What Price Replication?

The Story

In an ideal world, we would have the time, money, and wide-eyed graduate students to run all the studies we desired. However, in the real world we don't have infinite time, money, or labor to conduct enough experiments to be completely sure of our claims.

We'll play the same game as last time, in which replication can help us to zero in on the truth of a hypothesis. This time, however, science is costly, so the number of tests you can conduct is limited. Choose wisely.

Get to Work

Your grant allocates $100,000 for research expenses. You can conduct experiments until the money runs out. Each experimental test of a hyothesis costs $1,000.

To replicate an experimental test, mouse over the hypothesis so that it is highlighted, and click the mouse cursor. You can test each hypothesis a maximum of 12 times. Each test has the same power and false positive rate.

When you're ready, we will once again let you know which hypotheses are actually true.

Up Next...

We'll look at some other factors in how we assess the truth of published research, including accounting for bias in the submission and publication of negative findings.


Test new hypotheses and replicate them

Base rate: Power: False positive rate:
Novel hypotheses: Expected to be TRUE:

Cash remaining: $

Hyp#   Research Findings What do you think?