I worked on statistical studies /systems for many many years, not only in the highly-regulated pharmaceutical world but prior to that in various marketing and financial planning environments.
So although not a statistician myself feel that my experience is not completely irrelevant. So I’ll try a simple answer to your question.
Because in the interests of openness and transparency one does not only publish the data and results that clearly and obviously support the final conclusion, but all of those that do not as well, together with the rationale as to why such-and-such a figure may not be meaningful in the broader scope of the study or simply “not statistically significant”.
That latter term is often mis- or completely un-understood by the press, such that you will even see headlines making outrageous claims, with the caveat of “the results, while not statistically significant, show that…” . I always want to grab them by the throat and shout in their faces “No, if they’re not statistically significant, they do not show anything”.
And for a lay-person take on “not statistically significant” it can be thought of as “the numbers calculated are just as likely to have come about by a random selection as they are by a genuine underlying cause”.
But we still publish them, it’s part of the scientific method to show that you’re not just ignoring the stuff that doesn’t agree with your hypothesis.
Do you understand scientific results? They can never show anything. They can offer evidence in support of or against a theory or hypothesis. The statistically significant cut-off of 95% certainty in the result is totally arbitrary. This is a real problem, as many results that meet the criteria but are from badly designed experiments are judged OK, but results from much better experiments may not pass this test and are difficult to publish. Unfortunately it is easy to look at the stats and difficult to judge the quality.
Yes, that would be nice. Peer review tries, but it is naturally very difficult. IMO, the science community does a good job in the long run. The pressure to publish and the mis-use of results to further agendas of all kinds is a big negative. Then there is the lay public trying to make sense of it all and the politicians…
Just from the way you phrase that it’s clear that you do not understand it.
Go away and study statistics for a while before you make such ridiculous comments.
Just as a hint, the 95% is not, in any way, a measure of “certainty”. As I mentioned a few posts ago, it’s a measure of confidence as to whether any results, positive or negative, could have occurred by chance.
How significant (in a more everyday sense of the word) those results may be is in no way related to this measure.
I can see we are not going anywhere with this discussion. Certainty was the wrong word, sorry. I was using the 95% cut-off as an example. Obviously most statistical treatments of medical data use other measures that vary with the data set. In the end, a cut-off will be used for statistical “significance”.
As for studying statistics - I’ve done my bit. Planning experiments with a statistician was always a pain in the butt, my experience was that they often had little appreciation of the science. I stand by my comments.
A 95% confidence interval (CI) is used becauseit balances the trade-off between the precision of an estimate and the reliability of the result, providing a reasonable compromise between a too-narrow (and potentially unreliable) interval and a too-wide (and potentially useless) interval.
It represents a range where the true population parameter is expected to fall in approximately 95% of repeated sampling studies, and its use is largely a historical convention established by statisticians like Sir Ronald Fis
A 95% confidence interval does not have a fixed “cut-off” in the same way a p-value does, but rather it represents a threshold for rejecting the null hypothesis, with the corresponding 5% non-confidence level often considered a cut-off for statistical significance
Yes, we are getting a bit into semantics, threshold/cut-off.
Regarding p-values:
“n null-hypothesis significance testing, the p-value is the probability of obtaining test results at least as extreme as the result actually observed, under the assumption that the null hypothesis is correct. A very small p-value means that such an extreme observed outcome would be very unlikely under the null hypothesis. Even though reporting p-values of statistical tests is common practice in academic publications of many quantitative fields, misinterpretation and misuse of p-values is widespread and has been a major topic in mathematics and metascience”. In other words, the statement that a p value has a fixed cut-off is wrong. A cut-off is where one chooses to put it. p-value testing is generally a low bar. The quality of the the data is a much more important parameter, however quality is very difficult to measure.
The problem with statisticians is that they are useless at actually producing any data. They just seem to get off on slagging those that actually do the hard work.
At least in Europe, the number of cases has a seasonal behavior with peaks in later winter / spring because with lower temperatures people spends more time indoors.
Great iniative, at a minimum will provide an alternative view
"A coalition of West Coast, Democratic-led states on Wednesday announced a new public health alliance formed in defiance of Health and Human Services (HHS) Secretary Robert F. Kennedy Jr., with the states saying it will provide “science-driven” advice on vaccines.
California Gov. Gavin Newsom (D), Oregon Gov. Tina Kotek (D) and Washington Gov. Bob Ferguson (D) announced the launch of the West Coast Health Alliance just days after Kennedy sacked the director of the Centers for Disease Control and Prevention (CDC). The health secretary has also overhauled the CDC’s vaccine advisory board, appointing multiple like-minded vaccine skeptics.
“President Trump’s mass firing of CDC doctors and scientists — and his blatant politicization of the agency — is a direct assault on the health and safety of the American people,” the governors said in a joint statement. “The CDC has become a political tool that increasingly peddles ideology instead of science, ideology that will lead to severe health consequences. California, Oregon, and Washington will not allow the people of our states to be put at risk.”
“Bayesian meta-analysis revealed a link between vaccines and CAD risk (OR, 1.70; 95% CrI: 1.11–2.57), particularly after BNT162b2 (OR, 1.64; 95% CrI: 1.06–2.55) and second dose (OR, 3.44; 95% CrI: 1.99–5.98). No increased risk of heart attack, arrhythmia, or stroke was observed post-COVID-19 vaccination. As the only noteworthy point, a protective effect on stroke (OR, 0.19; 95% CrI: 0.10–0.39) and myocardial infarction (OR, 0.003; 95% CrI: 0.001–0.006) was observed after the third dose of the vaccine.”
So OP cherry picking directly !!
“Titbits “ ???
not my “assertions” . just quoting results (IN FULL) from same study as OP.
“However, no significant relationship was detected between vaccination and stroke, myocardial infarction, and arrhythmia. And just a protective effect on stroke (OR, 0.19; 95% CrI: 0.10–0.39) and myocardial infarction (OR, 0.003; 95% CrI: 0.001–0.006) was observed after the third dose of the vaccine.”
if you mean marton quoting this.. he should check what it means..
No they are not. What is immoral is allowing preventative diseases to spread unhindered through the community making children seriously ill and killing some.
