Jump to content

~*Official #COVID-19 Thread of Doom*~


Recommended Posts

41 minutes ago, Air_Delivery said:

They didn't know it. They took a chance based on what they thought but the data didn't come in till months later. 
 

 

The UK looked at the data in the AZ trials, as well as the mRNA vaccines, and recognized a pattern that was highly suggestive that a single dose provided ample protection and as such delaying the second dose was a good idea. The first big block of data out of Israel came out in the second week of January, only roughly 4 weeks after the UK began dosing.

 

I think the disconnect here is that the trials did not attempt to ascertain efficacy after 1 dose, but that doesn't mean there was no data to suggest the answer to that question in the process. There was, which is why the UK went the route they did. They had good reason to believe what they did, it wasn't like the accidentally came upon a good strategy.

Link to post
Share on other sites
4 hours ago, Joe said:

 

I mean are we actually following the trials as conducted? We don't give the doses at exactly 21 and 28 day intervals, for example. I joke about that as a schtick, but people smarter than you and I concluded that it was extremely reasonable to not be super precise with the 2nd dose date and give 4 or so days leeway. Same thing with the single shot strategy. It's not like some dude just threw it out there for shits and giggles. 

 

I think it was always obvious that +/- a day was fine. I think everyone figured even a week or two of delay on the second dose was fine. The real uncertainty was on how long dose 1 immunity would last and if you could still get the booster effect from it with the second shot if you waited too long.

Link to post
Share on other sites
23 minutes ago, Jason said:

 

I think it was always obvious that +/- a day was fine. I think everyone figured even a week or two of delay on the second dose was fine. The real uncertainty was on how long dose 1 immunity would last and if you could still get the booster effect from it with the second shot if you waited too long.

 

Why? We don't have data on that. It's not just +/- a day. It's 4 days.

Link to post
Share on other sites

What the UK scientists were really banking on was that the efficacy of the first dose in an environment with MUCH higher infection derived immunity plus the rapid increase on vaccine derived immunity would be significantly higher than even what they had found in the trial data on antibody levels over time.


You have to remember the conditions at the time these trials were done we significantly different than where we were even in mid December when they decided this was the likely path they would take.

 

Brilliant move and courageous in the face of unfounded mockery of their plan. They saved a lot of lives.

Link to post
Share on other sites
1 hour ago, skillzdadirecta said:
woman-sitting-on-luggage-3943882.jpg?ito
WWW.IPSOS.COM

Latest Axios/Ipsos poll shows removing the one-dose shot from the market has not impacted people's likelihood to get vaccinated

 

Guess enthusiasm in the country hasn't been as dampened as it was HERE. 


That poll doesn’t say what you think it says.

  • Upvote 1
Link to post
Share on other sites
23 minutes ago, b_m_b_m_b_m said:

"The polls aren't wrong" guy trying to draw conclusions within the margin of error

 

 


What?

Link to post
Share on other sites
31 minutes ago, b_m_b_m_b_m said:

"The polls aren't wrong" guy trying to draw conclusions within the margin of error

 

 


b_m: The polling guy should stay in his fucking lane!

 

Also b_m: The polling guy doesn’t understand polls, according to me!

 

Lmao

Link to post
Share on other sites
2 minutes ago, b_m_b_m_b_m said:

Overlapping margin of error means the results are not that different actually. And only one data point does not a trend make.

 

I mean it's from the same pollster using the same methodology. It's not comparing different polls from different pollsters. I think that 6% is significant.

Link to post
Share on other sites
6 minutes ago, b_m_b_m_b_m said:

Overlapping margin of error means the results are not that different actually. And only one data point does not a trend make.

 

Also, the overlap you are talking about here is on the absolute edge of both polls. You are really stretching to try to dunk on Nate, mi amigo.

Link to post
Share on other sites
56 minutes ago, Joe said:

 

Also, the overlap you are talking about here is on the absolute edge of both polls. You are really stretching to try to dunk on Nate, mi amigo.


As if we’d expect anything less from him? Come on, Jack. 

  • Haha 2
Link to post
Share on other sites
25 minutes ago, Joe said:

 

Also, the overlap you are talking about here is on the absolute edge of both polls. You are really stretching to try to dunk on Nate, mi amigo.

The amount of overlap is irrelevant when trying to determine the true value, they overlap, meaning sampling variability could cause this difference. And yes, that is true even with the same pollster.

 

Anyway, the +/-3.3% is for each subcategory. So very likely is 24% +/-3.3% and somewhat likely is 22 +/-3.3%. so really it looks like not very likely might have ticked up a bit and everything else is pretty much the same. This poll is good to see trends over time in opinion, and this change between the new poll and the one from earlier this month could just be noise, there's not enough data to say that opinion has shifted due to the pause. 

Link to post
Share on other sites

Actually I take that back, 3.3% is for the whole population of adults. The subgroup for the question regarding children under 18 is a smaller group than "all adults" so the margin is larger for this subgroup. Whoops!

Link to post
Share on other sites

Again, the trend matters when it’s the same pollster using the same methodology. If a pollster has Biden up 3 and then up 6 two weeks later, they have him up in their polling. You don’t say that they are the exact same poll because the ends of their MoE overlap.

Link to post
Share on other sites

Two points don't make a trend lmao

 

And the subgroup "adults with children under 18" has a greater margin of error than the 3.3% for the total sample population of "all adults" and is not published in the poll.

 

Quote

The margin of sampling error is plus or minus 3.3 percentage points at the 95% confidence level, for

results based on the entire sample of adults. The margin of sampling error takes into account the

design effect, which was 1.19. The margin of sampling error is higher and varies for results based on

sub-samples. 

 

Link to post
Share on other sites

Two points don't necessarily make a trend, but they certainly could and it's not completely outrageous to comment on it. All you want to do is dunk on Nate, which is obvious with the bs strawman you put in quotes.

Link to post
Share on other sites

This is a good post about what we are talking about. You're saying that the overlapping region pictured here is enough to dismiss the difference, but it's not. Also should be noted that the overlap you are talking about is even smaller than this picture.

 

1*dPwBA_t1q9jJEqa-HmfQaA.png

 

1*4xJeTki9ww4ZM9aBNXxixQ.png
TOWARDSDATASCIENCE.COM

“The confidence intervals of the two groups overlap, hence the difference is not statistically significant” — A lot of People

 

  • Upvote 1
Link to post
Share on other sites

 

Just now, Joe said:

This is a good post about what we are talking about. You're saying that the overlapping region pictured here is enough to dismiss the difference, but it's not. Also should be noted that the overlap you are talking about is even smaller than this picture.

 

1*dPwBA_t1q9jJEqa-HmfQaA.png

 

1*4xJeTki9ww4ZM9aBNXxixQ.png
TOWARDSDATASCIENCE.COM

“The confidence intervals of the two groups overlap, hence the difference is not statistically significant” — A lot of People

 

Quote

The margin of sampling error is plus or minus 3.3 percentage points at the 95% confidence level, for

results based on the entire sample of adults. The margin of sampling error takes into account the

design effect, which was 1.19. The margin of sampling error is higher and varies for results based on

sub-samples. 

Again, 3.3% is for the full population, and the subgroup of "adults with children under 18" has a higher margin, as stated in the poll itself. Drawing a conclusion based on two data point with this information is fucking stupid, and Mr polls should know this stuff.

Link to post
Share on other sites

Noting changes to a poll, even within the MOE, pretty standard fare analysis. More data points will obviously help shed light on noise vs. trend. I do think it’s important to note that results falling within overlapping MOEs does not mean they are the same, only that they could be the same.

  • Upvote 1
Link to post
Share on other sites
37 minutes ago, sblfilms said:

Noting changes to a poll, even within the MOE, pretty standard fare analysis. More data points will obviously help shed light on noise vs. trend. I do think it’s important to note that results falling within overlapping MOEs does not mean they are the same, only that they could be the same.

 

Exactly. This dunk is predicated on the overlap being the only possibility when it is in fact the less likely possibility. Also, you can't exclude the value of Bayesian inference here (using outside knowledge to determine probability, something UK employed in their one-shot approach). Given that we know an event happened that would make people doubt J&J more, we have even more of a reason to believe that this shift isn't just noise. A third data point will definitely make this point iron-clad, though.

Link to post
Share on other sites

You can't rule out it is statistical noise because the data is not there. You don't know the margin for the subgroup. That anyone can make a statement explaining differences in the data without the complete data is bad if your brand is data

Link to post
Share on other sites
1 minute ago, b_m_b_m_b_m said:

You can't rule out it is statistical noise because the data is not there. You don't know the margin for the subgroup. That anyone can make a statement explaining differences in the data without the complete data is bad if your brand is data

 

This is going to look like a very silly discussion when the third data point confirms the trend in 2 weeks lol.

 

And homie isn't about to put a full disclaimer on every tweet. Come on now!

  • Upvote 1
Link to post
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

  • Recently Browsing   0 members

    No registered users viewing this page.

×
×
  • Create New...