topic Re: How to choose "Level Tested" in Binomial Probability test in Discussions

How to choose "Level Tested" in Binomial Probability test

ms_Ni — Sat, 10 Jun 2023 20:43:29 GMT

Hello,

I am a new user of Jmp. I currently using it to run binomial hypothesis testing, and have encountered what I believe is a "settings issue" that I cannot seem to solve.

I am testing binomial hypothesis for data that has either "satisfied" or "dissatisfied"; when I run the "test probabilities" and enter the given threshold (say, 90%) for the "satisfied" value and 0.1 for "dissatisfied", respectively, JMP returns the probabilities for "dissatisfied".

So, is there a way to choose which "level" is tested? I understand that the values give me the result I am looking for, but I am looking for the data table to show said value.

Any help is much appreciated - a screenshot is uploaded to show which figure I am talking about.

Re: How to choose "Level Tested" in Binomial Probability test

ih — Tue, 23 Feb 2021 14:07:04 GMT

It doesn't look like your screenshot made it onto your post and that might help us understand your goal. Could you try posting it again? Also, it would be helpful to have some example data to work with. Could you recreate your chart with one of the example data sets, or using the table below?

Names Default to here(1);

New Table( "Example Data",
	Add Rows( 50 ),
	New Column( "Responses", Numeric, "Continuous", Format( "Best", 12 ),
		Formula( Floor( Random Uniform( 1, 9.99 ) ) )
	),
	New Column( "Satisfied", Numeric, "Ordinal", Format( "Best", 12 ),
		Formula( If( :Responses < 3, 0, :Responses > 7, 2, 1 ) ),
		Value Labels( {0 = "Dissatisfied", 1 = "Neither", 2 = "Satisfied"} ),
		Use Value Labels( 1 )
	)
);

Copy this into a new script window and press run to create the table.

Re: How to choose "Level Tested" in Binomial Probability test

Mark_Bailey — Tue, 23 Feb 2021 15:56:44 GMT

You enter the hypothesized probabilities for the levels of interest. The null hypothesis, that is. So does 0.9 represent the assumed probability of satisfied responses for which the alternative hypothesis is another probability? These are tests for a difference (i.e., decide that probability is not the assumed value).

See this example: Test Probabilities