Microbenchmark for different collection types in Python. What do you expect to happen?

demo.py

def set_up_test(count):
  l = []
  d = {}
  s = set()
  for i in range(count):
    l.append(i)
    d[i] = True
    s.add(i)
  return l, d, s

def stopwatch(container, count):
  import time
  start = time.time()
  n = 0
  for i in range(count):
    if i in container:
      n += 1
  return time.time() - start, n

import sys
count = 10**5
if len(sys.argv) > 1:
  count = int(sys.argv[1])

l, d, s = set_up_test(count)
print "dictionary", stopwatch(d, count)
print "set", stopwatch(s, count)
print "list", stopwatch(l, count)