Oxford CompSoc banner graphic
1st Week

MT3 Newsletter: A Fascinating Talk on Vaccine Survey Stats, Learn to Code Session 2 & Board Game Bash!

Hi everyone,

It's been a weird week news-wise right!? Well, despite all the political chaos right now, there'll always be one thing you can be sure of... cool CompSoc events. Great to see lots of you last week - stay tuned for our first academic Thursday Talk and more!

- the CompSoc committee

Still not a member? Membership only costs £1 for LIFE, with tons of benefits - discounts on merch and tickets, free food at all socials, our exclusive Discord server and more - so it's definitely worth signing up!

Be a Member

Thursday Talk: Unrepresentative Big Surveys Significantly Overestimate US Vaccine Uptake, by Seth Flaxman

Lecture theatre picture

(Bradley et al, Nature 2021, https://www.nature.com/articles/s41586-021-04198-4)


Surveys are a crucial tool for understanding public opinion and behaviour, and their accuracy depends on maintaining statistical representativeness of their target populations by minimizing biases from all sources. Increasing data size shrinks confidence intervals but magnifies the effect of survey bias: an instance of the Big Data Paradox. Here we demonstrate this paradox in estimates of first-dose COVID-19 vaccine uptake in US adults from 9 January to 19 May 2021 from two large surveys: Delphi–Facebook (about 250,000 responses per week) and Census Household Pulse4 (about 75,000 every two weeks). In May 2021, Delphi–Facebook overestimated uptake by 17 percentage points (14–20 percentage points with 5% benchmark imprecision) and Census Household Pulse by 14 (11–17 percentage points with 5% benchmark imprecision), compared to a retroactively updated benchmark the Centres for Disease Control and Prevention published on 26 May 2021. Moreover, their large sample sizes led to miniscule margins of error on the incorrect estimates. By contrast, an Axios–Ipsos online panel with about 1,000 responses per week following survey research best practices provided reliable estimates and uncertainty quantification. We decompose observed error using a recent analytic framework to explain the inaccuracy in the three surveys. We then analyse the implications for vaccine hesitancy and willingness. We show how a survey of 250,000 respondents can produce an estimate of the population mean that is no more accurate than an estimate from a simple random sample of size 10. Our central message is that data quality matters more than data quantity, and that compensating the former with the latter is a mathematically provable losing proposition.

About our speaker:

I am an associate professor at the University of Oxford in the Department of Computer Science and a Tutorial Fellow in Jesus College. My research is on scalable methods and flexible models for spatiotemporal statistics and Bayesian machine learning, applied to public policy and social science. Active application areas include public and global health and machine learning for science. I co-founded the Machine Learning & Global Health Network (MLGH.net) and I help run the WHO-associated "Global Reference Group on Children Affected by COVID-19." My research is currently supported by an EPSRC Fellowship (2020-2025), “Spatiotemporal Statistical Machine Learning (ST-SML): Theory, Methods, and Applications.”

TL;DR: Big surveys are often far from reliable, especially when related to Covid-19 - come to the talk if you're interested in vaccines, statistics or big data and want to find out more!

When: 5-6PM, Thursday 27th October

Where: Lecture Theatre B, Department of Computer Science 7 Parks Rd, Oxford OX1 3QG

Click 'Going' here: https://fb.me/e/2H2J9YxXh

Learn to Code Session 2 (Bring your laptop!)

Code picture

Join us for the second session of our hugely popular Learn to Code series, aimed at complete beginners! Our amazing coding officers will guide you through learning Python, from the absolute basics to making your very own game by the end of the course. If you've ever wanted to pick up coding, but felt it was hard, frustrating, or just didn't know where to start, this is the perfect event series for you. Non-members and members alike are free to come along! Bring your laptop!!

Course Materials: https://github.com/oxcompsoc/learntocode

When: 6-7PM, Saturday 29th October

Where: Lecture Theatre B, Department of Computer Science 7 Parks Rd, Oxford OX1 3QG

Click 'Going' here: https://fb.me/e/2lfwCxOla

Saturday Social: Board Game Bash!

Gaming picture

What kind of board gamer are you - a massive strategy nerd, with the perfect winning algorithm falling into place in your mind? A mental minmaxer, tallying up the probabilities and making the wildest leaps in Codenames? Or just prefer to relax with a nice game of snakes and ladders (which technically has 0 players, but it's all about the drama right)? However you choose to play, you'll definitely have fun at our mega board game bash this Saturday! We'll have a huge selection of various games and decks of cards... and pizza for members ;)

p.s. yeah, this is the same night as Oxford Board Games Soc, but we'll be less crowded AND have food! and all the nerdiest strat chat.

When: 7-11PM, Saturday 29th October

Where: Undergraduate Social Area, Department of Computer Science 7 Parks Rd, Oxford OX1 3QG

Click 'Going' here: https://fb.me/e/2oezN9LEI

And More...

Laptop Stickers still available!

Gaming picture

The misprinted white ones are surprisingly popular huh...

Check out OxWoCS!

Make sure to join OxWoCS (Oxford Women in CS), a society we're working closely with, if you identify as a woman or woman-adjacent person in CS! They have a host of wonderful events on, including talks, panels and socials. Computer Science is unfortunately a subject with one of the biggest gender gaps in Oxford and worldwide, and addressing this is at the core of OxWoCS.

They have a (free) lunch social coming up on the 28th, in the CS Department atrium!

Follow Us

contact us at committee@ox.compsoc.net

Our Sponsors
Jane Street

Manage your email preferences or unsubscribe here

Designed with BEE Designed with BEE