1 00:00:02,550 --> 00:00:06,350 - [Instructor] Hello, and welcome to this video lecture, 2 00:00:06,350 --> 00:00:08,613 which is on Research Design. 3 00:00:13,140 --> 00:00:15,260 We'll be talking about some of the options 4 00:00:15,260 --> 00:00:17,070 that our researchers have. 5 00:00:17,070 --> 00:00:22,070 First of all time, how long, and in what way over time 6 00:00:22,810 --> 00:00:25,490 does the researcher collect data? 7 00:00:25,490 --> 00:00:29,160 And then I wanna walk you through the basics of project 8 00:00:29,160 --> 00:00:30,933 and proposal design. 9 00:00:34,240 --> 00:00:37,710 So when I'm talking about time dimension, 10 00:00:37,710 --> 00:00:41,943 I'm talking about collecting data over time. 11 00:00:46,970 --> 00:00:50,140 There's two main options cross-section, 12 00:00:50,140 --> 00:00:51,990 and then longitudinal, 13 00:00:51,990 --> 00:00:56,920 and a number of options under longitudinal. 14 00:00:56,920 --> 00:01:01,133 And note that as I go through each of these, 15 00:01:02,060 --> 00:01:05,490 that I'm using the notation, 16 00:01:05,490 --> 00:01:10,310 that T is the number of years or number of time periods 17 00:01:10,310 --> 00:01:12,930 that you are collecting data, 18 00:01:12,930 --> 00:01:15,763 and N is the number of subjects. 19 00:01:21,010 --> 00:01:22,390 The first one, and maybe the one 20 00:01:22,390 --> 00:01:26,340 you're most familiar with, is cross-section. 21 00:01:26,340 --> 00:01:31,123 And this is when T equals one that you're collecting data 22 00:01:33,240 --> 00:01:35,810 at one time and one point in time, 23 00:01:35,810 --> 00:01:38,400 it's sort of a snapshot in time. 24 00:01:38,400 --> 00:01:42,450 The survey that we'll be doing in class is cross-section, 25 00:01:42,450 --> 00:01:45,070 but you have many subjects, 26 00:01:45,070 --> 00:01:48,510 so you maybe do a survey and go out 27 00:01:48,510 --> 00:01:53,510 and collect data from a large N number of subjects, 28 00:01:55,500 --> 00:01:57,000 but just one time. 29 00:01:57,000 --> 00:02:00,673 So T equals one, N equals many is cross-section. 30 00:02:04,040 --> 00:02:06,730 The rest of them will be longitudinal 31 00:02:06,730 --> 00:02:09,410 where you don't just collect data one time, 32 00:02:09,410 --> 00:02:14,410 you collect data many times over specific points in time. 33 00:02:14,670 --> 00:02:17,360 And in many cases it can be graphed. 34 00:02:17,360 --> 00:02:20,700 So very often, when we do this, 35 00:02:20,700 --> 00:02:25,700 we look at the time being on the x-axis 36 00:02:26,320 --> 00:02:31,180 and the variable of interest being on the y-axis. 37 00:02:31,180 --> 00:02:32,537 And you can see that 38 00:02:36,490 --> 00:02:38,530 graduation rate here, 39 00:02:38,530 --> 00:02:42,330 it increases real greatly over time 40 00:02:42,330 --> 00:02:45,460 and then sort of flattens out over the last, 41 00:02:45,460 --> 00:02:47,643 looks like 60 years or so. 42 00:02:52,500 --> 00:02:57,110 So one common longitudinal strategy is time series, 43 00:02:57,110 --> 00:03:00,830 and this is where N equals one, but T equals many. 44 00:03:00,830 --> 00:03:04,420 So here, we're looking at one variable, 45 00:03:04,420 --> 00:03:08,040 one phenomenon over time. 46 00:03:08,040 --> 00:03:12,610 So it could be prices over time, 47 00:03:12,610 --> 00:03:16,350 GDP is a very well known one, 48 00:03:16,350 --> 00:03:20,560 inflation, it could be things like graduation rates 49 00:03:20,560 --> 00:03:24,983 at the university, tuition rates at the university, 50 00:03:25,900 --> 00:03:30,760 carbon in the atmosphere, there's many, many options, 51 00:03:30,760 --> 00:03:34,720 but is a single variable, 52 00:03:34,720 --> 00:03:39,720 looking at a single subject like the economy, 53 00:03:40,660 --> 00:03:42,123 but over many years. 54 00:03:49,360 --> 00:03:52,070 Here's the first one where both T and N 55 00:03:52,070 --> 00:03:53,840 are greater than one. 56 00:03:53,840 --> 00:03:56,690 So here, this is a trend study. 57 00:03:56,690 --> 00:04:01,690 So here we're measuring how does a given attitude 58 00:04:01,750 --> 00:04:04,373 or belief change over time. 59 00:04:06,390 --> 00:04:09,340 Usually it involves sampling. 60 00:04:09,340 --> 00:04:13,060 So every time we draw a new sample, 61 00:04:13,060 --> 00:04:17,730 so it might be Vermont residents, UVM students, 62 00:04:17,730 --> 00:04:21,000 small business owners, but the key here 63 00:04:21,000 --> 00:04:24,240 is that we're drawing a different sample every time. 64 00:04:24,240 --> 00:04:29,240 So if we are looking at the Vermont poll 65 00:04:30,050 --> 00:04:34,800 and ask the same question over time, 66 00:04:34,800 --> 00:04:37,960 and here are just some I thought of, gay marriage, 67 00:04:37,960 --> 00:04:41,330 handgun restrictions maybe 68 00:04:41,330 --> 00:04:45,997 what's the most important problem 69 00:04:46,880 --> 00:04:48,523 facing the state. 70 00:04:50,260 --> 00:04:55,260 So again, it's measuring a trend of a variable, 71 00:04:56,110 --> 00:05:00,730 but looking both over time period 72 00:05:00,730 --> 00:05:04,540 and drawing data from many people, 73 00:05:04,540 --> 00:05:06,920 but with a changing sample each time. 74 00:05:06,920 --> 00:05:11,920 So each time we do this such as the Vermont poll, 75 00:05:12,150 --> 00:05:13,853 we have a new sample. 76 00:05:18,580 --> 00:05:23,150 A cohort study is like that except now, 77 00:05:23,150 --> 00:05:25,640 we are drawing from a specific cohort. 78 00:05:25,640 --> 00:05:29,580 So again, T equals many and N is still many, 79 00:05:29,580 --> 00:05:32,110 but we're sampling from a cohort. 80 00:05:32,110 --> 00:05:35,500 And cohort is a group of people 81 00:05:35,500 --> 00:05:38,150 who have something in common. 82 00:05:38,150 --> 00:05:42,920 So it might be the UVM class of 2021, 83 00:05:42,920 --> 00:05:47,500 or it might be people who graduated high school in 1983, 84 00:05:47,500 --> 00:05:48,713 which would be me. 85 00:05:50,320 --> 00:05:52,693 And in a cohort study, 86 00:05:53,530 --> 00:05:57,980 if it was folks who graduated high school '83, 87 00:05:57,980 --> 00:06:01,840 they would draw a new sample each time, 88 00:06:01,840 --> 00:06:03,990 but from that same cohort. 89 00:06:03,990 --> 00:06:08,730 So they would draw a bunch of folks in one year. 90 00:06:08,730 --> 00:06:12,750 And the next year, all of us having in common 91 00:06:12,750 --> 00:06:16,570 that we graduated high school in the same year, 92 00:06:16,570 --> 00:06:20,950 but one year I may be in the sample, 93 00:06:20,950 --> 00:06:23,060 but in the next year, not so. 94 00:06:23,060 --> 00:06:26,230 Again, it's a new sample every time, 95 00:06:26,230 --> 00:06:29,243 but it's drawn from a cohort. 96 00:06:33,350 --> 00:06:35,910 And then there is panel studies. 97 00:06:35,910 --> 00:06:40,540 So here it's T equals many 98 00:06:40,540 --> 00:06:41,870 and N equals many. 99 00:06:41,870 --> 00:06:44,050 But it's the same people over time. 100 00:06:44,050 --> 00:06:49,050 So now we are following the same individual. 101 00:06:49,480 --> 00:06:54,480 So we interview or survey them over time, 102 00:06:55,410 --> 00:06:57,040 or it might be the same business 103 00:06:57,040 --> 00:07:01,800 or the same city 104 00:07:01,800 --> 00:07:04,173 or state or other. 105 00:07:06,020 --> 00:07:07,150 You didn't have analysis, 106 00:07:07,150 --> 00:07:10,980 but we track what they say over time. 107 00:07:10,980 --> 00:07:14,900 So we can actually measure how does an individual person 108 00:07:14,900 --> 00:07:19,900 or an individual city or business change over time, 109 00:07:21,200 --> 00:07:25,920 because we have data that are linked to that individual 110 00:07:26,770 --> 00:07:28,183 over many years. 111 00:07:33,900 --> 00:07:35,890 So just as a review, 112 00:07:35,890 --> 00:07:39,940 you should be able to compare and contrast 113 00:07:39,940 --> 00:07:41,933 these three strategies. 114 00:07:45,320 --> 00:07:46,820 That is the end of part one. 115 00:07:46,820 --> 00:07:48,923 Please go to part two, thank you.