stats help PLEASE

sarahbarron

hi my name is sarah and im in urgent need of some help. im doing an undergraduate course in radiography but for some crazy reason i have to do statistical analysis as part of my dissertation. i did an experiment where i took a number of x-rays on a phantom in two different postions.

my dissertation supervisor told me that I have to do the anova test and the wilcoxen test. the poor man spent over two hours explaing how to do it but i came out and still hadnt a clue. my dissertation is due in a couple of weeks and i cant finish it unless i get some help.

the data i have is the dose recieved by four sensitive organs, the testes, the ovaries the eyes and the thyroid gland. i have data for both positons.

if anyone can help me even by telling me which data i use for which test id really appreicate it. i dont even know what probability is

sarah

Thomas_S_Hunterson

Ok what software have you got?

The data you use depends on what you are trying to test. I'm guessing you should be trying to compare the two treatments (positions) with a paired (by organ) test?

Any statistical software package will do the tests with no hassle. minitab is probably easiest to use for a beginner.

sarahbarron

i have excel does that help. seriously i just dont have a clue. im definately trying to compare both positions to see which one gives the least amount of dose to the patient. but then i also think im trying to compare the dose to the individual organs i.e the eyes in one position and the eyes in the other.

so i think i have to use the anova one for the comparing of both positons and then the wilcoxen the results of each positon. but then again i could be completely wrong.

ive also just discovered that there is three types of anova test?????? and i dont know which one i would use AAAAAARRRRRRRRRRRHHHHHH

Thomas_S_Hunterson

Seriously, get a book on this stuff. It will stand to you. Statistics will be very important to you later on. I'm surprised you haven't been taught.

What you're trying to do takes a couple of seconds in minitab but it's important that you actually understand what you're doing. I don't think excel is really a serious option. Your college will more than likely have the necessary software on its computers.

Thomas_S_Hunterson

Sean_K wrote: »

Ok what software have you got?

The data you use depends on what you are trying to test. I'm guessing you should be trying to compare the two treatments (positions) with a paired (by organ) test?

Any statistical software package will do the tests with no hassle. minitab is probably easiest to use for a beginner.

just ammended this, meant minitab, but somehow typed matlab. Hope that didn't confuse you:o

2Scoops

sarahbarron wrote: »

i have excel does that help. seriously i just dont have a clue. im definately trying to compare both positions to see which one gives the least amount of dose to the patient. but then i also think im trying to compare the dose to the individual organs i.e the eyes in one position and the eyes in the other.

so i think i have to use the anova one for the comparing of both positons and then the wilcoxen the results of each positon. but then again i could be completely wrong.

ive also just discovered that there is three types of anova test?????? and i dont know which one i would use AAAAAARRRRRRRRRRRHHHHHH

Are the data scale or ordinal - i.e. if one value is 1.5 and another 3.0, does that mean the second value is exactly twice as big as, or just 'bigger than,' 1.5? If it's not a scale meaure, this will indicate that you need to use non-parametric statistics like Wilcoxon straight off the bat. Otherwise, ANOVA is safe enough (though it relies on a number of assumptions that are commonly ignored by medical researchers :eek:).

A straight forward repeated measures ANOVA could be used with 'position' as the independent variable and then run it 4 times - one for each organ. This is essentially a paired t-test which MS Excel could handle, although it tends to be off sometimes.

If you need to compare organs to each other as well, e.g. eyes vs thyroid, a 2-way ANOVA with repeated measures on 'position' but not 'organ type' would do the job. Excel won't be able to do this, but it may be more analysis than you need to answer your research question anyway.

The researcher forum can help you run through your statistics as a case study thread, if you like. :pac:

sarahbarron

Sean_K wrote: »

What you're trying to do takes a couple of seconds in minitab but it's important that you actually understand what you're doing. I don't think excel is really a serious option. Your college will more than likely have the necessary software on its computers.

I got minitab 15, if that's any help.

2Scoops wrote: »

Are the data scale or ordinal - i.e. if one value is 1.5 and another 3.0, does that mean the second value is exactly twice as big as, or just 'bigger than,' 1.5? If it's not a scale meaure, this will indicate that you need to use non-parametric statistics like Wilcoxon straight off the bat. Otherwise, ANOVA is safe enough (though it relies on a number of assumptions that are commonly ignored by medical researchers :eek:).

A straight forward repeated measures ANOVA could be used with 'position' as the independent variable and then run it 4 times - one for each organ. This is essentially a paired t-test which MS Excel could handle, although it tends to be off sometimes.

If you need to compare organs to each other as well, e.g. eyes vs thyroid, a 2-way ANOVA with repeated measures on 'position' but not 'organ type' would do the job. Excel won't be able to do this, but it may be more analysis than you need to answer your research question anyway.

The researcher forum can help you run through your statistics as a case study thread, if you like. :pac:

I'm still not exactly sure what you're talking about. My lecturer told me to use both the ANOVA and Wilcoxen tests. I've followed your advice and created a thread in the researcher forum, so hopefully everyone can help me step by step in there!