Experimental design (1)

Just get some review from 1980 to 2018, there are some changes regarding the methods part. For example, univariate to multivariate, self-report to experiment, the single level/single source of data to multi-level/ multiple sources of data, cross-sectional to longitudinal, qualitative/quantitative to mixed methods, single discipline to inter-disciplinary and multi-disciplinary. And this is not to say that we need to master it at university, but to know how to use it, that’s its development. How we move that to a new direction? We still use single level, single source data, we still use self-report, use cross-sectional, but just let you know there’s new thing happening, because it can add more benefits and increase the quality of your research.

I would like to emphasize on the experimental study today, but just the basic part, for other parts, Because each item could be at least one complete course, so I’m not able to do that, just give you a brief introduction, if you are interested, you can get more in detail for different areas. First one, univariate and multivariate, I just come up a list, there are about thirty seven methods, actually they’re all the statistical methods and multivariate methods listed here. I’m not going to go over it, because if you want to go over, it will take long time, but just give you a general idea.

These are the methods, the multivariate methods, probably you can use them for different studies, for different purposes. Second, the self-report and experimental research design, and I want to spend a little more time on this part, but still is only introduction. We’re now going to the detail abut how to do it, how to analyze it, but I just give you brief introduction to see what it is. Quite many academic researchers or some great students all know self-report biases because of self-report, which means you report what you feel. For example, if you do a study, you want to assess tourists’ perception about a certain destination, you can ask how happy you’re? What do you think about service quality? Are you satisfied?

So people just report by themselves. But self-report has some kinds of biases, because their responses could be influenced by all the different situations, so the reported data probably has not true value. Which means they do not report the true data, do not reflect the true perception or the attitude. The bias could be random, could be systematic, could be constant, or could be variable. If that happens, which means the final results you got from your study probably did not reflect the true people’s perception and feeling. Now experiment design is a little bit different. Just like probably you will see the difference by social science and science.

In terms of science, for example, if you study chemistry use of the medical, the scientists will use the mouse to do the lab test, after test is successful and they moved the test to human subject, finally they got the new medicine, new treatment, or new finding in the market. But at hospitality and tourism area, most time, I should say ninety five percent of time, we use self-report, which is a question survey. So some people call us paper pencil research, because we also use the paper and pens. Of course right now, quite many studies are done, surveys done through the internet, but we still use self-report.

The experimental design can get a true variation and a certain condition, the variation is very important. At the very beginning, I’m the same, when I studied research, I did not notice how important variation is, but the variation can give you true picture of your study. In the self-report, if you see the results all towards one side, for example, if you see the one to five Likert scale, the people all towards five, or maybe all towards one, there’s no valuation. So you really need to have a variation under certain condition, and what kind of condition, people how to respond.

The experiment design can provide some kinds of situations that they can manipulate, the situation that people can express their feeling and perception, and their expectation. There are three purposes of experimental design. One is the causation, because most of the time we study is trying to see one variable influences other things, and relationship could be simple relationship or copy an effective relationship. The second one is control, which means that kind of relationship happened under what kind of condition. If a condition changes, people’s perception will probably be different. And I have one example later on to show you how we did control and manipulate condition to try to assess customers’ responses. The last one, variability.

It reduced variability within the treatment conditions, which means under same condition the variability is reduced, but among the between conditions variability actually is high. I would like to give a few examples here. One is that we called video manipulation, second is the virtual reality, number three is neural marketing, number four is eye tracking, the last one is vignette. I give some examples, I’m not going to discuss the detailed technical side, but just give you some cases to tell you how you can use technology. Video, VR, neural, eye tracking are all new technologies, but the vignette is different.