Not sure which test to use...


New Member
Hi everyone, thanks for reading. I'll try and make this quick, so that hopefully once I know which test I should use I can go and have a go rather than keep bugging you! :eek: :eek:

I am doing a research project investigating the specificity and sensitivity of urine dipsticks to measure ketone levels in cow urine.

I took a single urine sample from each cow and read the dipstick ketone result at 5, 15, 30, 45 and 60 seconds, recording it as negative, 1+, 2+ or 3+ according to the reference chart supplied with the dipsticks. The limitation of this simple test is that I although I can convert these to quantitative results as shown on the reference chart (0, 1, 5 and 15 mmol/L respectively), I cannot accurately convert them to a range of values.

I also have a blood ketone level from each cow to compare the urine results with, in order to find the specificity and sensitivity of the dipstick test.

Now, I can work out the specificity and sensitivity of the test when read at 5, 15, 30, 45 and 60 seconds easily enough. But I'm a bit stuck as to what statistical test(s) to use. I was looking at a Repeated Measures Anova (seeing as I have been obtaining repeated measurements of the same parameter) but the change in values is caused by nothing more than time; ie there are no other factors determining the result than 'number of seconds the dipstick is read after' so I don't think that test would work. :confused: Should I simply compare the specificities and sensitivities to aid my conclusions?

Sorry to bother youse with this. I have been trying to read up on it but just feel more confused than when I started. :(