Can you explain please …
A goal of an innocent until proven guilty justice system is to place a higher priority on:
A) avoiding type II errors. B) the null hypothesis. C) avoiding type I errors.
I would say C, avoiding Type I errors. If they are assuming innocence and trying to prove guilt, then H0: individual innocent; Ha: individual guilty.
If they don’t want to wrongly accuse someone, they would prefer to minimize the probability of calling someone guilty when, in truth, he or she is innocent [rejecting H0 when H0 is true (Type I error)].
I wish that these question writers would learn their subject matter: it’s _ presumed innocent until proven guilty_. (Whether they’re actually innocent doesn’t change merely based on what can or cannot be proven.)
Which error type you avoid depends not on whether they’re presumed innocent until proven guilty or presumed guilty until they’re proven innocent; it depends on the threshold of proof (another thing this author bollixed). A high standard of proof (high confidence level, low level of significance) avoids a Type I error, while a low standard of proof (low confidence level, high level of significance) avoids a Type II error.
Anyway, the answer they want you to reach is C), but it’s an incredibly poor question. I’d be ashamed for having written a question such as this one.
But that depends on the threshold of proof, not on the null hypothesis.
(Oh, and I believe that what they don’t want to do wrongly is _ convict _ someone, not merely accuse someone.)
Sorry, I was a little distracted, but good catch. Also, I know it would depend on the threshold of proof (I agree a pretty crummy question), but I thought of what the author was (probably) trying to test.
As did I. And we’re probably correct, but it’s an awful question.
People who read that question and its explanation (that would provide a good giggle), will mistakenly think that how you choose your null hypothesis has some bearing on what type of error you’re trying to avoid, which is absurd. So they’ll have a poorer understanding of how hypothesis testing really works than they did before the question.
Grrrrr.
I agree, and unfortunately, there is a lot of incorrect or misleading “authoritative” information published for people to “learn” from. Another issue is that too many people miss their mark when they try to write tricky questions (professors, textbook writers, prep providers). I just think content providers have to produce thousands of questions and so, a few bad questions like this pop up.
thank you very much all. all these questions are really making me confused & also i am becoming uncomfortable with hypothesis.thanks
Which prep provider did that question come from? In addition, I would recommend looking at Khan Academy (free videos) to see if they are helpful for you to understand hypothesis testing better; S2000 has a website as well. Lastly, I would recommend using a legitimate statistics textbook if you want further clarification/practice.
Here is a link to Khan (in case you want to check it out)… https://www.khanacademy.org/math/probability/statistics-inferential
You’re quite welcome.
I’ve been working on a series of articles on hypothesis testing, but it’s not quite complete yet. Ask me at the end of January.
thank you very much…
sure & thanks
Glad to help!
that question was from kaplan schweser…thanks