What assumptions about market efficiency are typically adopted in capital markets research? What do we mean by market efficiency?

Evidence shows that share prices might not fully react to financial accounting information immediately and that abnormal returns might persist for a period of time following the release of information (a case of post-announcement drift). Does this indicate that securities markets are not efficient and that assumptions about market efficiency should be rejected?

What, if any, effect would the size of an entity have on the likelihood that the capital market will react to the disclosure of accounting information?