In the previous segment, we alluded to the problem of comprehension in the early examples, and that is the, last piece here that we want to feed to you as one that is related to recall error. It is important to know and think about, how is an object defined in the respondent hand. If you have to retrieve from memory, it is dependent on what the respondent understands what is searched, you know? So what memory traces are in the respondent hand? Does the object line up with any of these, and do they feel, that they know the answer to this? You know, it can help to think about what are, relevant memory clusters. So if you ask about a specific thing and you can ask about a set of things that's sort of similar to that specific thing, then it might be easier to recall these things. And that, you know, we summarized in terms of, you know, setting a context for retrieval. You should think about where in the questionnaire, you have the possibility to set such contacts. So this would be the introduction to the item. The surrounding items themselves. The wording of the question, and often also the answer categories, if it is a closed ended question. You can help with the introductory statement for example, to facilitate the accuracy of recall questions. So, in the context of victimization surveys, you could do this similar to the introduction of crime and victimization survey, that says, please tell me about any crime that may have occurred int he last six months. That's a pretty short one. Not too much in the introduction yet and we'll show you, a variation on that similar segment. The introduction to a, the national violence against women survey, on the other hand, has a introduction here that says, we are particularly interested in learning about violence women experience, either by strangers, friends, relatives or even husbands and partners. And so here, you set a context. You in the introduction reminding them that, you can be victim of crimes that come from a lot of other people not just a patic, particular stranger which might be the first thing that people think about. So you change the focus and the attention to what could be in memory. This leads to, what we pointed out earlier as a detailed cueing. It can help define object. You know, you, when you decide your questionnaire, you want to use cues, to focus retrieval, on those objects. But when you do that, be careful and mindful of two things. The cues might interfere with retrieving other information. So you can, you know, focus too much on the cues and block everything else. And, the cues also implicitly might exclude objects that are not mentioned. So if respondents think that this is a finite set, and not just some memory help, then you can be in trouble as well. so, you know, this is a suggestive improvement for your questionnaire, but how it works in practice has again be need to be tested in before fielding the questionnaire. That's why we have a whole unit on question testing in this set. So, let me show you an example here for the Summary vs detailed Cueing approach, for the Crime and Victimization survey. So here's a whole set of questions that you could have, where the questions themselves provide context and cues so, did you have your pocket picked? First question. Did anyone steal things that belonged to you, blah, blah, blah. Was anything stolen from you while you were away from your home, and so on? And then, finally, was anything at all stolen from you during the last six months? You know would be a summary question of those individual ones. The detailed cueing on the other hand, what you see on the right is a long question, was something belonging to you stolen such as and then all the kinds of things are listed. In fact so many that I couldn't put them on this slide here. David Kent as I said, did a lot of studies on The National Crime and Victimization Survey and here again, we have rates of victimization per 1000 people and you know, asked with the two different approaches. The summary questions that we had up here, was anything at all stolen from you and then, this detailed Cueing question. And you do see that we have, a higher rate of reporting on these detailed Cueing questions. Both are violence, poverty and household crimes. Now let me summarize, the methods for improving answers to factual questions. First of all, we learned short recall periods are good. Rate of forgetting can be overcome. And shorter lengths of the survey is always good. You overcome fatigue on the respondent side and help with the memory. An appropriate structure of the questionnaire, helps setting Q in context. And helps the respondent remember. The user records, in addition, can help. There, can be, an, attempt to increase salience and motivation of particular items. And the motivation to, even, do, that task of remembering. You can, use lists to prompt the respondent or diaries. You know, TV watching for example will often be asked with the help of a diary. Because otherwise it gets, too much to remember everything that happens even in two weeks. You can have trigger question or decompose a question into various subparts so that you can, instead of having cues as we saw with the, the crime and victimization survey. Just help the respondent by asking about all the different elements that belong to that particular question. So in the next segment, we will be moving on, to asking sensitive questions. And looking forward to seeing you there. [BLANK_AUDIO]
