Extra Exercise 12
-----------------

PSLS "Statistical significance" applet.

Table of P-values and significance for different observed sample means:

            P-value (significance) for H0:mu=0 &
Xmean       Ha:mu>0        Ha:mu<0       Ha:mu<>0
0.0       0.5000 (no)    0.5000 (no)    1.0000 (no)
0.1       0.3745 (no)    0.6255 (no)    0.7490 (no)
0.2       0.2643 (no)    0.7357 (no)    0.5286 (no)
0.3       0.1711 (no)    0.8289 (no)    0.3422 (no)
0.4       0.1038 (no)    0.8962 (no)    0.2076 (no)
0.5       0.0571 (no)    0.9429 (no)    0.1142 (no)
0.6       0.0287 (yes)   0.9713 (no)    0.0574 (no)
0.7       0.0136 (yes)   0.9864 (no)    0.0272 (yes)
0.8       0.0057 (yes)   0.9943 (no)    0.0114 (yes)    
0.9       0.0022 (yes)   0.9978 (no)    0.0044 (yes)
1.0       0.0008 (yes)   0.9992 (no)    0.0016 (yes)

When Xmean moves further away from 0 (the hypothesized mean value)
IN A DIRECTION COMPATIBLE WITH THE ALTERNATIVE HYPOTHESIS, the P-value
drops down, and the test result shifts from non-significant to
significant at some point. The value required for significance is a 
bit larger for a two-tailed alternative hypothesis than a one-tailed 
alternative (in the right direction), because the two-tailed P-values 
are twice the one-tailed P-values.

For Xmean-values less than zero, the P-values in the two columns for 
one-sided alternative hypothesis switch roles, but the two-sided 
P-values stay the same. (Try it, if you're not sure what is meant!)