So continuing on, another type of question that we sometimes use in surveys, whether they be paper or electronic would be a visual analogue scale. We discuss these briefly in our project looking, looking at measuring pain in Morphine Marinol studies earlier. But, again the the idea here being that, you can ask an individual a question, you have a known scale. If it's on paper or maybe that scale is ten centimeters in length, you have one side of that scale with one label, another label on the other sort, sort of the banding the responses. If it's on paper you ask the individual to rate something by making a mark on the paper. And then you can go, go forward then and measure where that mark is. And, and so you're turning something into a quantitative measurement, a continuous variable. In that case, between zero and ten centimeters. If you're doing it using a website you can have a slider type bar and have them, do essentially the same sort of thing. One nice thing about visual analogue scale questions is that, you can enable less respondent perception and memory bias in, in the answers. And what I mean there is, you know, well I remember when I, when they asked me about this question earlier this, this week. Or, or yesterday or maybe even earlier in this survey around this concept. I said seven out of ten, so I'd better make this an eight. so, so, this way you get a little bit more objective response from the from the participant. You can consider Likert squa-, scales, we use those a lot in surveys. A Likert scale would be a question, followed by a series of answers. like how often do you experience pain on a daily basis. And then the individual would be, would choose one of the following options, never, very rarely, rarely et cetera, on up to very frequently. [COUGH]. When you're using Likert questions and again they're very common. When you're using them make sure that you're coding the responses for later analysis, so that zero, whenever they choose never, that you're coding that in the database that you'll analyze later as a zero. very rarely could be one, et cetera. The respondent doesn't have to see and probably shouldn't see those codes as we mentioned earlier. But from an analysis standpoint, it's a good idea to to do that coding. typically you see anywhere from five to nine responses in a Likert scale type response. Some expert likes, experts like even number of responses, some like odd number of responses. and, and really you can ask very strong experts in this field and they will have different opinions on it. as I mentioned earlier, I'm not an expert in this field. But my opinion is, I always like to use an even number, because that way you can force people off of the neutral. Rather than you know, people have a tendency to want to sort of migrate towards the, the middle or the mean. And so you'll, you'll find that a lot of times you'll, you'll have a few on the edges categories. They'll bunch up in the middle and if you if, if you have an even number of responses at least you can, you can get them off of that dead center. scale should be balanced and gaps between the answer choi-, choices should be equidistant. So in other words, you shouldn't say how much do you like ice cream? Hate it, moderately love it, and love it. You know, there's, there's, there's too much room between hate it and moderately love it there. So make sure that things are equidistant re-, regardless of how many question answers that you give, there are choice, choices that they have. consider branching logic to increase data validity and decrease survey fatigue. So, I think we covered this in an earlier module but, but briefly again, because we use it a lot in surveys. Or we should be using a lot in surveys, where, where the individual is spending a lot of time giving us the data that we're asking for. if they answer that first question, what is your gender, male or female. you know, if, if, if we're using branching logic, we can may have either the computer or maybe sometimes on paper surveys. You'll see, you know, wording of if you answer female, answer this next question. With a computer, you can actually make the web, websites, et cetera. It's pretty common to just make the the next question disappear and only have it appear if they in fact chose female for question one. So again, these are important in surveys sometimes you can, you can completely hide a whole section of questions use, using branching logic. And so, so you'll greatly reduce survey fatigue and and, and increase your participant satisfaction by, by leveraging branching logic. So, so on to our road map here, our, our guiding map and navigating the survey process, next topic would be assembling the data collection instrument. Again, going back to that principle again that that depending on how you deliver these things, whether it be by mail, or an email, or handing them off in person. Or maybe they're just sitting on a, clinic clinic coffee table, some, something of that nature. assembling those and, and making sure that those are done, it's done in a professional way. Doing it in a way that will, actually make, make people want to take that survey is, is important. So first thing I always think about is making sure that we're introducing the survey with sufficient information for respondents to understand and the intent and also the expectations of the survey. So, so, I like any time I'm using surveys in research, I like to make sure that then in that introductory mess, message, whether that be in the email or at the top of the survey itself or, or, you know, maybe both. I like to explain, well what is, what is the purpose of this study? What are, what are you contributing to by, by donating your time? here or, or participating using your time in in, in, in taking this survey? I like to tell them how much time is required to complete the survey. And this is really important. We'll talk about testing in, in, in just a few slides. But it's really important that, that you actually have an honest assessment of how long it's going to take to complete the survey. Nothing is more irritating than to get a message saying that, hey, we'd like you to fill, fill out this particular survey and it only takes a couple minutes of your time. Nothing is more frustrating than getting 20 minutes into one of those and not seeing the end in sight. So be, be realistic, be honest with the participants. telling them how the data will be used. whether the responses will be traceable back to them or not. there are ways that you can do identified surveys. There are ways you can do de-identified surveys. you can even do de-identified surveys where you're collecting information on who actually took the surveys. So, and to storing it different and indexed in a different way. So that you can't trace the responses back to the to the participants. But, but it's important that the, the volunteer or the participant in this case, knows what to expect, or what's being done with my data? Can it be traced back to me? it's important to, to let them know upfront, you know, what are the incentive opportunities for participating if I donate my five minutes of time for this? any, any additional instructions needed before starting. One, one of the I, I was taking a survey just recently for a work related survey, not a research survey. Well, one of the things that would of have been really nice taking this particular survey that, that I got annoyed with. Was knowing, you know, kind of the domains of information that I needed to collect before I started taking the survey. This really was one that I got five or six minutes into and realized, hey, I actually need to go pull a couple of other folks from my team to finish this out. Because they've got the details here, here that I would need to finish it out. So make sure that when you're, when you're assembling that data collection instrument, that when you're publishing information to the participant, saying here's what it is. Here's what it does, here long, here's how long it's going to take. Make sure you also give than any additional instructions needed before they start the process. when, when they finish the survey, you know, just as, as you might do in person. Make sure that you, you know, you're setting things up so that it's, it's a respectful and a courteous conclusion. So again, I'm sort of thinking about it from the, from the software standpoint and a web based application. Make sure that the system thanks the participants. Make sure that you know, they you let them know that what, what's going to happen with the study results, and how that might be disseminated. optional you know, again if you are doing some sort of an incentive. Let them know again, you know, that here's how that's going to happen. Here, here's about when it's going to happen and, and, and you'll be notified this way if, if you are If you are you know, selected or, or, or you know, if it's not a drawing after, after you participate. And, and you're automatically going to get some sort of an incentive type type, type reward. Make sure that you're letting them know right then, you know, how, how to redeem that or, or when it's coming. consider breaking up a very long survey into several clearly identified sections. And I've seen this work well, whether, where a very long survey could, could, could reasonably go on to one web page. if, if you were collecting data electronically, as long as you're breaking it up into sections and so there's clearly demarcated areas. it cou-, might also be that you want to sort of have one section per per, per web page as, as your doing the submission. So you can, you can play around with it anyway that you want. But as you assembling that instrument, make sure that you're, you know, you're really always looking at it with, with an eye of being courteous. And re-, respectful of what your end user participant would, would want to be seeing because it, it'll greatly enhance the participation. Ca-, consider formatting. again, you know, the same types of rules apply on paper as they do on on web based survey instruments you know, make sure that it's easy to read. You know I, I, I always in, in, insist on using sort of large font size and high contrast. because you don't want to antagonize your participant, by giving them something that's, that's that's hard to read or that's, that's hard to distinguish. Leave lots of white space align the answers in consistent ways. So if you're always right justifying in, in, you know, always right justify, don't, don't go back and forth. Because it's just sort of annoying and, and, and, just a little bit of cognitive dissonance on, on your participants side. It kind of builds up and they'll get tired of taking your survey or, or, or think that it's not very professional and maybe not do their best job. And lea-, leave plenty of space for answers in open-styled questions. So, one of the things I've seen here, particularly in companies asking for feedback. might, might be to have a survey on a website ask you to please describe you, you know, your negative experience with our company or, or maybe even a positive experience with our company. Where you know, they give you this much space to, to, to fill in the details. You see it sometimes even like a, like I've shown there in that bad example, you know, where you just have a very small line to fill out. And so from those, you know, you can typically infer that, that the people making the survey. You know, they had the right idea, they knew they needed to ask that question. That they knew maybe even what they were going to do with it, but they didn't do the testing. They didn't go through and sort of pretend they were a participant, see how this really looks and feels as, as you're going through and providing answers. And so that sort of thing will absolutely kill you when, when you get the answers back. And you realize that, hey we just, we just popped that survey out to 400 people. And the answers are all coming back very, very short. At this point, there's nothing you can do about it. And so again, going back to that, get it right before you start. this is one of those areas where it's really important. So that's it for this section, we'll pick it up in the next.
