WEBVTT Kind: captions; Language: en 1 00:00:00.460 --> 00:00:03.360 Welcome to the Advanced Statistical Research Methods course 2 00:00:03.360 --> 00:00:06.770 or the advanced course like I like to call it. 3 00:00:06.770 --> 00:00:11.430 This course is continuation from my basic course or statistical research 4 00:00:11.430 --> 00:00:14.650 methods, which many of you have already taken on this video. 5 00:00:14.650 --> 00:00:18.080 I will briefly go through what this course is about, what kind 6 00:00:18.080 --> 00:00:21.360 of topics we address on this course and so on. 7 00:00:21.360 --> 00:00:26.240 Other videos in the introductory set will explain the practicalities or 8 00:00:26.240 --> 00:00:30.600 how you work through the course and what kind of tools we use, how you 9 00:00:30.600 --> 00:00:33.360 grade it, what kind of credits you get and so on. 10 00:00:33.360 --> 00:00:36.200 Let's get to business. So what is this course about? 11 00:00:36.200 --> 00:00:39.260 What will you learn on the basic course? 12 00:00:39.260 --> 00:00:45.040 My idea was to teach you basics of research design and teach you how to use regression 13 00:00:45.040 --> 00:00:50.830 analysis and factor analysis and some other very basic tools and understand them on 14 00:00:50.830 --> 00:00:56.720 more fundamental level so that you can apply them in an informed way instead of simply 15 00:00:56.720 --> 00:01:00.590 following this kind of cookbook or recipe approach. 16 00:01:00.590 --> 00:01:05.610 On this course we will expand from that lot, so we will 17 00:01:05.610 --> 00:01:08.630 go through some of the more advanced analysis. 18 00:01:08.630 --> 00:01:14.550 So we will do structural equation modelling, multi level modelling, missing data and so on 19 00:01:14.550 --> 00:01:18.970 and we will try to understand these techniques or more fundamental level. 20 00:01:18.970 --> 00:01:25.750 Going beyond this kind of like procedures and checklists and best practise 21 00:01:25.750 --> 00:01:28.890 isn't really understand what these techniques are about. 22 00:01:28.890 --> 00:01:30.750 What are the assumptions they are based on? 23 00:01:30.750 --> 00:01:35.520 What are the principles they're based on, and how they should be used in an informed 24 00:01:35.520 --> 00:01:39.950 way instead of simply following what others have been doing. 25 00:01:39.950 --> 00:01:44.710 After this course you should be able to carry out this analysis efficiently, 26 00:01:44.710 --> 00:01:50.270 effectively and reproducibly using using R or Stata. 27 00:01:50.270 --> 00:01:53.490 I don't support SPSS for two reasons. 28 00:01:53.490 --> 00:01:58.230 While you could do some of this analysis with SVSS, most of the things that 29 00:01:58.230 --> 00:02:01.890 we do on this course are simply not possible with that software. 30 00:02:01.890 --> 00:02:08.850 Also I have no idea how to do for example multi level model with spaces, so be warned. 31 00:02:08.850 --> 00:02:13.910 This is about advanced staff using RRSP or data and these 32 00:02:13.910 --> 00:02:16.140 are the software that you should be using. 33 00:02:16.140 --> 00:02:20.700 You can also do some of the work with M Plus, and if any of the students 34 00:02:20.700 --> 00:02:24.610 want to do that, I can provide the instruction on that. 35 00:02:24.610 --> 00:02:28.710 Another important thing beyond effectively, efficiently 36 00:02:28.710 --> 00:02:32.330 and is the reproducibility of the analysis. 37 00:02:32.330 --> 00:02:37.830 It is very important, and this has been highlighted in many, many different articles recently, 38 00:02:37.830 --> 00:02:42.690 that we are able to reproduce the analysis that we present in an article. 39 00:02:42.690 --> 00:02:47.400 So you should always document every single step from the 40 00:02:47.400 --> 00:02:51.690 raw data to the final table in an article. 41 00:02:51.690 --> 00:02:55.390 The ideal way of accomplishing that is through automation. 42 00:02:55.390 --> 00:03:01.210 So we will learn how to construct your state of New file or our file in way 43 00:03:01.210 --> 00:03:06.630 that it starts by loading the data set and then exports all tables and then we 44 00:03:06.630 --> 00:03:11.010 can just copy, paste or link that table to Word document. 45 00:03:11.010 --> 00:03:13.970 So there is no manual work involved. 46 00:03:13.970 --> 00:03:21.640 This automation guarantees that your results will be reproducible in way that you can always 47 00:03:21.640 --> 00:03:27.030 show the exact steps that you took from the raw data to the final output. 48 00:03:27.030 --> 00:03:29.190 This is this is really important. 49 00:03:29.190 --> 00:03:31.960 We'll go through why later on the course. 50 00:03:31.960 --> 00:03:38.300 Then we'll go through reporting in contrast to the basic course where you did 51 00:03:38.300 --> 00:03:42.640 one to three data analysis assignments and and wrote reports. 52 00:03:42.640 --> 00:03:50.300 On this course we will focus on different kinds of interpretations of analysis techniques 53 00:03:50.300 --> 00:03:54.480 and how to do, for example, graphical displays of results. 54 00:03:54.480 --> 00:04:01.760 So the idea here is that we try to make our results understandable also to people 55 00:04:01.760 --> 00:04:06.370 who don't know the method that we're applying, because if we are realistic, if 56 00:04:06.370 --> 00:04:12.550 we use for example our Allano Bond estimator for dynamic panel models, not many 57 00:04:12.550 --> 00:04:15.510 researchers understand what that is about. 58 00:04:15.510 --> 00:04:23.140 Finally, you should be able to review articles using basics, basic and applied techniques 59 00:04:23.140 --> 00:04:28.750 and use be able to read and produce methodological evidence. 60 00:04:28.750 --> 00:04:32.490 This final point is important because your career as professional 61 00:04:32.490 --> 00:04:36.750 researcher does not stop on this course and you'll probably have to 62 00:04:36.750 --> 00:04:40.590 learn some techniques that go beyond this course. 63 00:04:40.590 --> 00:04:44.730 For example, there might be specialised techniques that are applicable 64 00:04:44.730 --> 00:04:47.810 to certain niches that I don't cover on the course. 65 00:04:47.810 --> 00:04:52.460 Or there might be new techniques that come out after you have taken this course. 66 00:04:52.460 --> 00:04:55.640 You need to understand what is the evidence behind those 67 00:04:55.640 --> 00:04:58.140 techniques and be able to read that evidence. 68 00:04:58.140 --> 00:05:02.100 Because not all techniques are not all particularly guideline 69 00:05:02.100 --> 00:05:04.340 type articles are evidence based. 70 00:05:04.340 --> 00:05:08.040 They might be based simply on observing with researchers too, 71 00:05:08.040 --> 00:05:10.940 and then presenting that as best practise. 72 00:05:10.940 --> 00:05:16.040 But as we will find on this course, and as you've seen perhaps on the basic 73 00:05:16.040 --> 00:05:20.040 course, not everything that researchers do is justified. 74 00:05:20.040 --> 00:05:24.700 People make mistakes, and sometimes a mistake is institutionalised 75 00:05:24.700 --> 00:05:26.500 as the best practise. 76 00:05:26.500 --> 00:05:30.740 We need to understand evidence to be able to avoid that problem. 77 00:05:30.740 --> 00:05:36.120 The best way to understand evidence is to be able to produce evidence yourself, 78 00:05:36.120 --> 00:05:40.060 and for this reason we'll do simulations on the course. 79 00:05:40.060 --> 00:05:45.160 We start by simple simulations, just simulate 1 sample, run ricos analysis, 80 00:05:45.160 --> 00:05:50.320 see that we get the correct result from large sample size. 81 00:05:50.320 --> 00:05:52.360 We also do Monte Carlo simulations. 82 00:05:52.360 --> 00:05:58.790 Extend that simple case by simulating multiple samples and we will see that regression analysis 83 00:05:58.790 --> 00:06:04.510 for example, produces the correct result on average from small samples. It's unbiased. 84 00:06:04.510 --> 00:06:06.830 We will learn these kind of things and this is the reason 85 00:06:06.830 --> 00:06:09.170 why we do simulations on the course. 86 00:06:09.170 --> 00:06:14.290 Personally, I do simulations all the time, not only for methodological research, 87 00:06:14.290 --> 00:06:17.890 but also if I want to understand how technique works. 88 00:06:17.890 --> 00:06:21.890 I simulate sample and then I try the technique on that sample. 89 00:06:21.890 --> 00:06:26.020 If I get the correct result, then I can be more confident 90 00:06:26.020 --> 00:06:29.210 that I'm using the technique correctly. 91 00:06:29.210 --> 00:06:32.490 Now let's go through the the schedule and the content. 92 00:06:32.490 --> 00:06:38.730 The course runs for full academic year starting from September and ending in May. 93 00:06:38.730 --> 00:06:40.530 We have 10 units. 94 00:06:40.530 --> 00:06:44.390 The first unit like in the basic course is the introduction and the purpose 95 00:06:44.390 --> 00:06:47.510 of the introduction is just to get to know each other. 96 00:06:47.510 --> 00:06:52.450 Familiarise you with what this course is about, How do we work on the course 97 00:06:52.450 --> 00:06:56.330 and you will also do pre exam which covers some books. 98 00:06:56.330 --> 00:06:59.690 I'll explain that in another video. 99 00:06:59.690 --> 00:07:04.390 Then are we will have 8 normal units units 2:00 to 9:00 100 00:07:04.390 --> 00:07:07.100 that address different analysis techniques. 101 00:07:07.100 --> 00:07:13.040 We'll start with more research design related things, more fundamental foundational 102 00:07:13.040 --> 00:07:18.700 things like causality and structuralism modelling which is special generalisation 103 00:07:18.700 --> 00:07:22.630 of many of the techniques that we work on on the course. 104 00:07:22.630 --> 00:07:27.370 And then we'll go through more advanced course building up to multi level models 105 00:07:27.370 --> 00:07:31.510 which forms the foundation for longitudinal data analysis and so on. 106 00:07:31.510 --> 00:07:35.150 Then at the end of the course we will have student 107 00:07:35.150 --> 00:07:37.400 presentations and course conclusions. 108 00:07:37.400 --> 00:07:41.130 Every student on this course makes 2 presentations. 109 00:07:41.130 --> 00:07:47.270 You present results of 1 Monte Carlo simulation that you did on your own, 110 00:07:47.270 --> 00:07:51.710 of course with the help of other students and myself. 111 00:07:51.710 --> 00:07:55.930 And then you will also get to pick a methodological article that we have not 112 00:07:55.930 --> 00:07:59.910 covered on the course and then present that article to others. 113 00:07:59.910 --> 00:08:05.150 The purpose here is to teach you how to evaluate evidence again, 114 00:08:05.150 --> 00:08:08.170 and the best way is to produce and present evidence. 115 00:08:08.170 --> 00:08:11.030 Then you really start to understand what the evidence is 116 00:08:11.030 --> 00:08:13.230 is about and how it should be interpreted. 117 00:08:13.230 --> 00:08:17.910 Finally we have final exam where you get to put your skills to the test and 118 00:08:17.910 --> 00:08:21.130 I can guarantee that that is going to be pretty challenging. 119 00:08:21.130 --> 00:08:26.450 You don't have to get it all correctly to get 5, but if you get it all correctly, you 120 00:08:26.450 --> 00:08:31.720 are probably in at least top 1% of quantitative researchers in the world. 121 00:08:31.720 --> 00:08:35.130 So welcome to the course and in the following videos I will 122 00:08:35.130 --> 00:08:37.620 explain the practicalities in more detail.