Different input types

[lbenedetto]

Currently, you can only input a % likelihood if hypothesis is true/false.

If someone is using this app for medical diagnosis, it would be more useful to be able to input Sensitivity and Specificity of the test, and then indicate if the test result was positive or negative.

Another useful input method would be to input the Likelihood ratio directly.

It is trivial to convert between these with code, but can be confusing to do in your head, or even on a pocket calculator if you aren't very experienced with this.

Positive Likelihood ratio = Sensitivity% / (100 - Specificity%)
Negative Likelihood ratio = Specificity% / (100 - Sensitivity%)

[jackokaiser]

Gotcha for the medical test use case, I'll also take it in. The likelihood is already the default input type, or do you mean something else?

[lbenedetto]

What I mean by likelihood ratio is a single number that is sometimes provided instead/in addition to sensitivity & specificity.

It's most useful in the odds form of the Bayes equation, because the equation is super simple. Prior odds * likelihood ratio = posterior odds.

So you can input the positive likelihood ratio if the test was positive, or the negative likelihood ratio if it's negative.

For example, this site has the option to display the likelihood ratios instead.
http://getthediagnosis.org/diagnosis/Acute_Sinusitis.htm?mode=display&id=303&sort=0&lr=1