Social experiments are powerful sources of information about the effectiveness of interventions. In practice, initial randomization plans are almost always compromised. Multiple hypotheses are frequently tested. "Significant" effects are often reported with p-values that do not account for...

"When studying income differences and income distribution, the self-employed are often excluded from the population studied. There are several good reasons for this, for example that incomes from self-employment are not reported to the same extent as incomes from being an employee. On the other...