Stats: Determining significance level in problem

Join the discussion
Ask a follow-up here, or get your own question answered by working scientists, mathematicians and engineers — people, not an autocomplete.
Real named experts · corrections over time · the nuance an AI answer skips
Sign up Log in
5 replies · 2K views
physicslady123
Messages
10
Reaction score
0
TL;DR
The answer is 50% but why isn't it 5%?
Question:
In finance, the strong form of the efficient market hypothesis states that studying financial information about stocks is a waste of time since all public and private information that might affect the stock price is already reflected in the price of the stock. However, a study of 450 stocks found that only about 8% had price movements that could be accounted for in this way. At what significance level could you accept the strong form of the efficient-market hypothesis?

I know that significance levels are usually 5%, 1%, or 10% (as confidence levels are 95%, 99%, and 90%). So, why is the answer 50%? Isn't this too large of a significance level?
 
Physics news on Phys.org
I don't think the question has a meaningful answer.
Giving up and saying "what if it's even?" might be an answer but I don't think it is a good one.
 
physicslady123 said:
Summary:: The answer is 50% but why isn't it 5%?

However, a study of 450 stocks found that only about 8% had price movements that could be accounted for in this way.
Hint: what is the p-value of this data set?
 
I'm confused by the question too. If the strong market hypothesis is that 100% of stocks should "have price movements that could be accounted for this way", and only 8% of them do, isn't it just wrong? Where does a significance level even come into play here?
 
You have a more fundamental issue - the Joint Hypothesis Problem. Who is to say that your model is correct? To test the EMH you need a model of what ‘efficient’ values should be. However if you find violations of this model, to whatever significance level you choose, you don't know whether the EMH or your model is incorrect.

however, nobody believes the strong form of the EMH, which states that stock prices reflect even non-public information. If that was true, no one could make an economic profit trading on insider information
 
I think there is some confusion that this is a question about the Efficient Market Hypothesis; it isn't, it is a (rather poorly constructed) question on significance testing and it belongs in the homework section.
 
  • Like
Likes   Reactions: Dale