In elicitation, a social engineer will not simply walk up to a target and request him or her to give out sensitive credentials. While this information is the main goal, it must be sought in a careful manner to avoid raising suspicion. Therefore, social engineers must learn how to ask questions in the right way. Mostly, social engineers will use questions that will appear as useless from the target's perspective and, therefore, they will not be reluctant to answer them. It is only when these tiny bits of information are combined that the big picture is visible.

Let us take a close look at the types of questions that social engineers use in order to gather information from their target:

• Open-ended questions: These are questions that require a little bit more than a yes or no response. Therefore, if posed to a target, more information can be gathered. Open-ended questions are commonly used in interviews so as to elicit respondents to keep giving answers. Each question is followed by a follow-up question inquiring about the why and how of the response given. Therefore, a target will keep adding more details to the responses given earlier. Since some targets may show some resistance when answering questions, it is best to use the pyramid approach, whereby the social engineer will start with narrow questions and save the broader questions for the final parts of the interview.
• Close-ended questions: These types of questions are used to lead respondents down a certain path. They usually can only be answered using a few options. The goal is not to gather information but rather to lead the respondent to a certain corner where other questions can be used to gather the information needed. Attorneys and police officers commonly use these questions when they are trying to pin them to a particular path. The questions might appear short and segmented but they slowly restrain a respondent from diverting to another version of the story down the line.
• Leading questions: These usually borrow the best of both open and close-ended questions. They are open-ended but give the respondent a hint at the expected answer. For example, I saw the printer company send in a few technicians the other day; your company must be facing some technical issues with the printers, isn't it?. This question invites a yes or no answer followed by a narrow explanation. It also suggests to the respondent the idea that you are knowledgeable about what you are saying. Even when a respondent only gives a yes or no answer, the leading question will have some other impacts. A lot of information gets planted into the respondent's brain and more facts are suggested for him or her to agree or disagree with. The respondent is elicited to talk more in the end. Leading questions are also particularly powerful tools in that they can distort memories. They can be used to manipulate a target's memory. A study done in 1932 by a psychologist called Bartlett came to the conclusion that people can hardly recall an event accurately and entirely. This is because people make memories in different ways based on what is regarded as important by their brains. When the same memories are referred to by the respondent, some will have been lost and the ones regarded as important will be recalled. To cover up the deficit, some generalizations are done to fill up the gaps.

Therefore, if a respondent is preloaded with some information before the question is asked, the preloaded information will have a distorting effect on the memory. Leading questions are, therefore, very important tools for seeking information forcefully but unconsciously driving a respondent down a particular path.