Urgent statistics help needed for a confused lady

ConfusedLady

Hi there

This forum was recommended to me by a friend for the kind help she received. I'm an undergraduate studying radiography and currently doing my research dissertation. My study the effect of increasing FFD (focus to film distance...which means..distance from the xray tube to the film) on dose to the patient. I tested varying heights from 100cm to 140cm in 10cm increments. I took readings of the entrance dose, exit dose, dap and optical density. I want to compare the results I have for each height with each other.

I have done the basics...mean, standard deviation, variance, correlation coefficients, correlation of alienation, z score but I am having a nightmare when it comes to which statistical test I should use!! Its driving me a bit insane :mad: and I only have a couple of weeks until its due! ahh! :eek:

So any help would be greatly received

Many Thanks

Lynsey

2Scoops

What are you trying to find? What would be the ideal result from this investigation - is one of these FFDs going to 'better' than the others and, if so, how? The statistics you use are wholly dependent on what question you are asking.

Also, please briefly explain the following and state what units they are measured in: entrance dose, exit dose, dap and optical density. This will indicate what kind of statistical test you should use.

ConfusedLady

Thank you for getting back to me so quickly.

Yes its too see which distance gives the lowest dose to the patient whilst maintaining image quality.

Previous studies have shown a distance of 130cm is best.
My results agree with this statement as the values are at their lowest and it has maintained optical density.

The Entrance dose units is mGy.
The Exit dose units us mSv.
The DAP units is Gy cm².

The optical denisty is being assessed by the AGFA LGM reference value on the image. This is a value given regarding to the densities on the image. I'm not to worried about using a test on this data set as all the results fall into the range where they should be. Which can just be stated.

I just want to compare the above dose readings for each Distance. 100,110,120,130 and 140cm.

I hope that makes some sense

Thank you again

Lynsey

2Scoops

These are all scale variables and can be analyzed with a simple 1 way repeated measures ANOVA. It might be worthwhile seeing if the data are normally distributed before applying this test. You will also have to see if sphericity is a problem (this test is usually done with the RM ANOVA itself).

If you would like help doing this analyze or interpreting the output let me know.

Basically, assuming all goes well, the ANOVA will let you if one or more of the 4 distances is different from another. If it is not significant, then job done - they are all the same (assuming your study is adequately powered). If it is significant, then you jump in and compare each distance to the rest to find out where the differences lie (and apply some correction for multiple comparisons to keep type 1 error in check).

If there are problems with normality of sphericity, or if you want to analyze the densiometry variable, you can use the non-parametric Friedman's test, which does not rely on those assumptions, but is less well able to detect significant differences.