Managing Hype in Statistics

[00:14:49] Alexander: Yeah, I think the, when I talk about influencing others, yeah, don’t just rely on the facts. Speak about [00:15:00] what’s in it for them, drive emotions, yeah, that is so important. And last but not least, one really critical thing is people need to trust in you and your organization. Yeah. I, when all this data science hype came [00:15:20] up a couple of years ago I was seeing Steven Rupert talk, talking about this and he was saying, if you don’t buy the time, this hype comes, have a good reputation within your company.

[00:15:39] [00:15:40] If you’re not regarded as a trustworthy partner, people will not believe you that, you know, this external things are overhired and things like that. So you need to build your trust, your reputation. Before the hype comes, when the hype is coming. And you [00:16:00] know, the question is not whether there will be another hype.

[00:16:03] The question is when there will be another hype. Yeah. You need to build a trust in your organization, in your individual people before the hype comes. And that by the way, will also help you to [00:16:20] implement all these, let’s say lower hanging fruits. like a futility analysis much more consistently. So think about the branding of your rep of your biostats department of your data science department, whatever you kind of call it, where you’re [00:16:40] working at.

[00:16:41] So reputation is absolutely critical. 

[00:16:44] Kaspar: It’s a, it’s a very good point you bring up. And it reminds me that in, in June, there is the FSPI stats leaders meeting, which happens every year. This year it happens in Switzerland and the FSPI statistical methodology leaders have been invited to run a session there [00:17:00] and we discuss innovation, commercialization of innovations in organizations and before the event, we sent out a questionnaire.

[00:17:08] And one of the questions was if some fancy. Startup is approaching your chief medical officer with some new quantitative method. They promised to cut, I don’t know, costs [00:17:20] by half or whatever these promises typically are. Who is your CMO approaching? Is it kind of the data science head in your organization?

[00:17:27] Or is it, is he or she just believing that startup and, and pushes this into the organization? And I think that exactly, relates to the point you’re making. Is there trust in the internal data [00:17:40] science and statistical capabilities that they will provide an honest and rigorous assessment of the method and that they’re capable of doing that?

[00:17:49] Or is there no such trust? And then as a data science organization, you’re left out. And that’s not where we want to be as statisticians and data science organization. So yeah. And you have to [00:18:00] build this trust in say, In quiet times, when, when there is no such things coming your way and then that you have it once the things come your way and and yeah, and can make this rigorous and honest assessment for the organization.

[00:18:14] Alexander: Yeah, completely agree. So a couple of things, what are the things that [00:18:20] you think you can do to improve set reputation beforehand? So you don’t get into this challenge.

[00:18:28] Kaspar: I mean, act. Consistently as a rigorous voice, which is not just saying no to everything because it’s [00:18:40] maybe complicated or you need to look into new things. Always try to make an honest assessment, provide your expertise in such a way. And then build that reputation over time. But I think we should not underestimate it takes time and you need, I mean, that’s what you said.

[00:18:59] You need to [00:19:00] take the other side seriously. They’re also, they have, they are under pressure. They need to justify. Why they’re not doing things, why they are doing things, you can help them make, guide that decision. Yeah, these are all kind of lofty, lofty comments. But yeah, I think, I mean, how can you, [00:19:20] it boils down to the question, how can you build that reputation?

[00:19:23] And I think it’s by consistently being honest, By consistently being rigorous. Also, sometimes I feel, and that’s maybe sometimes a bit unfortunate, but you’re listened [00:19:40] to in your own organization, sometimes more when you come from externally, when people see you comment on a panel with FDA folks, for example try to get in there.

[00:19:51] So that’s kind of this outreach piece, this collaboration piece as well. If you work together across companies and then six, seven companies. [00:20:00] write a white paper, maybe that carries more weight than when you internally try to push for something or build that reputation. So those all goes together. And I think there is no standard recipe.

[00:20:13] And it, it, I think it takes time and, and effort to build. That [00:20:20] credibility and then even if you have that credibility, depending on who’s on the other side, how much they trust statisticians, data science, it’s still no guarantee you are the go to person for them, but I don’t think there’s much more you can do.

[00:20:34] And Steve, of course, is a is a template example. For that and and [00:20:40] for his influence but there are other such figures, but it’s not like they show up one day and then the next week they have this role. Yes. I think that builds over years and years through multiple activities, being involved in the scientific discourse, publications Attending [00:21:00] conferences, speaking at conferences, speaking in panel discussions, doing webinars recording effective statistician podcasts, you name it.

[00:21:08] All these activities help to build a certain reputation first in your field and then hopefully at some point. Beyond your field, but yeah, I, I, [00:21:20] I don’t think there is kind of a secret recipe or secret sauce here. It’s just hard work. 

[00:21:25] Alexander: Well, there’s a couple of things that you just need to know. Yeah. So it is a lot about building trust, networking, linear reputation, and definitely having an [00:21:40] external network and an external reputation that you can.

[00:21:43] Pull from will hugely increase your internal trust. Another thing is also that when you also spearhead these innovations internally. Yeah, so [00:22:00] never think about, oh, that’s not, that’s not cool or we’re skeptic about it. Yeah, leverage things internally, educate internally. You need to be on the top of these things.

[00:22:14] Yeah. So is that you have already [00:22:20] communicated first within the organization. So that’s the big advantage in being the first mover here in terms of communication, because if you already set the frame, if you already have the story here, then people will be much less likely and susceptible to. [00:22:40] external kind of stories and external hype.

[00:22:43] However, if you kind of just stay always to kind of we have all always done it this way, then of course you set yourself up for being made redundant, being kind of put into the that’s, you know, statisticians are just those that [00:23:00] do the tables for the clinical trials niche. So I think being, you know, innovative Always kind of on the lookout for how you can improve things, how you can, you know, leverage these waves rather than by being hit by them is very, very important.

[00:23:19] Kaspar: I agree to [00:23:20] that. So this kind of being proactive instead of reactive is very, very important. I think what one should not underestimate is that an organization needs to provide a at least to a certain extent, resources for that, or, or kind of incentivize and encourage people to, to stay abreast. And if it’s, if it’s just [00:23:40] with the statistical literature to see maybe what’s coming and then that you can make an assessment.

[00:23:47] One of my favorite examples is estimates for a long time there, there were lots of discussions happening and there was no requirement from health authorities yet. But if you started [00:24:00] early on to bring your organization into this direction and to, to start to generate at least understanding, if not excitement for that, this concept actually will help us doing drug development.

[00:24:15] Then you might have. eventually, when then at some point regulators [00:24:20] started to request it. And there are other such examples. And I think organizations also need to be robust enough because sometimes you invest time into topics that do not become important. I think, I mean, the more senior people you have, the more they have a nose for what’s maybe will be important [00:24:40] moving forward.

[00:24:40] But sometimes you can also be wrong. And then you’ve invested time into something that turned out not to be so relevant, but you never really know you have no certainty. And but if you want to stay ahead of the curve, you need to allow for at least a certain amount of time. [00:25:00] Some people or some people with some time to, to kind of try to identify what will be important moving forward so that you can be proactive when then something is coming from outside and you already can very quickly develop an opinion on it, or already have that opinion and have the expertise.[00:25:20] 

[00:25:20] Alexander: Yep. Completely agree. If all of your people are already kind of 100 percent booked with all the standard stuff then it’s pretty hard to be innovative and thinking forward. You need to invest time so that you can get time out later on. [00:25:40] And this time investment is really worth it. And when you do these things, I think one important thing is that you have some kind of forum, platform, communication channel within your organization, via which you can shape the organization, via which [00:26:00] you can show how innovative you are, how you can bring in external experts, how you can, you know, collaborate with others.

[00:26:10] That is really, really fundamental to have. Because this kind of channel is what, how you build your reputation. [00:26:20] Yeah. Of course, all the one to ones your stat statisticians, data scientists have is really important. And I would never underestimate these kinds of things. And you also need to have some kind of internal communication channels so that you can, you know, do all these kinds of things.

[00:26:39] Kaspar: Yes. [00:26:40] And I mean, I can mention Roche as an example, we have internal seminars, of course, where people can present and also try to generate an atmosphere that you don’t need to have like conference style presentations where everything is perfect. You can just come and maybe also say, well, I got until here and then I got stuck.

[00:26:57] And what do people think? And, and [00:27:00] we’re not always successful in this, but yeah, also kind of have these forum where you can just discuss and have a scientific discussion about things and, and, and seek input from your colleagues so that you can learn together. I think that is a very. Important piece and generate the confidence in people that they can also learn new things [00:27:20] and that they’re not overwhelmed by something coming from outside and understand, well, it’s maybe a little bit outside of what I know, but maybe it’s not so far outside.

[00:27:29] And maybe a few of us together can figure out what this thing is really doing. And then you can stay on top of the [00:27:40] developments rather efficiently. 

[00:27:41] Alexander: Yes, you definitely need these. Statistics, data science, internal things, and you also need the external facing seminars and trainings and things like that.

[00:27:54] Presentations where you as a biostatistics [00:28:00] organization present your expertise and your collaborations to others. Yeah. So when you see there’s a new hype. around X coming in, yeah? Invite someone that is an expert in X and have a panel discussion [00:28:20] internally or something like this, yeah? That will give you a lot of visibility and a lot of credibility and that will prevent that people, or at least minimize risks that people will do without you, yeah?

[00:28:37] That you then need to kind [00:28:40] of correct and Well, reset expectations, what always usually is lowering expectations and nobody wants to hear that. So be proactive in these kind of things. It is absolutely fundamental.

[00:28:58] Thanks so much Kaspar. [00:29:00] That was an awesome discussion about managing hypes. And we have mentioned a couple of these at the beginning. I don’t know which ones will be the next I just know that maybe already next year on two years, we’ll have the next big things that is coming around. So [00:29:20] really invest in your reputation internally, in your trust, in your network, in your visibility, all these things.

[00:29:28] will help you a lot. And as you said, also commercialize the things that are already existing. That will also gain you a lot of reputation. [00:29:40] Any last thought on this topic? 

[00:29:44] Kaspar: No. Thanks for, for having me and give me the opportunity to share my thoughts and yeah, and encouragement or just kind of wrapping up.

[00:29:52] I think two points for me are important. Stay abreast of what’s going on, build that reputation, as you mentioned, and And don’t [00:30:00] forget to do the simple things right. And sometimes we tend to miss the simple things just because we are too busy with trying to do the fancy things. But the simple things like utilities, they can carry a lot of value and also focus on those.

[00:30:16] Alexander: Thanks so much.