1 00:00:00,000 --> 00:00:01,320 INSTRUCTOR: Single-case design. 2 00:00:01,320 --> 00:00:05,380 This is probably one of the most common types of quasi-experimental designs, 3 00:00:05,380 --> 00:00:08,980 and sometimes it gets a little confused with case studies for obvious reasons. 4 00:00:08,980 --> 00:00:11,545 It includes single case design. 5 00:00:11,545 --> 00:00:13,520 It sounds like a case study. 6 00:00:13,520 --> 00:00:15,380 But unlike a case study, 7 00:00:15,380 --> 00:00:17,980 so case studies usually just describe people. 8 00:00:17,980 --> 00:00:23,090 There's no real manipulation or experiment going on, 9 00:00:23,090 --> 00:00:25,820 there's no exposure to some type of stimulus, 10 00:00:25,820 --> 00:00:27,430 and then to measure a reaction. 11 00:00:27,430 --> 00:00:29,160 So with an actual case study, 12 00:00:29,160 --> 00:00:30,860 you're just following them in their life, 13 00:00:30,860 --> 00:00:33,460 seeing how they react normally to things, 14 00:00:33,460 --> 00:00:36,450 and describing that very special case. 15 00:00:36,450 --> 00:00:38,190 With a single-case design, 16 00:00:38,190 --> 00:00:39,765 again, this is quasi-experimental, 17 00:00:39,765 --> 00:00:43,515 you are looking at manipulating them in some way oftentimes. 18 00:00:43,515 --> 00:00:47,265 How does someone with depression react to this type of therapy? 19 00:00:47,265 --> 00:00:52,380 Or how does a pack-a-day smoker react to this type of medication or treatment? 20 00:00:52,380 --> 00:00:56,790 Again, you are running a type of experiment, it's a quasi-experiment. 21 00:00:56,790 --> 00:01:00,135 You're not just following these people around and recording their everyday lives, 22 00:01:00,135 --> 00:01:02,340 you're trying to change them in some way, 23 00:01:02,340 --> 00:01:06,255 usually, or treat them is maybe a better way of saying it. 24 00:01:06,255 --> 00:01:10,740 Also, single case, and this is where I don't like the way this is named— 25 00:01:10,740 --> 00:01:14,990 it's misleading—it implies that you're following one person. 26 00:01:14,990 --> 00:01:19,000 That's really never the case with a single-case design. 27 00:01:19,000 --> 00:01:22,330 The reason why we do call it a single-case design is because 28 00:01:22,330 --> 00:01:26,190 we're usually analyzing data on an individual basis. 29 00:01:26,190 --> 00:01:30,180 Pretty much all the analyses that we've been doing so far have been on group levels. 30 00:01:30,180 --> 00:01:32,530 Comparing group averages, seeing if there's 31 00:01:32,530 --> 00:01:35,595 a correlation between two variables on average. 32 00:01:35,595 --> 00:01:38,300 The difference with a single-case design is you're 33 00:01:38,300 --> 00:01:41,615 measuring change in individuals over time. 34 00:01:41,615 --> 00:01:48,440 You often hear things like two out of three people benefited from this therapy. 35 00:01:48,440 --> 00:01:52,025 Well, that's a quasi-experimental single-case statement. 36 00:01:52,025 --> 00:01:54,320 You're measuring things on an individual basis. 37 00:01:54,320 --> 00:01:57,020 You're not saying, on average, people got 30% better, 38 00:01:57,020 --> 00:02:00,380 you're saying two out of three people got better. 39 00:02:01,610 --> 00:02:05,095 I'm going to talk about one of the most common types of 40 00:02:05,095 --> 00:02:09,745 single-case designs which involves some type of pre-post measure. 41 00:02:09,745 --> 00:02:13,830 Usually, you're measuring the DV at a baseline. 42 00:02:13,830 --> 00:02:15,860 I'll go and put this up at a baseline period. 43 00:02:15,860 --> 00:02:18,530 We call that baseline because it's before manipulation, 44 00:02:18,530 --> 00:02:20,040 it's just the regular level. 45 00:02:20,040 --> 00:02:21,920 Then after the manipulation, 46 00:02:21,920 --> 00:02:23,600 usually what we call a treatment period, 47 00:02:23,600 --> 00:02:27,935 because some type of treatment or effect has been undertaken. 48 00:02:27,935 --> 00:02:29,820 A real classic one, 49 00:02:29,820 --> 00:02:31,650 one that's received a lot of interest lately as 50 00:02:31,650 --> 00:02:36,210 a possible treatment for depression is called deep brain stimulation. 51 00:02:36,210 --> 00:02:40,350 This is like having a pacemaker attached to your brain. 52 00:02:40,350 --> 00:02:42,330 In a real way, it really is like that. 53 00:02:42,330 --> 00:02:46,365 You have a pacemaker battery-like device implanted behind your collarbone, 54 00:02:46,365 --> 00:02:49,630 and then electrodes go into very deep centers of your brain, 55 00:02:49,630 --> 00:02:52,845 especially in the limbic system where depression is thought to originate. 56 00:02:52,845 --> 00:02:57,040 These deep brain stimulations actually shut down 57 00:02:57,040 --> 00:03:01,495 or calm down areas of the brain that might be involved with depression. 58 00:03:01,495 --> 00:03:04,500 In 2005, this was a very experimental manipulation. 59 00:03:04,500 --> 00:03:08,895 This had previously only been done with lab rats and some types of monkeys. 60 00:03:08,895 --> 00:03:11,220 What they did, they wanted to test it on sick, 61 00:03:11,220 --> 00:03:13,420 severely depressed patients, which is 62 00:03:13,420 --> 00:03:16,020 what we often do with new medications or treatments. 63 00:03:16,020 --> 00:03:17,710 We pick those people who, 64 00:03:17,710 --> 00:03:20,040 for lack of a better term, have nothing to lose. 65 00:03:20,040 --> 00:03:22,250 They're at the bottom of the barrel. 66 00:03:22,250 --> 00:03:26,050 Really, they could use any type of possible treatment. 67 00:03:26,050 --> 00:03:30,560 Basically, beforehand, they took measures of the depression as using 68 00:03:30,560 --> 00:03:32,460 themselves as a comparison group to see if 69 00:03:32,460 --> 00:03:34,860 they get better over time with this new type of treatment. 70 00:03:34,860 --> 00:03:37,260 Then, of course, they undergo the surgery necessary to 71 00:03:37,260 --> 00:03:40,825 implant this deep brain stimulation device. 72 00:03:40,825 --> 00:03:45,760 Again, we're looking at individual changes in depression, not group-level changes. 73 00:03:45,760 --> 00:03:48,400 That's part of the single-case design. 74 00:03:48,400 --> 00:03:50,680 With some of the individuals, they found that it really 75 00:03:50,680 --> 00:03:53,835 made not much of a difference or no significant difference. 76 00:03:53,835 --> 00:03:56,100 Maybe they only went down a depression score, 77 00:03:56,100 --> 00:04:00,405 maybe from eight to seven on a one to ten depression score, 78 00:04:00,405 --> 00:04:04,020 whereas other people showed really dramatic decreases in depression, 79 00:04:04,020 --> 00:04:06,780 in some cases alleviating depression all the way. 80 00:04:06,780 --> 00:04:09,685 Again, we're looking at this in an individual basis. 81 00:04:09,685 --> 00:04:12,100 Each line represents a different person. 82 00:04:12,100 --> 00:04:14,870 In the end, the data looks something like this with 83 00:04:14,870 --> 00:04:19,590 basically two people showing negligible benefits or non-significant benefits, 84 00:04:19,590 --> 00:04:21,110 thankfully not getting any worse, 85 00:04:21,110 --> 00:04:26,790 but no real change and four of the six showing good benefits and in some cases, 86 00:04:26,790 --> 00:04:28,830 even complete alleviation from depression, 87 00:04:28,830 --> 00:04:30,855 which is, of course, what you'd want to see. 88 00:04:30,855 --> 00:04:34,380 Again, that's looking at an individual basis here. 89 00:04:34,380 --> 00:04:38,560 It's a quasi-experiment because there was no control group. 90 00:04:38,560 --> 00:04:41,200 A true control group in this case, 91 00:04:41,200 --> 00:04:43,990 and it would be unethical which is just why it wasn't done, 92 00:04:43,990 --> 00:04:48,700 would be to have people undergo surgery to implant this device, which is, of course, 93 00:04:48,700 --> 00:04:51,060 very risky, and then basically to 94 00:04:51,060 --> 00:04:54,495 think the device is on but to never actually turn it on. 95 00:04:54,495 --> 00:04:58,010 In this case, that's just too much to ask for a control group. 96 00:04:58,010 --> 00:05:01,320 Again, depression is a very detrimental thing. 97 00:05:01,320 --> 00:05:03,160 These are people who are severely depressed, 98 00:05:03,160 --> 00:05:06,400 so it would be unethical to withhold treatment from them. 99 00:05:06,400 --> 00:05:08,560 But then without this control group, 100 00:05:08,560 --> 00:05:10,360 there's lots of confounds that may have 101 00:05:10,360 --> 00:05:13,630 alleviated depression for these four out of the six individuals. 102 00:05:13,630 --> 00:05:16,300 The biggest one being a placebo effect. 103 00:05:16,300 --> 00:05:20,460 Lots and lots of studies show that when people think they're being treated, 104 00:05:20,460 --> 00:05:23,380 oftentimes, their symptoms will be lessened. 105 00:05:23,380 --> 00:05:30,510 This can be anything from an actual treatment to just a sugar pill or a placebo pill. 106 00:05:30,510 --> 00:05:32,440 Even with no treatment whatsoever, 107 00:05:32,440 --> 00:05:36,720 just the thought of being treated can oftentimes help people that mind over matter, 108 00:05:36,720 --> 00:05:38,480 especially with mental illness. 109 00:05:38,480 --> 00:05:40,840 If someone thinks they're being treated, they might think, well, 110 00:05:40,840 --> 00:05:43,080 I think I'm less depressed than I was before, 111 00:05:43,080 --> 00:05:45,675 but they're just thinking more optimistically. 112 00:05:45,675 --> 00:05:51,090 Another confound, and this is especially prevalent with research on the mentally 113 00:05:51,090 --> 00:05:52,810 ill is that lots of people with 114 00:05:52,810 --> 00:05:55,930 mental illness actually get over their mental illness on their own. 115 00:05:55,930 --> 00:05:59,110 This is something that we therapists don't talk about a lot because, of course, 116 00:05:59,110 --> 00:06:02,530 we like to tout that therapy and medication are really good things, which they are. 117 00:06:02,530 --> 00:06:07,090 They do help people oftentimes get over mental illness sooner than they would otherwise. 118 00:06:07,090 --> 00:06:09,230 But typically, about 60% of people with 119 00:06:09,230 --> 00:06:12,030 mental illness do get over it completely on their own. 120 00:06:12,030 --> 00:06:14,170 This includes anything from depression, 121 00:06:14,170 --> 00:06:16,830 usually not severe things like schizophrenia or bipolar, 122 00:06:16,830 --> 00:06:20,335 but depression, lots of mood disorders included. 123 00:06:20,335 --> 00:06:23,380 Now, this is unlikely in this group, 124 00:06:23,380 --> 00:06:25,680 though, especially at a 60% level, 125 00:06:25,680 --> 00:06:29,110 because they selected people with severe and prolonged depression, 126 00:06:29,110 --> 00:06:32,810 people who have had it for years, were previously untreatable. 127 00:06:32,810 --> 00:06:36,120 It is unlikely that these individuals would just snap out of 128 00:06:36,120 --> 00:06:40,640 it in the first initial weeks of this treatment. 129 00:06:42,260 --> 00:06:44,720 Again, I don't think this is such a problem in 130 00:06:44,720 --> 00:06:47,020 this study because these were very severe patients. 131 00:06:47,020 --> 00:06:50,920 But sometimes just getting a baseline measure of mental illness 132 00:06:50,920 --> 00:06:55,850 can help people motivate themselves to make changes in their life to get better. 133 00:06:55,850 --> 00:06:58,600 Especially for people with maybe mild to moderate depression, 134 00:06:58,600 --> 00:07:00,940 taking a depression inventory where they're listing 135 00:07:00,940 --> 00:07:03,940 all the things that's been wrong with them so obviously, 136 00:07:03,940 --> 00:07:05,760 depressed mood, maybe insomnia, 137 00:07:05,760 --> 00:07:09,120 not eating enough, isolation from friends and family. 138 00:07:09,120 --> 00:07:11,630 People might take that as a wake-up call and go, 139 00:07:11,630 --> 00:07:14,300 wow, I really need to eat more, 140 00:07:14,300 --> 00:07:15,840 and I need to rest more, 141 00:07:15,840 --> 00:07:18,200 and I need to contact my friends and family and actually 142 00:07:18,200 --> 00:07:21,040 can help them to snap themselves out of that depression. 143 00:07:21,040 --> 00:07:23,960 Again, I don't think this is a problem here because these were, 144 00:07:23,960 --> 00:07:26,990 again, very severely depressed patients. 145 00:07:26,990 --> 00:07:31,000 Probably is a case that this wouldn't play that big of a role. 146 00:07:31,250 --> 00:07:34,695 But one way to deal with these confounds, 147 00:07:34,695 --> 00:07:37,020 is to do what's called a reversal design. 148 00:07:37,020 --> 00:07:40,600 It's basically if you see an effect, which they did here, 149 00:07:40,600 --> 00:07:45,200 four out of six patients showed positive results, 150 00:07:45,200 --> 00:07:47,820 basically alleviations in their depression scores, 151 00:07:47,820 --> 00:07:49,380 then you might want to, 152 00:07:49,380 --> 00:07:51,500 in essence, take away the manipulation, 153 00:07:51,500 --> 00:07:54,620 take away the treatment, and see if they return back to their baseline scores, 154 00:07:54,620 --> 00:07:58,825 see if those four out of six return back to depressed scores. 155 00:07:58,825 --> 00:08:01,995 Now, this is called an ABA design. 156 00:08:01,995 --> 00:08:05,790 In this case, A really just means what's happening before their baseline and B, 157 00:08:05,790 --> 00:08:07,955 what happens after their treatment condition. 158 00:08:07,955 --> 00:08:10,710 ABA, going from baseline to treatment, 159 00:08:10,710 --> 00:08:14,405 and then hopefully back to baseline once you remove the treatment. 160 00:08:14,405 --> 00:08:17,565 They did this with one of the individuals. 161 00:08:17,565 --> 00:08:19,590 Whether or not this is ethical, 162 00:08:19,590 --> 00:08:22,810 this is a gray area, 163 00:08:22,810 --> 00:08:24,610 but at this point, this is one of the only ways to 164 00:08:24,610 --> 00:08:26,440 really rule out things like placebo effect. 165 00:08:26,440 --> 00:08:28,670 With one of the individuals who was showing 166 00:08:28,670 --> 00:08:32,090 a very positive result from the deep brain stimulation, they basically, 167 00:08:32,090 --> 00:08:34,550 because these people got a week-by-week checkup, 168 00:08:34,550 --> 00:08:38,670 where they fine-tuned the pacemaker inside them and all that, 169 00:08:38,670 --> 00:08:42,835 with this individual, they basically turned it off without telling him. 170 00:08:42,835 --> 00:08:45,130 Low and behold, within a couple of weeks, 171 00:08:45,130 --> 00:08:47,160 the patient returned back to baseline. 172 00:08:47,160 --> 00:08:49,470 This is a good indication 173 00:08:49,470 --> 00:08:53,050 that this is the actual treatment that's doing it and not just a placebo effect, 174 00:08:53,050 --> 00:08:54,230 because for all this person knows, 175 00:08:54,230 --> 00:08:55,600 they're still receiving treatment, 176 00:08:55,600 --> 00:08:58,510 so it rules out things like placebo effect, 177 00:08:58,510 --> 00:09:04,550 but also things like just self-recovery from depression. 178 00:09:04,700 --> 00:09:08,535 Of course, this is maybe unethical, 179 00:09:08,535 --> 00:09:11,730 because now you're withholding treatment for someone who really benefits from it. 180 00:09:11,730 --> 00:09:15,750 What most often they do with these types of designs is not just an ABA design, 181 00:09:15,750 --> 00:09:17,470 but an ABAB design. 182 00:09:17,470 --> 00:09:20,190 If they find that a person does return back to baseline, 183 00:09:20,190 --> 00:09:26,515 they again instill the treatment as soon as they start showing symptoms back to baseline. 184 00:09:26,515 --> 00:09:28,850 Hopefully, so that this individual doesn't 185 00:09:28,850 --> 00:09:31,605 have to suffer for any prolonged amount of time. 186 00:09:31,605 --> 00:09:33,750 In this case, again, they did this. 187 00:09:33,750 --> 00:09:36,700 They restarted the deep brain stimulation, 188 00:09:36,700 --> 00:09:38,220 again, without this person's knowledge, 189 00:09:38,220 --> 00:09:40,600 and within even a shorter period of time, 190 00:09:40,600 --> 00:09:44,415 a few days, he returned back to his treatment levels. 191 00:09:44,415 --> 00:09:48,690 This is even better for the whole anti-placebo argument because, 192 00:09:48,690 --> 00:09:49,880 again, this is without his knowledge. 193 00:09:49,880 --> 00:09:51,835 It's being turned on and off. 194 00:09:51,835 --> 00:09:57,230 This is a replication of the effect multiple times in the same individual. 195 00:09:57,230 --> 00:10:01,060 Really good indication that it really is the deep brain stimulation and not something 196 00:10:01,060 --> 00:10:03,460 else just random changes in the person 197 00:10:03,460 --> 00:10:07,120 over time or placebo effect causing this difference. 198 00:10:11,170 --> 00:10:14,710 [NOISE] Just some characteristics of the single-case design and 199 00:10:14,710 --> 00:10:17,725 a good way to differentiate it from regular types of experiments. 200 00:10:17,725 --> 00:10:19,270 For one, of course, 201 00:10:19,270 --> 00:10:21,830 it examines people individually as opposed to a group, 202 00:10:21,830 --> 00:10:24,110 so it's not going to talk about things like group averages. 203 00:10:24,110 --> 00:10:27,710 It's going to be saying more like four out of six people, for instance, 204 00:10:27,710 --> 00:10:31,925 in this case, did show significant improvement in their depression scores. 205 00:10:31,925 --> 00:10:35,025 Most of them do employ this reversal of design so 206 00:10:35,025 --> 00:10:38,280 either an ABA or an ABAB design. 207 00:10:38,280 --> 00:10:41,370 Of course, what makes it a quasi-experiment is, 208 00:10:41,370 --> 00:10:42,790 well, in this case, very clearly, 209 00:10:42,790 --> 00:10:44,310 there was no control group. 210 00:10:44,310 --> 00:10:47,050 Without a control group, you necessarily can't have 211 00:10:47,050 --> 00:10:50,970 experimental control or randomization because those require multiple groups. 212 00:10:50,970 --> 00:10:52,710 But even with multiple groups, 213 00:10:52,710 --> 00:10:54,150 usually with a quasi-experiment, 214 00:10:54,150 --> 00:10:57,450 you don't have any type of true control over 215 00:10:57,450 --> 00:11:01,590 what patients are experiencing other than the independent variable or the treatment, 216 00:11:01,590 --> 00:11:03,300 and there's no random assignment, 217 00:11:03,300 --> 00:11:07,115 because usually, you're dealing with person factors like depression. 218 00:11:07,115 --> 00:11:10,970 Most often these things are used to test the effectiveness of treatments. 219 00:11:10,970 --> 00:11:13,930 They're very common in therapeutic designs, 220 00:11:13,930 --> 00:11:17,030 especially involved with medications.