Read on Twitter
I just spent three amazing days with @tfblanken and @CharlotteCTanis in the @kromhouthal. We measured social distancing behavior during the Art Fair organized by #smartdistancelab. This is how the biggest behavioral experiment I've ever done came to be.
In May, the Science versus Corona team ( @tfblanken, @CharlotteCTanis, @fdabl and I) came up with the idea of measuring contact networks through wearable electronics to figure out which social distancing interventions work best: https://psyarxiv.com/53ey9/ .
The central idea is that we can use sensors to measure distances between people. We can think of these distances as a network. In that network, we represent each individual as a node, and we link two nodes when people are within a distance of 1.5 meters of each other.
We can analyze the network to assess how risky it is with respect to virus transmission. The simplest way is to count the number of links between people: the more links, the riskier. We can also use more advanced measures. Piet Van Mieghem ( @tudelft) helped us develop these.
So suppose we have two experimental conditions: say, one in which people wear a device that buzzes if they get too close, one in which they don’t. We compare the networks in both conditions to determine which is safer. We call this a BEhavioral COntact Network (BECON) indicator.
By using BECON indicators, we could for instance compare different floor plans, walking directions, nudges, signs, etc. In other words: BECON indicators provide a methodology to measure the success of behavioral interventions. They connect behavioral science to epidemiology.
Contact measurement is unreliable, however. We used simulations to assess whether BECONS would work in situations where one has many false positives and false negatives in the measured network. This was promising: BECONS proved robust to noise (see https://psyarxiv.com/53ey9/ )
In June, after writing up the paper, we started looking for technology that could make this work. @harmerbe of @CapgeminiNL guided us through possible technological solutions.
In July, more or less by accident, we found out about the people from #smartdistancelab who were organizing an Art Fair in the @kromhouthal to test how to safely organize large public events. We got in touch; they were enthusiastic and wanted to try our BECONs.
Meier Boersma of the #smartdistancelab organization then discovered Focus Technologies https://www.findfocus.nl/ . It turned out that they actually had developed a sensor to measure distances between people.
The Social Distancing Sensor ( https://socialdistancingsensor.nl/ ) developed by Focus is produced by the company Sentech ( @Sentech, https://www.sentech.nl/en/ ) where you can actually buy this device. People at Sentech were enthusiastic as well, and provided us with 250 sensors for free. Thanks!
In August, with @goafm and Sanne de Wit, we developed an experimental design to compare a baseline condition (no intervention) to conditions in which a) people wear face masks, b) there are walking directions on the floor, c) the sensor buzzes if people get too close.
@goafm and Sanne de Wit also contributed a questionnaire to assess people's attitudes towards social distancing rules and how much they worry about the virus. We administered this to all visitors of the Art Fair when they ordered their tickets. There's also an exit questionnaire.
And now we actually did it! Over the past three days, we measured hundreds of visitors in the @kromhouthal visiting the #smartdistancelab Art Fair and collected ~200.000 data points. It was an amazing experience.
From an idea in May to a paper in June to technology in July to a large scale behavioral experiment in August... I don't think I have ever worked so fast in my life. @tfblanken and @CharlotteCTanis what a ride!
Here’s some Dutch media coverage: https://nos.nl/l/2346053 , , https://www.volkskrant.nl/cultuur-media/bezoeker-wordt-proefkonijn-smart-distance-lab-test-het-gedrag-van-kijkers-op-the-art-fair~ba7f0860/?utm_source=link&utm_medium=app&utm_campaign=shared%20content&utm_content=free
What comes next? With Science versus Corona, we have formed a Team Data to organize the (anonymized) data and to make them openly available through @BrianNosek's @OSFramework. We also plan a paper in @JelteWicherts' Journal of Open Psychology Data ( @up_jopd).
And of course, after we analyze the data, the substantive results will be reported in a paper as well. (This may take a while, because we don't just have to statistically analyze the data, but also have to invent some of the statistical analyses themselves.)
I want to thank everyone who helped us out: Floor Nauta, @SandraJGeiger, Melle van der Linde, Nadza Dzinalija, @ninaleach13, Marieke Wagenaar, Elenika Schalm Boiko, Trijntje van Doorn, @HenkNieweg, Meier Boersma, @mayaVirg, Mette Samkalden, @helmavandenberg…
…Yuri Jansen, Nathan Wiersma, @svpsfeerbeheer, @QEastbird, Rick Bouten, Hendrik-Jan Dreuning, @harmerbe, Sentech, Focus Technologies all the people at the @kromhouthal, artists, visitors, and everyone I forgot!
Many people at the @UvA_Amsterdam helped us as well - even if just putting up with me missing appointments and forgetting half my to do list. @dora_matzke thanks for having my back, and @EJWagenmakers thanks for the @JASPStats t-shirts ;-)
The Science versus Corona project is supported by an NWO fast track grant ( @NWONieuws), and the #smartdistanclab is supported by @MinisterieEZK.
The research project was executed in keeping with the ethical regulations of the @FMG_UvA and evaluated by the Ethics Committee of the Psychology Department. All participants filled in an informed consent form, participation was voluntary, and all data will be fully anonymized.
Curious about the scientific results of the research project? Follow Science versus Corona @versuscorona to stay tuned!
