Retrieval

Basic memory processes fundamentally affect the answers survey respondents give to survey questions, and retrieval is one of these memory processes. Retrieval refers to the active recovery of previously encoded information from long-term memory into conscious, working memory. Long-term memory that requires conscious retrieval is divided into (a) memory for facts and (b) memory for events. It is important for questionnaire designers to be aware of the limitations that retrieval processes place on the survey experience for the respondent.

The process of retrieving information and memories for events is closely related to the initial process of encoding, when a memory for the event is first stored by transferring information from active experience into long-term storage. The most difficult retrieval situation is unaided (free) recall, where there are no cues provided that relate to the desired information. However, questionnaires can avoid this situation by including cues in the survey questions that aid in recall. Retrieval is maximized when the cues present during encoding closely match the cues present in the questionnaire. If, for example, memory for a particular event is typically encoded in terms of time (i.e. filing your taxes), then survey questions that cue respondents to think about the event temporally will aid recall. The more cues there are, the more precise the retrieved information will be; for example, memory for a person's name may be aided by cueing specifie events relating to that person, picturing their physical characteristics, thinking about the sound of her or his voice, or thinking about common acquaintances. Although the match between encoding and retrieval cues is important, it is also the case that, all other things being equal, some cues are generally better than others in aiding recall for events. Cues relating to the type of event work best; next best are cues relating to location and people and, finally, cues relating to time.

The depth of initial encoding is also important to later success in retrieving that information from memory; this phenomenon is referred to as the levels-of-processing effect. It is easier to remember information that was initially processed more deeply in comparison to information that was processed superficially. For example, information that is processed both visually (i.e. pictorially) and verbally is easier to remember than information processed in only one fashion.

Retrieval is greatly affected by the type of information that a survey question requires respondents to access. Memory access is first and foremost affected by how long ago the memory was initially encoded. Retrieval for older information and events is more difficult and error-prone, whereas memory for more recent information is easier to access. Certain kinds of information are also easier to retrieve from memory. For example, dates and names are forgotten easily. Information about a unique event is easier to remember, whereas memory for a specifie event that is a repetitive and common experience of the respondent is more difficult to remember and more prone to errors. For example, it is easy to remember the details of meeting the president because this information is distinctive in comparison to other information in memory. In contrast, retrieving specifie information about your last trip to the grocery store is more difficult to remember accurately because there are so many similar events in your past that the details for each event seem to blur into a general memory for “going to the grocery store.” This difficulty leads to errors in identifying the source of a memory; that is, respondents are unable to accurately identify the specific originating event of the retrieved information.

...