VoG II 22:12 28 Jun 2006

Hi> I have 37 paired values. I want to check if there is a real difference between them. I have
used Student's t-test with a null hypothesis of 0, p=0.05, and assuming the same variance. Did I use the correct test - paired t-test?

  ade.h 22:15 28 Jun 2006

I genuinely wish that I could help, but the whooshing sound that you hear is your question going right over my head, VoG! LOL.

  wee eddie 22:44 28 Jun 2006

[email protected]?subject=Statistics Inquiry

or his page here

click here

I read the article and now know little more than I did when I started! But having read this one, I think that the skies are a little brighter.

click here

  VoG II 20:08 29 Jun 2006

Thanks guys - obviously no statisticians are members.

Anyway, they look the same to me click here and that's good enough till I can talk to my tame statistician :o)

  Forum Editor 20:46 29 Jun 2006

click here

Or it might not.

One of my clients is an oceanographic survey company. They blind me with equally incomprehensible stuff on a regular basis. Once, when pretending to understand what they were on about, I dropped my laptop over the side of a survey vessel into about 40 feet of water. It contained lots of data sets - they weren't amused, and neither was I. It's still down there.

Sorry I can't be helpful.

  VoG II 21:08 29 Jun 2006

for a very informative link.

It has reminded me that 'statistically significant' and 'scientifically important' are two different things. I can make a judgement about scientific importance - and there isn't any with this set of data.

Statistical analysis is a contractual requirement of this project and I have a professional statistician lined-up to do this later in the year. However, I wanted to report the interim results to the client but unfortunately the statistician is unavailable at present. He has pointed out to me in the past that Excel (and other software packages) offer statistical tools that in uninformed hands can give rise to nonsensical conclusions. Hence my question. However, I am now 95% confident that I have used the right approach - but will couch the conclusion in careful terms ('weasel words') to my client.

Thank you also for the laptop anecdote - I needed a laugh!

  Forum Editor 21:17 29 Jun 2006

did you know that your affection for Excel puts you in exalted company?

It's Bill Gates' favourite program by a long chalk.

  ade.h 21:19 29 Jun 2006

"I dropped my laptop over the side of a survey vessel into about 40 feet of water."

Well, it's reassuring to know that it can happen to the best of us.

"It's Bill Gates' favourite program by a long chalk."

Now, why doesn't that surprise me?

  wky 21:28 29 Jun 2006

has this anything to do with parametric and nonparametric, because in psychology statistics, a paired t-test would be parametric and there are certain rules for parametric tests.

  VoG II 21:28 29 Jun 2006

I hope that BG doesn't use Excel to deal with his personal finances since the maximum value that Excel can handle is 1.79769313486232E+308 !!

Of the Office applications, Excel has got to be the most versatile although I would not use it to write a letter. With a little ingenuity and a knowlege of VBA it is possible to write Excel applications that do not look like Excel at all. Fascinating for Excel afficianados; boring for the other 99.999% of the population.

  Jackcoms 21:29 29 Jun 2006

"It's Bill Gates' favourite program by a long chalk."

So can we now also expect VoG™ to step down from his working life and spend his time in charitable acts?

This thread is now locked and can not be replied to.

Elsewhere on IDG sites

HTC U12 Plus review: Hands-on

Brilliant (and Weird) World Cup 2018 Art and Design Projects

iPhone 8 vs OnePlus 6

TV & Streaming : comment regarder Roland Garros ?