Thread: paired samples

In an investigation of the techniques used by two different laboratories for the dating of prehistoric cereal seeds, 10 pairs of seeds were used. Each pair had been found close together at the same site; one seed from each pair was dated by laboratory A and the other by laboratory B. The table shows the ages in years.

Site 1 2 3 4 5 6 7 8 9 10
Lab. A 10124 10672 9491 10848 9277 9217 9294 10584 9983 10004
Lab. B 9902 10680 9474 11044 9315 9344 9696 10571 9712 10009

i) Is there any evidence of a systematic difference in the ages returned by the two laboratories?
ii) Give a 90% confidence interval for the mean difference between the age returned by laboratory A and the corresponding age returned by laboratory B?

are they independent?

It looks like a paired difference experiment where the data is dependent
take the difference of all 10
and do a one sample t test

Each pair had been found close together at the same site..

Does that make them dependent. that is actually what confuses me??

Originally Posted by rubic

Each pair had been found close together at the same site..

Does that make them dependent. that is actually what confuses me??

thats the intention here
It can be anything your instructors want it to be, lol
A much better example of a paired difference experiment is a diet
There you measure n people before and after a period of time
Clearly the two measurements are dependent
BUT the n people are independent
You wouldn't want to weigh 10 people and then 10 different people a month later.