Contact FutureLearn for Support
Skip main navigation
We use cookies to give you a better experience, if that’s ok you can close this message and carry on browsing. For more info read our cookies policy.
We use cookies to give you a better experience. Carry on browsing if you're happy with this, or read our cookies policy for more information.

Skip to 0 minutes and 1 secondSo, you know, the big question here was, how should we present our data? How should we present information? There are transparent ways of presenting data and there are ways that obscure the data. There are ways that sort of hide what's really going on in the data. We’re going to try to talk about transparent ways of presenting the data. Tufte is a professor of political science at Yale. He used to be at Princeton. Published, you know, empirical work in political science for a while. But he was always really good at and really interested in graphical depiction of data. That eventually, that sort of became his day job.

Skip to 0 minutes and 40 secondsAnd actually, what’s neat about Tufte’s book is it’s much more than just a book about graphic design. There are elements of graphic design, like what looks good. Some of that’s in there, for sure. But, there’s also sophisticated discussion of causality and research design in various points in the book. There’s also a big transparency thrust throughout. And we’ll see that in his principles and his bullet points about accurate ways of conveying information and data. That is a thread that is found throughout the book. These are a set of issues that I see as related to transparency points. So, among his 8 or 9 principles, he says, “show the data.” That is a pretty simple but important thing to do.

Skip to 1 minute and 24 secondsLike, when we see empirical papers or statistical analysis where we don’t really see the raw data in any form, we don’t see a scatter plot, we don’t see any relational graph between variables. You know, I think Tufte would say, “Maybe our antennae should perk up a bit.” I wanna see the data. I wanna see patterns in the data. And I think the best empirical work in the social sciences and in other fields is characterized by a certain openness about what the data looks like and what it is. So showing the data is important. Avoid distorting the data. A figure, a graphic, should serve a reasonably clear purpose.

Skip to 2 minutes and 2 secondsHe lists out, it might descriptive, it might be trying to capture a certain causal relationship with a certain research design. It may have some other goal. But when you look at it, it should be clear. With the figure description, it should be clear what the goal of the figure is.

Skip to 2 minutes and 19 secondsAnd then four, it should induce the reader to think about the substance. So there should be more than just some sort of statistical exercise involved. There should be a link to something important, something that matters. Something substantive. And this is one of the points that comes through the book again and again. That effective graphics are much more than just art and much more than putting statistics in an image. There has to be a real depth of intellectual understanding to create a good graphic. So, that’s something that comes through. These are really transparency related points that I think are good general principles. Then there’s a second set of issues that, you know, I guess speak to other issues.

Skip to 3 minutes and 5 secondsOne, the value of graphics. And we’ll show an example of this in a second, is they allow you to absorb a lot of data. They make large data sets coherent. You can present a lot in a small space. You can present a lot more data in less space. Good graphics, or figures, also reveal patterns in the data. Potentially at different levels of detail. Sometimes you get a different geographic aggregation. Sometimes different time scales are revealed all within one graphic. They encourage the eye to compare different pieces of data. Of course, that’s natural and much easier to do sometimes in a figure than in a table.

Skip to 3 minutes and 49 secondsThis is the last one, and we’ll come to this at the very end of the lecture today. Good graphics are able to integrate a lot of different elements. Good graphics integrate words, verbal descriptions, they integrate statistical concepts and statistical coefficient estimates. And then there are certain visual patterns they bring out. So, some of the most effective graphics are, in some sense, quite complicated. But the art, of course, in this is conveying that complicated information in a way that’s comprehensible. So, these are just some ideas. We’re going to go through these, not exactly one at a time. But we’re gonna touch on these in the course of the lecture with specific examples.

Skip to 4 minutes and 30 secondsAnd I think the hope of this is, a) to makes us all think about our own research and our own figures. And incorporate some of these ideas into our work. Both in terms of making our figures, and our tables too, more effective at doing what they’re supposed to do. And more transparent.

Introduction to transparent data visualization

Another way social scientists can improve transparency in their research is to use good visual representations of their data. This can be especially useful when presenting large datasets or when trying to illustrate patterns that are more easily revealed in visualizations than in tables. Statistician and political scientist Dr. Edward Tufte is well known for his book The Visual Display of Quantitative Information, which Dr. Miguel discusses in detail over the next five videos. This first video introduces the concept of data visualization, as well as some of the principles Dr. Tufte puts forth as integral to figures’ and graphics’ transparency and effectiveness in research. He also considers ways in which data visualization can enhance readers’ understanding of data. As you go through this Activity, think about how you might incorporate these ideas and principles into your own work.

Share this video:

This video is from the free online course:

Transparent and Open Social Science Research

University of California, Berkeley