A planned error-pattern table shows blank-prompt tests producing many omissions, while multiple-choice tests produce fewer omissions but some selections of familiar distractors. Which interpretation best fits retrieval concepts?