I remember you Very Happy

I am pleased the explanation helped you.

They are not easy these stats questions Razz

_________________
"Teale Fenning Medical Education" delivering evidence-based, exam-orientated learning since 1997