Hello everyone, I'm Mona Chalabi from FiveThirtyEight, and I analyse data on pubes and politics. Ask Me Anything!

Abstract

Hello everyone, I'm Mona Chalabi, a data journalist at FiveThirtyEight and I work with NPR to produce the Number Of The Week.

I try to think about data in areas where other people don't – things like what percentage of people pee in the shower, how many Americans are married to their cousins and (of course) how often people men and women masturbate. I'm interested in more sober topics too. Most recently, I worked on FiveThirtyEight's coverage of the UK election by profiling statistical outliers across the country. And I'm in London right now to work on a BBC documentary about the prevalence of racism in the UK.

I used to work for the Guardian's Data team in London and before that I got into data through working at the Bank of England, then the Economist Intelligence Unit and the International Organisation for Migration.

Here's proof that it's me.

I’ll be back at 1 PM ET to answer your questions.

Ask me anything! (Seriously, our readers do each week, so should you!)

I'M HERE NOW TO READ YOUR WEIRD AND WONDERFUL QUESTIONS AND DO MY BEST TO ANSWER THEM UPDATE: 30 MINS LEFT. KEEP THE QUESTIONS COMING!

UPDATE: My times up - I'd like to stay but the probability of me making typos/talking nonsense goes up exponentially with every passing minute. I'm so sorry I couldn't answer all of your brilliant questions but please do get in touch with me by email (mona.chalabi@fivethirtyeight.com) or on Twitter (@MonaChalabi) and I'll do my best to reply.

Hope the numbers are helping! xx

What is your take on the mathematician in Kansas who is seeking voting records to investigate fraud. http://www.kansas.com/news/politics-government/article17139890.html Do you know what type of analysis she has done?

Edit: s/he

EmceeDLT

I honestly don't know what sort of analysis she has done but I would like to. This is exactly the sort of story that makes people feel like maths matters in their lives. I know this sounds like a lame answer but it's also an honest one: I'll be looking into it.


Can you start a podcast? - "Pubes and Politics w/ Mona Chalabi"

djharrington88

Nate???! Are you listening to this?? The readers want what they want!


Most of the population seems to rely on other people to interpret data for them. What can we do as a society to make people more data literate?

brugaltheelder

This is a brilliant question and one that I think about every day.

I think the first step is to make people care about becoming data literate and to realise it's not difficult or boring to do. I think writing about non-conventional data topics like sex can help get people to care but they need to have some skepticism too.

I think articles like this can do a really good job of showing people just how ill-informed they are (or aren't!) and that can motivate people to care too.

http://www.theguardian.com/news/datablog/2014/oct/29/todays-key-fact-you-are-probably-wrong-about-almost-everything

Oh and the curriculum needs to be changed. On that small suggestion, I'll pass to the next question!


Can you remember a time where the use of statistics dramatically changed your opinion on something? A scenario where the stats disproved many of your preconceived notions about a topic?

rhiever

Seeing as this is my first post I just want to say THANK YOU for all these amazing questions! I really hope I don't disappoint!

To answer the first, yes! I can think of lots of examples - the one that comes to mind is a bit of an uncomfortable one though… I had an argument with someone about Muslim women. They told me that they were more likely to be unemployed than other women in Britain. That doesn't match my own experience and I feel like it supports some really negative stereotypes about Muslim women. But I did some research and statistically, they're right. That doesn't make the stereotypes ok but I do think it's a really good example of how data can change your mind even when it doesn't match with your own personal experience (or the argument that you instinctively want to make).


2 questions:

  1. What was your biggest "Why does this data set exist?" moment?
  2. What's your favourite correlation between two nominally unrelated phenomena?
ForLackOfAUserName

  1. I guess I can see the rationale for creating ANY data set so I've never really been too surprised. But I do find academic research super weird (qualitative and quantitative). Most recently my research led me to this… "Twinship, incest, and twincest in the Harry Potter universe" http://journal.transformativeworks.org/index.php/twc/article/view/576/457 which ¯_(ツ)_/¯
  2. There are so many good ones here! http://www.tylervigen.com/spurious-correlations - will try to think of a personal favourite. But don't you think people are noticing them all the time in their everyday lives? (albeit with some cognitive bias) eg "why does it always rain the one day that I straighten my hair??"

What was your education and career path like that led you to where you are now?

How would you recommend someone already a decade out of school to get into this line of work?

halalf

Well! Seeing as there were 3 of you that were interested in this! (and there's a saying amongst us very experienced data journalists which is 3 is greater than 1 - that's a bad joke sorry)

At the risk of sounding cheesy, it's never too late! Unlike most data journalists, I didn't start out in journalism and was never really drawn to the profession. I used to work in something called monitoring and evaluation in the humanitarian sector (sorry, that's all UN jargon but it's basically evaluating the level of need among vulnerable populations) it made me passionate about the importance of accurate numbers but frustrated with communicating them to a small group of so-called "experts" (who rarely include the individuals best positioned to actually do the fact-checking).

I ended up doing doing unpaid work experience at the Guardian 2 days a week (so that I could earn $$$$ the rest of the time) and I suppose they just got used to having me around. I recommend you do whatever you can to build up your experience - happy to give you more advice however I can (email me at mona.chalabi@fivethirtyeight.com) for now I have loads more qs to answer!!


Is there a 'Dear Mona' question that was just unpublishable because of ethical or moral reasons, but that you badly wanted to research and answer (and maybe you did)?

moebio

The only thing that would make a question unpublishable is if I can't find enough reliable data to answer it. The last time I remember that happening was about 2 weeks ago. A reader asked me if people were spending more time on the toilet now because they're using smartphones and handling business while they handle business. I could only find some really bad surveys and one study that had about 6 participants so that idea went down the drain (sorry, my colleagues will testify I love a terrible pun).


Given the steep rise of data analysis in just about every field imaginable, is there any area where you feel that data collection is most often manipulated?

Wierd_Carissa

I'm always really distrustful of data that we get sent by PR representatives of brands. ESPECIALLY when they're not willing to share their actual underlying data with us and only send over a series of neat looking statistics. Which is a shame because the private sector is collecting some of the most fascinating information about every aspect of our lives - if there was better oversight about its veracity, we as journalists could do some really great stuff with it!

Hope that answers your question!


I’m a huge fan of your “Dear Mona” series; I think it’s a brilliant modernization of “Dear Abby.” What are some of the weirdest or funniest questions that you’ve received from your readers? (Oh, and what did John Oliver ask you?!)

rhiever

Thanks Randy! I have had some very, very weird questions - someone recently asked me "what happens to the egg after it gets fertilised by the sperm once an individual is an adult i.e. is my mother's egg now in my arm or my lung??" (permit me one WTF!) I get cute ones too though. Someone told me he was going to ask his boyfriend to marry him and wanted to know the chances he would get a yes (unfortunately I couldn't help out)

I'm sorry, that's between me and John ;)

/ I really don't think it was that John Oliver.


Hello Mona, I think that Nate Silver and FiveThirtyEight are a very important addition to the media landscape today. Having what is usually solid data and analysis shown instead of the usual partisan parroting is refreshing. However, it seems that the team Nate Silver has put together includes a few journalists that don't really put much effort behind their data (and no this is not aimed at you). What is FiveThirtyEight doing to police it's own behavior, and how does such a small group of staff look objectively at the work of people that very well could also be a close friend? Do you see this as a potential danger to FiveThirtyEights journalistic integrity? I have been very disappointed at some of the articles coming from a few authors that can't even pass a basic logic test. And when this comes from the same few journalists over and over I have a hard time seeing how they still work there.

MonkRome

This is a great question. Data journalism can sometimes be surrounded in an undeserved halo. What we do is not perfectly objective (nothing is, not even science I'm afraid) but we truly strive to be accurate and honest in everything we do - and that comes before personal relationships and friendships in the newsroom.

Personally, I always read the comments below an article. If you feel I've messed up, I really want readers like you to tell me. Or if you don't want to publish publicly, please email me.


How surprised were you by the UK election result? I remember the outcome being outside the confidence interval predicted by 538. Any ideas why the prediction was so far from the outcome?

Captain_Wozzeck

I think there are lots of reasons and the ones I would have given you on May 8th are probably different to the one I'll give you now… right now though, the answer that comes to mind is that journalists took far too much heart from Nate's success in 2012. Britain is so very very different to America (the past 19 months has left me in no doubt about that!) and so the data and the way that you understand it isn't the same at all. I think that's sort of a moral in so much data journalism - people assume that objectivity means one size fits all. It doesn't. Context is everything.

Hope that makes sense!


What data do you have on pubes?

cowlick206

http://fivethirtyeight.com/datalab/au-naturel-or-barely-there-the-data-on-pubic-hair-preferences/

Your username together with your question has given me a pretty gross mental image. I'd now like to pass to another question.


How big of a deal is Bernie Sander's inability to capture African American votes? Are African Americans really eager to support Hillary? Otherwise, why would it be such a huge deal? Also, if Sanders hypothetically won the nomination, wouldn't he likely pick up a large portion of African American voters? I've heard that well over half of African Americans are Democrats and since the Republicans seems to still alienate them, would that really be a problem in the general election?

Iam_a_Jew

So many questions here that are worth understanding between now and November (but I'm not able to right now without at least another hour). We'll be discussing our coverage of race and voting intention at 538 so I'll pass these qs onto our editors … thank you!


What is your favorite statistical anomaly?

rhiever

Total mental block - sorry Randy! The only thing that comes to mind right now (literally because two twins just walked past my window) is this article http://fivethirtyeight.com/datalab/more-twins-fewer-triplets/ where I was genuinely surprised to find data that showed twins were on the rise in America but triplets were declining and had to figure out why. It wasn't the explanation I was expecting at all!


We've all heard about "lies, damn lies, and statistics." What's your favorite example of someone using statistics to mislead people on a grand scale?

TeaRecs

Rather than citing a specific statistic here, I'd rather point to one sloppy method (of many that I come across)

Please do not tell me about some insane percentage change if you can't tell me the base. An increase of 200% doesn't mean much if the original number was 1 (unless of course the subject is like median number of penises on men, then I am all ears - now wondering if my AMA will beat the one on diphallia. I doubt it.)


What is the most beautiful data you have ever seen?

cybercuzco

Sometimes even spreadsheets can be beautiful so long as you remember the numbers are about people (e.g. recovery rates in physical and mental health)

As for visualisations: oof! We have an incredible team at 538 who elevate my work from meh to woah every day.

eg Baby's first profanity by Allison Mccann http://fivethirtyeight.com/datalab/babys-first-profanity/ and UK election maps by Reuben Fischer-Baum http://fivethirtyeight.com/features/all-politics-is-local-even-in-the-most-average-place-in-the-uk/


Was the intention of the Dear Mona series to focus on so many sex topics from the beginning? Or is that just the type of question people keep sending in?

jiggabot

It really, really wasn't but readers often want to ask me questions that they don't feel comfortable asking other people in their lives so sex comes up a lot


Are you related to Ahmed Chalabi of Iraq War fame?

doopadoopadoodle


Are there any questions you're waiting to have asked because you really want to dive into the stats?

Are there any questions you're surprised no one has asked?

What numbers should more people know that hardly anyone does?

So glad you're doing this AMA, by the way!

liverpud

I'm glad I'm doing it too!

I wish people would ask me more about hip hop and skateboarding but alas I think a lot of 538 readers might have other interests!

I'm sort of surprised people don't ask me more questions about welfare (that might be the Brit in me!) but it's a policy area around which data journalism could be really informative.

I don't think enough people understand the numbers on sexual assault - they understand it happens but they just don't appreciate the prevalence (some BJS data here for those of you in the US who are interested http://www.bjs.gov/index.cfm?ty=tp&tid=317)


First thing you look at or do when you're given a brand new dataset (let's assume it's already been cleaned)?

quitelargeballs

Never assume it's already been cleaned! To do that, I know it sounds like BS but I do a commonsense check - I remember looking at data on doctors' surgeries opening hours and noticing that some of the rows said that they were open on average for 30 hours per day!!


When approaching an issue, do you approach the data agnostically with no predetermined hypothesis as to the outcome or do you have a hypothesis that you then use the data to prove or disprove?

I've been doing these types of analyses for work and I feel like I should do it the former way, but my mind always wants to do it the latter way and I have to force it back.

FoonaLagoonaBaboona

Everyone ALWAYS has a hypothesis (even if your hypothesis is that the data will tell you nothing) and I think it's good to articulate that rather than pretending it isn't the case. Then come up with a counter hypothesis, go back to the data and see if that stands up too!


What are your fundamental rules for making a visualization? i.e. are there certain things that you need to have or do in each visualization?

sarahbotts

I know this might sound stupid but the biggest rule is that it has to tell you something. If you have to lean your nose closer to the screen, reread the paragraph before the visualisation or close the tab I've failed miserably!


I've always been surprised that all FiveThirtyEight graphs are formatted and styled in exactly the same way. Do you ever find this restricting?

dmeret

Not at all! I think having a default chart style can be really helpful because it makes you think about why you want to do something in the non-default way! When people just want a quick takeaway from a chart (e.g. weird, it went up in 1914, then back down in 2014) sometimes simple is really good. Sometimes though the data demands a different format eg what's the average age difference in a couple? http://fivethirtyeight.com/datalab/whats-the-average-age-difference-in-a-couple/

Before I joined FiveThirtyEight I experimented with some seriously weird graphs using photos as part of an exhibit on Iraq. I'm not sure it was actually successful or effective but it was a really good way to get me to think creatively about visual journalism! (https://www.youtube.com/watch?v=LX89EGtb2xs if you're interested)


What data from Dear Mona was the hardest to find ? Is there any topics you have to work extra hard on to get serious statistics ?

mathildeboireau

All the time! The newsroom wouldn't function without my colleague Andrew Flowers who helps all the time with digging up data and making it usable and incisive. I would have never been able to figure out the most common name in America without his help!

http://fivethirtyeight.com/features/whats-the-most-common-name-in-america/

ps Sorry about the incessant links to 538 posts - this honestly isn't intended as shameless self promotion just honestly trying to provide examples of what I'm talking about!


Additional Assets

License

This article and its reviews are distributed under the terms of the Creative Commons Attribution 4.0 International License, which permits unrestricted use, distribution, and redistribution in any medium, provided that the original author and source are credited.