 Okay, final speaker for this segment of the afternoon is Ian Yin with understanding public uses in funding of science and Just to know after this there's a break. We're running a little late But I think the break will try to make it a little shorter and if you're coming back Yeah, be sure to be back on time for a moment this is not Thank you. It's great pleasure to be here. So my name is Ian Yin and I'm currently incoming assistant professor at Cornell University and one of the recurring topics, you know during the last two days of discussion have been thinking about How production in scientific knowledge will also affect other important on social economic domains across General society and this has been a very canonical example thinking about the potential impact of science outside science Remain your geometry, which is you know for a abstract mathematical theory developed in the 19th century, but it later proved essential to high intense development of General relativity theory later also essential to GPS a key technological Components that made right sharing services such as Uber and Lyft possible So this is just one example, but it illustrates how science even vocational basic science can have a far reaching impact on our daily lives Even after two centuries right but this at the same time it is only one of the many takes people may have Regarding the science society interface and there are indeed the many forms of sketch them about the public usefulness of science One of them says that science is often ivory tower activity So what this means is that scientists that have particular interests that are a song note towards the general public really cares about and Even if when there are good alignments between, you know, public interest and scientific interest that with this Exponential growth of research, you know research is getting increasingly more Specialized the public may have little capacity to distinguish the high quality sciences from the low quality science Which is another layer of concern about the youth's public usefulness of science and beyond that There has also been discussions and the views about whether scientific funding But especially public scientific funding which views from taxpayers money really has a lot of relationship to its public use So it is therefore not surprising that if you look at decades of literature across all kinds of fields in social sciences As well as natural sciences and engineering people have been talking about this idea about how exactly do science stacked with you know other parts of the society from very different perspectives and What we are trying to you know facilitate this discussion. This is a paper We recently published on nature human behavior is trying to use Computational social science methods to build a large-scale data pipelines that help us empirical. They understand these relationships So what we did is to start from 200 million papers and use new data linkages to trace cell downstream applications how they are cited in government policy documents how they are mentioned in media news and how they are cited in technological patterns We also brought in our fifth data set which is dimensions including more than five million research projects and the resulting Publications funded by over four hundred funding agencies worldwide so that you not only know the downstream applications of these papers, but also upstream funding of these papers and this allows us to Empirically look at a framework that looks at the relationship between scientific production is public use as well as its public funding And here we discover three man sets of results first we look at the quantities of use So here with design measurement, which we call relative consumption index RCI Which looks as a relative Extended to which a specific research field is used by a specific public domain as it's normalized so the baseline rate is always one and we can look at you know For each scientific field to what extent it's used in government policy media news as well as patents For example economics it is very heavily used in policy as well as media news But not so much in patents and this is a very different pattern If you look at on computer science as well as many other stem related fields You see tremendous use in patents but they currently appear to be underrepresented in terms of direct uses In policy and the news another interesting example here is biology This is across all top level fields We studied biology is the only example that you see over representation across all three domains policy news and the patents and Extending these analysis to all top fields in our analysis What we are seeing is a large set a very diverse set of specialized relationships between specific domains of Public uses as well as specific research fields. So collectively, this is a very diverse picture of how different scientific fields are getting used by different public domains and Beyond the quantity of use themselves our second set of results Looked as a long-standing concern that the public may draw on poorly established scientific ideas But what we are seeing here is not the case So what we did here is to calculate a measurement We call hit rate this is a best on scientific impact normalized by the same Publication year as well as scientific field. So a measurement of scientific impact best line rate is 1% Looking at the subset of research that gets used by government news as well as patents We see each of these domains the subset of research that gets used are associated with the high hit rates That are about 10 small times 10 times higher than the best line rate 1% and the results gets even more Substantial if you look at the intersect of these public uses for example If you look at the paper that is used by all three domains Then this paper is associated with a hit rate that is about 80 times higher than the best line rate And this appears to be a result that is not only common across different public domains of use But it also appears to be for a nearly universal across all research fields Regardless of whether which specific research field which specifics like public use domain you look at the Subsets is also as always uncalculated by a hit rate that is higher than the 1% best line rate So this is a second set of results for a universally high impact in terms of public use And our last set of results looks at the role of public funding how it further relates to the public use and to do this We expand from the 19 top-level fields to about 300 subfields and for each subfield We look at a measurement which we call average Founding US dollar per paper. So this is a measurement We use to approximate the magnitude of public investment in each scientific subfield And what we are seeing here is that on the y-axis there's a huge heterogeneity in terms of public investment If you look it from subfield to subfield you find that very a lot You know going across more than five orders of magnitude yet for each of the RCI measurements There are relative uses in government in use in patterns We see positive relationships which means that which is funded by the public as for a consistent was which is used by the public And what's even more interesting here is that if we simply combine the three access three RCI Measurements here using a very simple linear regression model We immediately see very high predictive power a very high degree of agreement with public funding So what this means is that each public domain provides independent and complementary predictive power for allocation of public investment in science Although each research field differs significantly in their relative role and the contribution with in science and beyond science Correctively what we are seeing here is as a combination of their impact beyond science sharply predicts the funding Which suggests that ultimately what's a public uses was a public funds as well as what is used by scientific Communities themselves are remarkably consistent with each other While these different elements may be you know considered to be very remote from each other in some of the theoretical frameworks What we are seeing in current picture is a rather optimistic news They may appear closer than we a lot of have has imagined and What's very exciting for me about this project is that it provides a data pipeline and the measurement Framework said cannot be only used to study this specific domain We can think about how do we do case studies or how do we even generalize this framework and this is just a one example of it Another paper we also published recently thinking about a case study, which I also from my opinion This is more about you know a stress of our optimistic view about how society and science may be related Maybe related very closely. So what we are looking at is let's take the clock back to three years ago So this is a very first a couple of months is of the COVID-19 global pandemic And why do I see this as a stress test because by that time we see very Dynamical very uncertain frontiers on both the policy side as well as a science side on the one hand side What you see is very dynamical uncertain yet extraordinary consequential policy environment across the globe But as a same time our scientific understanding of the disease about the fact things about all kinds of Effectiveness of all kinds of public health measurements has also been evolving a lot So this is really the time that we have for a uncertain policy and scientific understandings and which raises a lot of question for example is our policy understanding of COVID-19 closely linked with evolving scientific Understanding or largely separated from it does policy development engage high-quality science and so on so forth And I don't think I will have a time to cover a lot of the details here Feel free to check out the paper if you are interested in more methods But overall what we are finding here is again rather I will say optimistic view What I'm looking at is that if you focus on the citation relationship between COVID-19 related the policy documents and COVID-19 related scientific papers we see they are they are much more Likely to start extremely recent science These are not scientific papers published ten years ago But as they are more likely to start scientific papers published after the pandemic starts So also my likely to engage with scientific papers So scientists themselves find important if you look at the scientific citations you see a remarkable correlation with the policy citations and the further we look at the Publication value of these papers that are used by policy makers What we are seeing here is that COVID-19 policy are also more likely to start a peer-reviewed journal articles Rather than pre print manuscripts although the later is playing an increasingly important role in pandemic related science Before I jump into the final take home message Let me be very clear about some of the limitations in our current studies So data says we collected here only represents a subset of domains that science may impact within each domain There could also be other channels where science and is used and we see this as a great opportunity for future research Also, we are largely relying on citation based on matrix Although we have tried our best to do all kinds of normalization and the controls As we haven't done a lot of systematic semantics analysis, which we also see a Major shortcoming of the current analysis and the further we are mostly presenting Correlational evidences there could be multiple maximums at work And this also cause for future research in terms of a causal maximums why science and the society Appeal to be very closely related to each other Regardless of these shortcomings our results presents a further set of larger scale empirical evidence Showing what is funded by the public what is used by the public and what Used by scientific themselves appear to be remarkably consistent with each other and this kind of alignment Appeal to be there in both regular and emergency situations, and that's all I want to bring to you today. Thank you Excellent talk. I recall from a paper that I read recently that papers that are mentioned on Wikipedia are far more likely to be cited in Court cases same thing with briefs from court cases and whatnot. Did you look at the interdependence of these? Different sources and what did you learn from that? Yes, thank you. This is a great question actually I think what's interesting here if you think about the public uses of science They take different roles policy patterns and as you imagine on social media We keep pdr cot citations, which is a little bit closer to policy citations What's interesting here is that they operate on very different timescales, right? If you think about social media mainstream news most of the coverage happened within a few weeks Like you don't tweet papers published 50 years ago a lot But at the same time if you look at the policy cross citations or even patent citations Most of these citations operate on a timescale of years It is not uncommon to see policy documents or court decisions citing papers published one or two decades ago And I think this creates a lot of interesting questions One is that how different are they what and at the same time is that whether there are some commonalities What are they are in the you know common quantitative frameworks that tracks a temporal revolution of how people use science This is pretty much the kind of things we are working on and I do see that if you think about different layers of public uses There may be interactions if your paper was previously You know have huge influences on social media on Wikipedia on mainstream media news There is definitely a possibility that this is where also boosts their future impact in classrooms You know in the White House. This is pretty much a kind of work We are working on as a follow-up, but yeah, this is definitely important thing to think about in this area Thank you