1 00:00:00,280 --> 00:00:00,940 To prepare for 2 00:00:00,940 --> 00:00:06,045 an experience sampling research study, you will need to follow these steps. 3 00:00:06,045 --> 00:00:06,690 1., 4 00:00:06,690 --> 00:00:09,290 carefully think about the question. 5 00:00:09,290 --> 00:00:13,130 It has to be something about an action that repeats itself. 6 00:00:13,130 --> 00:00:14,740 A good example is, 7 00:00:14,740 --> 00:00:19,490 what was the reason you recently used a piece of paper to write something down? 8 00:00:19,490 --> 00:00:23,060 A bad example, what was the most annoying thing for you, 9 00:00:23,060 --> 00:00:25,930 the last time you moved to a new house or apartment? 10 00:00:27,110 --> 00:00:30,710 It has to be about behavior, not opinion. 11 00:00:30,710 --> 00:00:35,490 A good example, what was the reason you recently updated your website? 12 00:00:35,490 --> 00:00:37,130 And a bad example. 13 00:00:37,130 --> 00:00:41,100 What do you think about hiring a web developer for updating your website? 14 00:00:42,480 --> 00:00:44,640 No yes-no questions. 15 00:00:44,640 --> 00:00:46,010 A good example. 16 00:00:46,010 --> 00:00:48,440 What did you want to know recently? 17 00:00:48,440 --> 00:00:49,780 A bad example. 18 00:00:49,780 --> 00:00:50,850 Did you buy milk today? 19 00:00:52,350 --> 00:00:54,750 No quantitative questions. 20 00:00:54,750 --> 00:00:56,210 A good example. 21 00:00:56,210 --> 00:00:59,920 What was the reason for the last phone call you initiated? 22 00:00:59,920 --> 00:01:01,250 A bad example. 23 00:01:01,250 --> 00:01:03,526 How many phone calls did you receive in the past hour? 24 00:01:03,526 --> 00:01:05,303 2. 25 00:01:05,303 --> 00:01:06,320 [SOUND] How long? 26 00:01:06,320 --> 00:01:10,793 Carefully consider the time it'll take your study participants to 27 00:01:10,793 --> 00:01:14,850 put into answering your question each time you ask. 28 00:01:14,850 --> 00:01:18,710 Keep in mind you can definitely ask a lead question as well as 29 00:01:18,710 --> 00:01:20,610 some follow up questions. 30 00:01:20,610 --> 00:01:22,490 As a general rule of thumb, 31 00:01:22,490 --> 00:01:25,720 try not to take more than one minute of your participant's time. 32 00:01:26,760 --> 00:01:29,340 If you ask the question five times a day, 33 00:01:29,340 --> 00:01:32,810 the required effort of participation is five minutes a day. 34 00:01:32,810 --> 00:01:34,240 If you increase this time, 35 00:01:34,240 --> 00:01:37,830 you increase the likelihood of getting less answers from your participants. 36 00:01:38,980 --> 00:01:40,360 3. [NOISE] How many notifications? 37 00:01:40,360 --> 00:01:46,040 Another thing to consider is the number of notifications you send each day. 38 00:01:46,040 --> 00:01:49,640 The number should be a trade-off between the times you think the behavior you 39 00:01:49,640 --> 00:01:51,480 ask about is happening. 40 00:01:51,480 --> 00:01:53,200 And a number that would annoy or 41 00:01:53,200 --> 00:01:57,660 overwhelm your participants to a point they choose not to respond. 42 00:01:57,660 --> 00:02:02,580 For example, if you think people take notes on a piece of paper ten times a day. 43 00:02:02,580 --> 00:02:05,670 Ask them about it three to five times a day. 44 00:02:05,670 --> 00:02:07,360 Ten times is too much. 45 00:02:07,360 --> 00:02:12,270 Also your estimation might be wrong, and if you ask ten times a day about something 46 00:02:12,270 --> 00:02:17,320 that happens only five times a day your participants will feel uncomfortable. 47 00:02:17,320 --> 00:02:21,390 Another example is about behaviors that occur relatively rarely. 48 00:02:21,390 --> 00:02:26,190 For example, if you estimate your audience updates their website twice a week. 49 00:02:26,190 --> 00:02:28,270 Don't ask about it every day. 50 00:02:28,270 --> 00:02:32,420 My recommendation in this case would be to ask the question once a week for 51 00:02:32,420 --> 00:02:34,350 a period of ten weeks. 52 00:02:34,350 --> 00:02:39,690 If you do that with 50 people, you get 500 reasons people update their sites. 53 00:02:39,690 --> 00:02:41,700 That's a good number of data points to learn from. 54 00:02:42,800 --> 00:02:43,910 4. 55 00:02:43,910 --> 00:02:46,260 How long should the study last? 56 00:02:46,260 --> 00:02:50,740 The answer is based on a combination of the number of times you ask per day and 57 00:02:50,740 --> 00:02:52,940 the number of study participants. 58 00:02:52,940 --> 00:02:56,240 Think about a total number of data points you wish to gather, and 59 00:02:56,240 --> 00:02:57,485 calculate backwards. 60 00:02:57,485 --> 00:03:00,780 Take into account that about a third of the answers you 61 00:03:00,780 --> 00:03:06,290 want will be lost due to inability to answer, duplicates, or garbage. 62 00:03:06,290 --> 00:03:07,640 Here's an example. 63 00:03:07,640 --> 00:03:11,080 Let's say you want 1000 valuable responses. 64 00:03:11,080 --> 00:03:13,150 Considering a third will be lost, 65 00:03:13,150 --> 00:03:17,820 bring that number up to 1,500 responses you wish to collect. 66 00:03:17,820 --> 00:03:21,530 If you ask the question five times a day for three days, and 67 00:03:21,530 --> 00:03:26,970 100 people participate in the study, you potentially get 1,500 responses. 68 00:03:26,970 --> 00:03:27,540 Get the idea? 69 00:03:29,420 --> 00:03:30,170 5. 70 00:03:30,170 --> 00:03:31,680 Choose a medium. 71 00:03:31,680 --> 00:03:34,570 You need to decide how you're going to send the question to 72 00:03:34,570 --> 00:03:36,450 your study participants. 73 00:03:36,450 --> 00:03:41,750 SMS, email, app, or voice messages are all good options. 74 00:03:41,750 --> 00:03:44,340 The best way to go is the simplest way for 75 00:03:44,340 --> 00:03:48,440 you while thinking about how data will be collected. 76 00:03:48,440 --> 00:03:52,000 For example, if you sent text messages to participants and 77 00:03:52,000 --> 00:03:55,250 they reply with answers over text messages as well. 78 00:03:55,250 --> 00:03:58,860 How are you going to organize all of that data into one place? 79 00:03:58,860 --> 00:04:05,680 If you plan on copy and pasting 1,500 text messages into a spreadsheet, think again. 80 00:04:05,680 --> 00:04:07,540 You need something simpler. 81 00:04:07,540 --> 00:04:11,970 I recommend using email through which you send a simple Google form. 82 00:04:11,970 --> 00:04:16,760 If you're savvy with more technical stuff, which I'm not by the way, consider using 83 00:04:16,760 --> 00:04:21,910 text messages and an IFTTT recipe that will drop them in a Google spreadsheet. 84 00:04:23,170 --> 00:04:24,220 6. 85 00:04:24,220 --> 00:04:25,910 Plan the analysis. 86 00:04:25,910 --> 00:04:30,700 The main activity you will perform during the analysis phase is classifying all of 87 00:04:30,700 --> 00:04:33,020 the answers to the question. 88 00:04:33,020 --> 00:04:37,640 Classification is no easy task, especially if you're on your own. 89 00:04:37,640 --> 00:04:40,956 If you work in a team, have everyone participate. 90 00:04:40,956 --> 00:04:43,940 To prepare for a group classification you need to 91 00:04:43,940 --> 00:04:48,160 predetermine the categories by which you'll classify the answers. 92 00:04:48,160 --> 00:04:52,580 These categories will probably change after you start classifying answers. 93 00:04:52,580 --> 00:04:55,630 Yet preparing them in advance will save you time later on. 94 00:04:56,820 --> 00:04:57,950 For example, for 95 00:04:57,950 --> 00:05:03,330 the note taking exercise I shared earlier, I pre-determined the following categories. 96 00:05:03,330 --> 00:05:05,865 List of things to buy, things to do, 97 00:05:05,865 --> 00:05:10,510 in-classroom, expressing thoughts, and sketching. 98 00:05:11,520 --> 00:05:16,470 By following these steps you can ensure your experience sampling research study 99 00:05:16,470 --> 00:05:19,770 will collect the right data you can confidently use later.