statistics   206582

« earlier    

The statistical significance filter leads to overoptimistic expectations of replicability
"Treating a result as publishable just because the p-value is less than 0.05 leads to overoptimistic expectations of replicability. These overoptimistic expectations arise due to Type M(agnitude) error: when underpowered studies yield significant results, effect size estimates are guaranteed to be exaggerated and noisy. These effects get published, leading to an overconfident belief in replicability."
statistics  replicationcrisis 
just now by aapl
Fortnite's $100 million prize pool will take eSports to new heights: CHARTS - Business Insider
'Fortnite' is getting a $100 million prize pool — the biggest ever in eSports, and twice what was paid out in 2016 May 22, 2018 at 04:42PM

"Fortnite: Battle Royale" developer Epic Games announced this week that it will be putting $100 million towards prize pools for the game's upcoming competitive season— the largest single eSports prize pool to date. But while Epic's big prize package is sure to draw a lot of attention, it comes as eSports has become more lucrative overall.

As this chart from Statista shows, that total amount has been growing substantially every year. Epic's investment in "Fortnite," which is nearly as much as the $121 million given to eSports winners in 2017 overall, is likely to take the total amount of money in eSports prize pools to a new high in 2018.

Shayanne Gal/Business Insider

via Chart Of The Day
iftt  All  Feeds  Statistics  EVC 
2 hours ago by leconeyc
Still Not Significant | Probable Error
As well as being statistically flawed (results are either significant or not and can’t be qualified), the wording is linguistically interesting, often describing an aspect of the result that just doesn’t exist. For example, “a trend towards significance” expresses non-significance as some sort of motion towards significance, which it isn’t: there is no ‘trend’, in any direction, and nowhere for the trend to be ‘towards’.

Some further analysis will follow, but for now here is the list in full
statistics  humor 
6 hours ago by lena
The statistical significance filter leads to overoptimistic expectations of replicability - Statistical Modeling, Causal Inference, and Social Science
Treating a result as publishable just because the p-value is less than 0.05 leads to overoptimistic expectations of replicability. These overoptimistic expectations arise due to Type M(agnitude) error: when underpowered studies yield significant results, effect size estimates are guaranteed to be exaggerated and noisy. These effects get published, leading to an overconfident belief in replicability. We demonstrate the adverse consequences of this statistical significance filter by conducting six direct replication attempts (168 participants in total) of published results from a recent paper. We show that the published claims are so noisy that even non-significant results are fully compatible with them. We also demonstrate the contrast between such small-sample studies and a larger-sample study (100 participants); the latter generally yields less noisy estimates but also a smaller effect size, which looks less compelling but is more realistic. We make several suggestions for improving best practices in psycholinguistics and related areas.
9 hours ago by amy
How Not To Run an A/B Test – Evan Miller
Good advice on what not to do (don't look at results ahead of time!) when A/B testing. Points out that a Bayesian analysis makes more sense anyway.
statistics  analytics  testing 
9 hours ago by ruzel
Sample Size Calculator (Evan's Awesome A/B Tools)
Several statistical calculators especially helpful for testing.
statistics  tools  analytics  testing 
9 hours ago by ruzel
Robert Chapkin on Relationships Between the Gut Microbiome, Diet, and Colorectal Cancer | Cancer Network
1. Could you please provide some background on the beginnings of research into the relationship between gut microbiota and colon cancer? What are one or two pivotal studies in this area?

DR. CHAPKIN: Only recently has it been appreciated that intestinal microbiota influences the development of colorectal tumors in preclinical models and human subjects. There have been many studies highlighting the interplay between host genetics, gut microbes, and intestinal tumor initiation/progression. However, the seminal works of Arthur,[1] Tao,[2] and O’Keefe[3] stand out in my mind.

The Arthur laboratory findings provided novel insight into the complex effects of inflammation on microbial composition/activity and the host’s (intestinal mucosa) ability to protect itself from microorganisms with genotoxic capabilities. Also, from a mechanistic perspective, the Tao lab findings demonstrated that Frizzled proteins (receptors that regulate Wnt signaling, required for colonic stem cell maintenance, self-renewal, and repair of the epithelial lining) can be activated by toxin B produced by pathogenic bacteria (Clostridium difficile). These findings demonstrate that colonic stem cells are the target of some microbial toxins. Lastly, the O’Keefe lab findings demonstrated that dietary fiber and fat content have a remarkable effect on the colonic microbiota and its metabolic activity in high-risk vs low-risk cancer populations. His work supports the claim that total fiber intake needs to exceed 50 g per day to prevent colon cancer.

2. What key principles (eg, inflammation) or hypothesis models (eg, bacterial driver/passenger or keystone/pathogen models) are important to understanding how the gut microbiome modulates intestinal cells, immune cells, and malignant cells in the gut?

DR. CHAPKIN: There is ample evidence that the microbiome (eg, patchy bacterial biofilms) in the gut can modulate host immune cell function, promoting low-grade chronic inflammation. This, combined with intestinal barrier deterioration induced by (1) colorectal cancer–initiating genetic lesions, and (2) crosstalk between microbiota and diets low in fiber, results in the invasion of secreted microbial products (eg, oncotoxins), which can drive tumor growth. Intriguingly, the protective effects of dietary fiber on cancer development may be based on dramatic shifts in microbial community function—particularly short-chain fatty acid production—which stimulate gut mucosal metabolism and increase epithelial barrier function. These findings suggest that precision (personalized) nutrition might be utilized to favorably modulate an individual’s microbiota to reduce cancer risk.
cancer  nutrition  Science  genomics  Bioinformatics  statistics  HealthCare 
10 hours ago by cnk
Why User Reviews Matter For Your Business | Daily Infographic
Why User Reviews Matter For Your Business May 22, 2018 at 08:05AM

Why User Reviews Matter For Your Business

By Jerome | source:Here May 22nd, 2018

Just a few years ago, if you wanted to read a movie review, you had to read it in a newspaper from a professional movie critic. Yes, as hard as it is for younger generations to imagine, not everybody was a “professional” restaurant or movie critic before the early 2000s. It wasn’t until the Internet became popular and smartphone use exploded that we witnessed a huge influx of user reviews.

Websites like Yelp and Google’s Local Guides are just two of the websites that allow users to provide feedback. Now, don’t get me wrong. I love reading reviews. They are extremely beneficial for me when I’m making a decision about where to eat, and they’re also useful for business owners. You can learn a lot about a product or a service by reading what other people have to say, and businesses can evaluate that information to improve their customer service.

And that is exactly what today ‘s infographic focuses on. There’s a lot of data about user reviews and how they affect businesses. Did you know that user reviews increase sales by 18 percent, on average? I was shocked when I read that statistic, but user reviews help me decide whether to patronize a business. But then again, I, myself, usually skip buying a product, which wasn’t reviewed by a trusted source. User reviews are important — find out why.


When he's not chasing after his 2 rugrats, Jerome is looking for fast cars and toiling away on his bike.


via Daily Infographic
iftt  All  Feeds  Statistics  EVC 
14 hours ago by leconeyc
How a Kalman filter works, in pictures (2015)
Kalman filters are wonderful things and great for keeping track of where other things are in relation to yourself.
python  engineering  statistics 
17 hours ago by chris.leaman
Almost 40% of peer-reviewed dietary research turns out to be wrong. Here's why | New Food Economy
Food science has a huge statistics problem. The solution, for now? Stop treating new nutrition studies like they contain the truth.
statistics  food  health  research  science  article  bullshit  phacking 
19 hours ago by wwwald
Self-driving cars will be normal within 10 years, poll says - Business Insider
Roughly half of all Americans expect self-driving cars to become the norm within the next 10 years May 21, 2018 at 04:30PM

It's becoming increasingly common to see a self-driving car on the road these days, but we're still a few years out from considering them the norm.

In addition to the fact that car companies are still working out some technological flaws, a relatively large proportion of Americans are either against or hesitant about the concept of self-driving cars in general. And yet, as this chart from Statista shows, hesitant or not, a little more than half of the 1,503 Americans polled by Gallup expect autonomous cars to be commonly used within 10 years. Only a fourth of them expect it to take longer than 16 years.

Unsurprisingly, Western states are more likely than any other US region to believe it'll happen within the decade. Leading the way in testing driverless vehicles are Arizona and of course California, where Tesla owners are abundant, and companies have put a total of 409 autonomous test vehicles on the road. Considering the number of well-resourced entities putting money and time towards the effort, the 53% of Americans polled here might not be far off.

Shayanne Gal/Business Insider

Get the latest Tesla stock price here.

via Chart Of The Day
iftt  All  Feeds  Statistics  EVC 
yesterday by leconeyc

« earlier    

related tags

abtest  academic  ai  all  analysis  analytics  animals  apache  arguments  article  athletics  audio  awstats  babies  basketball  bayes  bayesian  best_of  bigdata  bioinformatics  book  books  bootcamp  bullshit  business  cairo  calculus  cancer  causality  charts  children  china  cities  coding  collaboration  community  correlation  courses  crash  crime  croydon  cs  cultural  culture  data  datascience  design  econometrics  economics  economie  economy  editorial  education  egypt  engineering  environment  estimation  evc  facebook  family  feeds  filter  finance  flexibleworking  food  free  futureofwork  genomics  geography  guncontrol  guns  health  healthcare  history  hn  humor  iftt  inequality  inference  insight  inspiration  javascript  journalism  jupyter  kalman  kids  lecture_notes  life  london  machine.learning  machinelearning  math  matrix  ml  money  nautilusmag  nl  nutrition  optimization  parenting  pca  phacking  pharmatest  philosophy  photography  police  population  poverty  predictions  prior  probability  programmer  programming  python  r  reference  regression  replicationcrisis  research  resource  science  sids  social_classes  society  sociology  sports  sql  stories  subscribers  syllabus  telecommuting  testing  tfl  time-series  tools  trains  transport  tube  tutorial  tweetstorm  uber  ubuntu  uk  ux  violence  visualization  wall_street  wealth  web  wikipedia  wishlist  women  writing  xgboost 

Copy this bookmark: