Nigel Cross, Essien, Fifteen Years discussed on Not So Standard Deviations

Automatic TRANSCRIPT

That's all I don't have to do these long lecture like writing essays. And you know, I do remember my first sociology class. I just have this sense of like, wait a minute. We just all get the room and talk. But like I want the answers like. That was so foreign to me and like now, I'm totally cool with it. But at the time, I was just like, no that's not how learning works like so. Yeah. I think you're you're high pop asus. I well, I don't know how it's data. There is a two to examine it. But I think I'm with you on that. I think I just maybe to frame it another way like. Dated houses processes like way, more high dimensional than we make it out to be. I think I think I think the here's the formula for how to do things it's like trying to project it onto a single dimension. And like let's throw people in the river is like its individuals. It's like. And I think it's funny because I see that kind of swig back and forth a couple of times. Now over the last like say fifteen years in terms of like how we teach did analysis, and it just it feels to be that like. Neither can be correct. Right. It's impossible either. What to be corrected? So it's, but that's it. That's all we've got basically. Yeah. Yeah. Well, I mean, that's one thing. I don't think he's covered it in this book. But Nigel cross talks about in. I think it's a paper. In fact, we should share the paper. For this episode. But he talks about how design went through like their version of null hypothesis testing where it was like, okay. Like, I think they call it scientific design or something where it was much more. Like, this is what you do now and not like the design sprint thing. Like, I think design's prince leave a lot of time for creativity. It was more. Like now, your brain is doing this prior. And now your brain is going to do this part. And like it happened in the sixties either leave or maybe the fifties. But I always wondered if that was at the same time that hypothesis testing was being developed as well like if that was just like a trend. Like, the rational scientific something or other theory. It's like a rational scientific approaches, and that's what it's called or something. I don't know like for for hypothesis Essien go for like for like the design for design. I think I think there was like a thing called rational scientific. Yeah. I think you're right. It was like, well, I thought of scientific design, but maybe rational is like on there too. Yeah. It was. But it was like a very like they were imitating the scientific method. Right. Yeah. But that like became I think in the sixties late sixties and seventies. That was like not how people were approaching the world. And so it fell apart for good reason. You know, so I think so here's what I think is happening in the data world is that like between then. And like now there was a swing somewhere in the middle to like, let's do whatever we want. Right. And I think we now we're swinging rear right now swinging back because for a variety of reasons, I think, and I think there is definitely a push towards like automation in this area. And which would which is like, a very complicated cookbook, basically. And because it will. So for example, in academia, this is whole talk about the replication crisis, and whatever, and I think there's one of the I think one people one of the root causes that people believe is that like people aren't trained to do. The correct date analysis. Right. And so if you automate all that, then you can cut it remove all the all the humans, humans are all the problem. Right. So so I think we're I feel like we're swinging back in the other direction now and away from the kind of like more exploratory approach. I guess. Yeah. I could see that..

Coming up next