Math Is Fun Forum

pod33

Member

Registered: 2012-04-05

Posts: 4

A time and a place for every test?

Hello
I'm a podiatry student working on a simple data analysis assignment for class. I'm not 100% sure if this forum is cool with students asking for help with their homework - I know that the podiatry forums hate it! - so if I'm making a bit of a faux pas, then I'm very sorry about that and please ignore my post.

Otherwise here is my predicament:

We have to look at some data and determine whether pulse is affected by exercise.

I have a group of people in my sample (n=20) and 3 data sets - pulse when: sitting, standing and after running.

I want to get values for variability within the data sets, as the mean for each data set seems to go up as expected, but the actual data points are all over the shop. I want to get the variability values so I can justify investigating why the variability is so high and whether this variability could undermines my hypothesis - that exercise makes your pulse go up.

What would be the best test to use to determine the variability for each data set?   

I think I have probably come across the answer online, but the shear variety and complexity of tests out there has got me stumped.
Can anyone give me any pointers?

N.B - most of my fellow students are working with very very simple values like mean and SD. Im keen to go a bit further as I have  been offered a summer job at the uni working on a Phd research project and I really want to flex/build some stats muscle and I'm trying to use this assignment as a training ground. I really do like statistics, but I have a long way to go........

P.S nice selection of smileys here!

bobbym

bumpkin

From: Bumpkinland

Registered: 2009-04-12

Posts: 109,606

Re: A time and a place for every test?

Hi;

Welcome to the forum. You might try a correlation coefficient.

---

In mathematics, you don't understand things. You just get used to them.
If it ain't broke, fix it until it is.
Always satisfy the Prime Directive of getting the right answer above all else.

Bob

Administrator

Registered: 2010-06-20

Posts: 10,621

Re: A time and a place for every test?

hi Max,

Standard deviation does measure variability, so it makes sense to calculate the three values along with the means.

I suppose that what you then want to know is

'Could the exercise data have come from the same population as the sitting data.'

Let's say you calculate a new variable:  the increase in pulse rate value = (exercise rate - sitting rate) for each of the 20 results.

Null hypothesis is these results come from a population with mean = zero.

Take n = 20 as a sample from that population.

Calculate the sample mean and sd.

Calculate how likely is it that that you get this sample from the population.

My statistics book "Advanced Level Statistics", by A Francis, says you should use a T-test because you don't know the population variance.

Estimate it with

where n = 20 and s is the sample variance.

So you compute

where xbar is the sample mean and s the sample variance, and we are assuming 'mu' =zero.

Then you need some T tables.  I expect you can 'google' for them.  And the parameter is n-1 = 19

How does that sound?

Bob

ps.  yes, a great set of smilies.  But you are supposed to use them to indicate how you are feeling, so showing them all at once is a bit much. 

---

Children are not defined by school ...........The Fonz
You cannot teach a man anything;  you can only help him find it within himself..........Galileo Galilei
Sometimes I deliberately make mistakes, just to test you!  …………….Bob

pod33

Member

Registered: 2012-04-05

Posts: 4

Re: A time and a place for every test?

Thanks! Yeah - smiely overkill! Sorry about that!

Last edited by pod33 (2012-04-06 04:51:13)

pod33

Member

Registered: 2012-04-05

Posts: 4

Re: A time and a place for every test?

Hello Bobs
I ran an independant T test using SPSS and got T values of
Sitting - 42.346
Standing - 47.519
exercise - 41.610

(P values of 0.00, does this mean there is almost certainly a a rise in pulse across variables? i.e reject my null hypothesis)

Im not sure how to interpret this information. I had a look at the T tables but I wasn't sure which T.x I should be using. Could you give me any pointers?

Im pretty confused to be honest. What I really want to do is get a P value that allows me to reject my null hypothesis that there will be no change in pulse across my independent variables. How do you think this is best achieved using SPSS? Or have I already achieved this with my sig of 0.00?

Thank you for giving me the formula to achieve these tests, but my teachers want us to use SPSS and provide output data.

Many thanks
Max

Last edited by pod33 (2012-04-10 04:45:20)