6/n A common method of doing this is to use the inverse of the variance of the studies. This essentially uses the confidence intervals of each study as a weighting tool, with tighter intervals getting a higher weightpic.twitter.com/yW7WF1YY8o
Epidemiologist. Writer (Guardian, Observer etc). "Well known research trouble-maker". PhDing at @UoW Host of @senscipod Email gidmk.healthnerd@gmail.com he/him
You can add location information to your Tweets, such as your city or precise location, from the web and via third-party applications. You always have the option to delete your Tweet location history. Learn more
Add this Tweet to your website by copying the code below. Learn more
Add this video to your website by copying the code below. Learn more
By embedding Twitter content in your website or app, you are agreeing to the Twitter Developer Agreement and Developer Policy.
| Country | Code | For customers of |
|---|---|---|
| United States | 40404 | (any) |
| Canada | 21212 | (any) |
| United Kingdom | 86444 | Vodafone, Orange, 3, O2 |
| Brazil | 40404 | Nextel, TIM |
| Haiti | 40404 | Digicel, Voila |
| Ireland | 51210 | Vodafone, O2 |
| India | 53000 | Bharti Airtel, Videocon, Reliance |
| Indonesia | 89887 | AXIS, 3, Telkomsel, Indosat, XL Axiata |
| Italy | 4880804 | Wind |
| 3424486444 | Vodafone | |
| » See SMS short codes for other countries | ||
This timeline is where you’ll spend most of your time, getting instant updates about what matters to you.
Hover over the profile pic and click the Following button to unfollow any account.
When you see a Tweet you love, tap the heart — it lets the person who wrote it know you shared the love.
The fastest way to share someone else’s Tweet with your followers is with a Retweet. Tap the icon to send it instantly.
Add your thoughts about any Tweet with a Reply. Find a topic you’re passionate about, and jump right in.
Get instant insight into what people are talking about now.
Follow more accounts to get instant updates about topics you care about.
See the latest conversations about any topic instantly.
Catch up instantly on the best stories happening as they unfold.
6/n A common method of doing this is to use the inverse of the variance of the studies. This essentially uses the confidence intervals of each study as a weighting tool, with tighter intervals getting a higher weightpic.twitter.com/yW7WF1YY8o
7/n Here's an example from the ivermectin literature - this is simply a weighted average where the weighting is derived from the inverse of the variance (calculated from the confidence intervals)pic.twitter.com/ux75gk6HFz
8/n But you can immediately see the issue here - this weighting is DERIVED FROM THE DATA If the data is unreliable, the weighting is meaningless!pic.twitter.com/7Noojnz1iQ
9/n If I were to, say, fabricate a large trial that had a heavy weight, it would completely mess up the meta-analysis and make the results unreliable For ivermectin, we KNOW that this has happenedhttps://gidmk.medium.com/is-ivermectin-for-covid-19-based-on-fraudulent-research-5cc079278602 …
10/n So this is what we mean by GIGO. If you incorporate bad numbers into a meta-analysis, by definition the results are also bad, because the model is simply an average of the numbers you input 
11/n This is why most of the work of meta-analysis is to carefully choose the studies you use, because if your model is based on garbage the results will also be garbage
12/n A good way to think of this for laypeople is to think of a simple, boring average Would you average out the included studies? If not, the meta-analysis probably doesn't make sense
13/n I think the website ivmmeta is actually a brilliant teaching tool for what not to do here. Can you imagine adding up the time it takes for people to recover and the proportion of people who had symptoms and dividing by 2? What would that even mean?pic.twitter.com/Cd9Bg2rWsV
"Can you imagine adding up the time it takes for people to recover and the proportion of people who had symptoms and dividing by 2?" Funny that the website *clearly* does not do that
They transform the values into odds ratios then combine with a random effects inverse variance model to generate a weighted mean. While this involves more transformations, at a theoretical level it's very similar
One way to try and think about this is to look at the included estimates and think about what that weighted average means for each of them. Does it mean a % reduction in death? Does that same % apply to symptoms? If we analyse only one of those, would we get the same result?
For sure can't *fully* believe these RR. But when 1 study finds seatbelts reduce deaths by x% (but p>0.05), and a 2nd finds they reduce hospitalization by y% (also p>0.05), would you dismiss seatbelts? When maybe the combined "prevent-damage-score" was stat.sig.
They should maybe stop calling it RR. Call it estimated combined "impact score", or something
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.