conlangs1-4  Exercise 3: Testing Linguistic Relativity
======================================================

One possible design.

  Measure: give two groups the same set of short logic puzzles and compare the
  proportion solved correctly and the time taken. Group A: people who have
  learned Lojban to a good level. Group B: matched English speakers who have
  not.

  A problem: self-selection. People who choose to learn a logical language
  may already be interested in logic and be better at puzzles. Any difference
  could come from who they are and not from the language. A better design
  would test the same people before and after learning, with a control group
  learning another language for the same time.

  What would support the weak form of the hypothesis: a small, consistent
  advantage on tasks that involve the structure Lojban makes explicit (for
  example, spotting how many things a relation involves), but no general
  advantage on all reasoning. A large across-the-board effect would be a
  surprise, and would still need an explanation other than the language.

Your design will differ. Check that you name a measurement, a confound and a
result that counts as support for the weak form specifically.