Paired-sample or Independent T-test?

#1
Hi, I have read all of the examples available but I can't tell which test to use because none of the examples use the kind data that I do! Please help, this is for my MA in Archaeology.

Basically, I want to figure out whether buildings are generally closer to shrines than they are to other buildings. I have 24 buildings, listed by an ID# in SPSS and two measurements for each. Distance to closest shrine and distance to closest building in metres.

I initially thought paired because each building has two measured variables. But the variables are independent of each other, i.e. how close a building is to a shrine is not likely to influence how close it is to another building.

I hope that makes sense and that someone out there can help! Thanks in advance
 

Karabiner

TS Contributor
#2
I have 24 buildings
Are these buildings completely independent from each other
(e.g. each building is in a separate city, so you did not measure
the distances from each of the buildings to the other 23)? If yes,
then you can simply perform a dependent samples t-test
(measurements are paired, for each building-shrine distance
there's exactely 1 paired building-nearest building distance).
Or perhaps a Wilcoxon signed rank test, if assumptions of the
t-test are not fulfilled.

With kind regards

K.