discuss issues with p-values

[d-morrison]

• small number = high evidence for H_A
• doesn't distinguish absence of evidence and evidence of absence