Exercises

The default values of the parameters are:
Probability of success in Condition 1 = 0.6
Probability of success in Condition 2 = 0.6
N = 10 for both conditions.


  1. Using the default values, simulate one experiment. Calculate Chi Square by hand and compare your results to those shown in the simulation. Click here for the computational method.
  2. Again using these default values, what proportion of the time should the Chi Square test be significant if the test were exact? Determine the proportion of significant outcomes by simulation. Does the test appear to be exact or approximate? Does Yate's correction for continuity improve the test?
  3. Find a set of parameters for which the Type I error rate is too high for the uncorrected test.
  4. Try out a variety of parameter values and see if you think the correction for continuity should be used as a general rule.
  5. A researcher is planning an experiment in which she believes that the population proportion of successes in Condition 1 is 0.6 and the population proportion in Condition 2 is 0.3. If the researcher wishes to have about an 80% chance of finding a significant difference (assuming her beliefs about the population are correct), what sample size should she use. Determine this for both the 0.05 and 0.01 significance levels.