WEBVTT 00:00:00.000 --> 00:00:01.960 Welcome to the Deep Dive. We're here to help 00:00:01.960 --> 00:00:04.440 you quickly grasp complex topics. Making sense 00:00:04.440 --> 00:00:07.339 of the complex fast. Exactly. And today, we're 00:00:07.339 --> 00:00:10.660 tackling something fundamental in research. Data 00:00:10.660 --> 00:00:13.919 quality. It's a big one. We're drawing on an 00:00:13.919 --> 00:00:17.120 article from HackScience .Education by Gary Ackerman. 00:00:17.719 --> 00:00:19.739 Yeah, Ackerman's got experience in both quantitative 00:00:19.739 --> 00:00:22.769 and qualitative research, which is, well... really 00:00:22.769 --> 00:00:24.530 useful here. It gives him a good perspective, 00:00:24.809 --> 00:00:27.050 right? Definitely. His main point is that data 00:00:27.050 --> 00:00:30.210 quality isn't, you know, one single thing. It 00:00:30.210 --> 00:00:32.210 really shifts depending on the research approach. 00:00:32.570 --> 00:00:34.570 OK, so that's our mission for this deep dive, 00:00:34.829 --> 00:00:37.049 to unpack those differences. We want you to feel 00:00:37.049 --> 00:00:38.929 more confident when you come across research 00:00:38.929 --> 00:00:41.469 findings. So let's start with the two main approaches 00:00:41.469 --> 00:00:44.670 Ackerman talks about, quantitative and qualitative. 00:00:44.829 --> 00:00:48.130 Right, quantitative first. Yeah, quantitative 00:00:48.130 --> 00:00:51.460 research, basically. tries to reduce things down. 00:00:52.000 --> 00:00:54.200 Control variables, isolate them. See how one 00:00:54.200 --> 00:00:56.840 thing affects another. Think controlled experiments, 00:00:57.200 --> 00:00:59.700 numbers, statistics. Measuring specific things. 00:00:59.719 --> 00:01:02.320 Precisely. And qualitative. Ackerman describes 00:01:02.320 --> 00:01:04.280 it differently. Yeah, it's more about looking 00:01:04.280 --> 00:01:06.560 at the whole picture. Studying things in their 00:01:06.560 --> 00:01:08.640 natural setting, letting all the variables interact. 00:01:09.659 --> 00:01:13.219 The goal is often deep understanding, context, 00:01:13.799 --> 00:01:16.670 perspectives. Like, observing a group to understand 00:01:16.670 --> 00:01:19.349 how they work together, maybe. Less about counting, 00:01:19.450 --> 00:01:21.890 more about understanding why. You got it. And 00:01:21.890 --> 00:01:23.730 that's the key takeaway Ackerman highlights. 00:01:23.849 --> 00:01:25.989 Yeah. Because the goals and methods are so different. 00:01:26.129 --> 00:01:28.349 What counts as good data must also be different. 00:01:28.409 --> 00:01:30.310 Makes sense. It really does. You can't use the 00:01:30.310 --> 00:01:33.230 same checklist for, say, survey results and interview 00:01:33.230 --> 00:01:36.109 transcripts. The very nature of the data is different. 00:01:36.450 --> 00:01:39.709 But Ackerman is clear that all researchers care 00:01:39.709 --> 00:01:42.219 about quality, right? whatever their approach. 00:01:42.620 --> 00:01:45.079 Absolutely. It's crucial. It's about the trustworthiness, 00:01:46.019 --> 00:01:48.700 the reliability and validity of the findings. 00:01:48.799 --> 00:01:51.019 OK, let's pause there. Reliability, that's about 00:01:51.019 --> 00:01:53.420 consistency. Like, if someone repeated the study, 00:01:53.459 --> 00:01:56.040 would they get similar results? Generally, yes. 00:01:56.340 --> 00:01:59.620 And validity is more, are you actually measuring 00:01:59.620 --> 00:02:02.579 what you think you're measuring or describing 00:02:02.579 --> 00:02:06.280 what you intend to describe? Got it. So bad data 00:02:06.280 --> 00:02:09.379 means shaky conclusions, regardless of method. 00:02:09.780 --> 00:02:12.360 Exactly. And Ackerman mentions using Hopeful's 00:02:12.360 --> 00:02:15.400 framework from back in 1997, specifically for 00:02:15.400 --> 00:02:18.039 understanding quality in qualitative research. 00:02:18.340 --> 00:02:21.659 Hopeful. OK, so that suggests there are established 00:02:21.659 --> 00:02:24.780 ways to think about quality beyond just numbers. 00:02:24.960 --> 00:02:27.340 Right. It introduces different ideas, different 00:02:27.340 --> 00:02:31.300 criteria suited for, say, interview data or observational 00:02:31.300 --> 00:02:34.139 notes, things like credibility, transferability, 00:02:35.080 --> 00:02:37.639 concepts that fit qualitative work. Which underlines 00:02:37.639 --> 00:02:40.900 Ackerman's point. We need different terms different 00:02:40.900 --> 00:02:43.639 concepts even to talk about quality in these 00:02:43.639 --> 00:02:45.939 two worlds We really do for quantitative you 00:02:45.939 --> 00:02:48.460 might focus on you know sample size statistical 00:02:48.460 --> 00:02:50.639 significance measurement error standard stuff 00:02:50.639 --> 00:02:52.800 for stats yeah, but for qualitative you might 00:02:52.800 --> 00:02:56.139 look at the richness of the description or Do 00:02:56.139 --> 00:02:58.020 the researcher check their findings from multiple 00:02:58.020 --> 00:03:00.599 angles triangulation they call it? Ackerman also 00:03:00.599 --> 00:03:03.460 touches on things like subjectivity and objectivity, 00:03:03.699 --> 00:03:05.539 referencing Philip's work. Seems like there's 00:03:05.539 --> 00:03:07.520 a whole philosophical layer, too, especially 00:03:07.520 --> 00:03:09.879 in qualitative research. There is. How does the 00:03:09.879 --> 00:03:11.419 researcher's perspective play into it? It's a 00:03:11.419 --> 00:03:13.560 key consideration. So bringing this back to our 00:03:13.560 --> 00:03:16.939 listeners, understanding this basic difference... 00:03:16.939 --> 00:03:20.360 quantitative versus qualitative quality criteria 00:03:20.360 --> 00:03:23.719 is actually pretty powerful, isn't it? I think 00:03:23.719 --> 00:03:26.020 so. It gives you a lens. When you see information, 00:03:26.199 --> 00:03:28.319 you can ask, OK, what kind of information is 00:03:28.319 --> 00:03:30.340 this? Is it based on measurement or interpretation? 00:03:30.680 --> 00:03:32.819 And then what are the right questions to ask 00:03:32.819 --> 00:03:36.479 about its quality? Not just is it good, but good 00:03:36.479 --> 00:03:39.240 in what way? Precisely. So the core takeaway 00:03:39.240 --> 00:03:42.990 is Data quality isn't universal. It's judged 00:03:42.990 --> 00:03:44.830 differently depending on whether the approach 00:03:44.830 --> 00:03:46.930 is quantitative, focused on numbers and control, 00:03:47.189 --> 00:03:50.229 or qualitative, focused on context and understanding. 00:03:50.729 --> 00:03:52.830 Right. Which leads to a final thought for everyone 00:03:52.830 --> 00:03:55.669 listening. Think about how you judge information 00:03:55.669 --> 00:03:59.250 in your daily life. When you see, say, a news 00:03:59.250 --> 00:04:01.789 report full of statistics versus hearing a personal 00:04:01.789 --> 00:04:05.810 story from a friend, do you maybe subconsciously 00:04:05.810 --> 00:04:07.930 use different standards to decide if you trust 00:04:07.930 --> 00:04:11.379 it, if it feels quality? That's a really interesting 00:04:11.379 --> 00:04:13.919 question to ponder. How the type of data shapes 00:04:13.919 --> 00:04:16.060 our own sense of its quality. Something to think 00:04:16.060 --> 00:04:17.819 about. Definitely. Well, thanks for exploring 00:04:17.819 --> 00:04:19.000 this with us today. Thank you.