In this study two small (25) groups of (mostly female) students were ask to keep their hands in ice water.
Their averages were 40.12 and 42.03 seconds. Which gets to show the range of variance with a small sample of 25.
They were told that they would do a different experiment after that: evaluating product design. One group handled Noodles, the other a bottle of Ibuprofen for 2 minutes.
After that they did another ice water test and now the results were:
45.28 seconds (instead of 40.12) for the Ibuprofen group, while the Noodles people lasted for 41.83 second instead of 42.03.
I wonder what would happen if one would repeat this experiment. As much one would want to have that handling pain killers will reduce pain the actual difference is so small that it could easily be part of noise to be expected in a sampling size of 25.