Statistics Help... (ANOVA/T-Test)

Page may contain affiliate links. Please see terms for details.

Spinal

MB Enthusiast
Joined
Sep 14, 2004
Messages
4,806
Location
between Uxbridge and the Alps
Car
x254, G350, Duster, S320, Mach1, 900ss and a few more
I need a quick hand...

I have some data, mainly falling into the following categories:
* Operating System (Linux, Windows, Macintosh, Other)
* Answer (Yes, No)

* Other questions (format: selection from a set of string)

I want to do some form of statistical analysis beyond averages... but stats isn't my forte!

I've been told to look at an ANOVA test or a T-Test, and am using SOFA for the analysis... but both the ANOVA and T-test seem to need a numerical column to analyse...

Can anyone give me a pointer? If I replace Yes with '1' and No with '-1' will that give valid results?

M.
 
Bump...

Had a look as SCSS... but I can't figure out how to get it to do a non-numerical type test...

Ideas?
 
I'm not a statistician but.....

Having a quick look at both ANOVA and T-tests, these appear to suit only those sets of results where there is a range of values. The data you're suggesting has only 2 values, yes or no, so I cannot see how these could be used.

What sort of analysis are you trying to make?

By the way, in substituting numbers for bi-value variables I've always found that 1 and 0 are easier to work with.
 
Last edited:
[FONT=helvetica, arial, ms sans serif]Is this of any use?? It does relate to situations which have only 2 possible outcomes and is therefore synonymous with your yes/no results.[/FONT]
[FONT=helvetica, arial, ms sans serif]
[/FONT]
[FONT=helvetica, arial, ms sans serif]Binomial distributions model (some) discrete random variables.[/FONT]
[FONT=helvetica, arial, ms sans serif]Typically, a binomial random variable is the number of successes in a series of trials, for example, the number of 'heads' occurring when a coin is tossed 50 times.[/FONT]
[FONT=helvetica, arial, ms sans serif]A discrete random variable X is said to follow a Binomial distribution with parameters n and p, written X ~ Bi(n,p) or X ~ B(n,p), if it has probability distribution [/FONT][FONT=helvetica, arial, ms sans serif]
bindist.gif
[/FONT][FONT=helvetica, arial, ms sans serif]where [/FONT][FONT=helvetica, arial, ms sans serif]x = 0, 1, 2, ......., n [/FONT][FONT=helvetica, arial, ms sans serif]n = 1, 2, 3, ....... [/FONT][FONT=helvetica, arial, ms sans serif]p = success probability; 0 < p < 1 [/FONT][FONT=helvetica, arial, ms sans serif]
bindist1.gif
[/FONT] [FONT=helvetica, arial, ms sans serif]The trials must meet the following requirements: [/FONT]

  1. [FONT=helvetica, arial, ms sans serif][/FONT][FONT=helvetica, arial, ms sans serif]
  2. the total number of trials is fixed in advance;
  3. there are just two outcomes of each trial; success and failure;
  4. the outcomes of all the trials are statistically independent;
  5. all the trials have the same probability of success.[/FONT]
[FONT=helvetica, arial, ms sans serif]The Binomial distribution has expected value E(X) = np and variance V(X) = np(1-p).[/FONT]
[FONT=helvetica, arial, ms sans serif]Examples

binomial_10_0.5.gif
binomial_10_0.25.gif
[/FONT]
 

Users who are viewing this thread

Back
Top Bottom