Home > Department of Statistics > Events > abstracts > Significance testing with no alternative hypothesis: a measure of surprise

 

Significance testing with no alternative hypothesis: a measure of surprise

When 2.00pm on Friday 19th January
Where B617, Leverhulme Library, Columbia House
Presentations  
Speaker John Howard
From London School of Economics and Political Science
Abstract A pure significance test would check the agreement of a statistical model with the observed data even when no alternative model was available. I will propose the use of a modified p-value to make such a test. The model will be rejected only if something surprising is observed compared to what else might have been observed. It is shown that the relation between this measure of surprise (the s-value) and the surprise indices of Weaver and Good is similar to the relationship between a p-value, a corresponding odds-ratio, and a logit or log-odds statistic. The s-value is always larger than the corresponding p-value, and is not uniformly distributed.

Difficulties with the whole approach will be discussed.

For further information Thomas Hewlett (Postgraduate Administrator) Ext. 6879
Department of Statistics, Columbia House
Share:Facebook|Twitter|LinkedIn|