New Book on Statistical Significance and p-Values
17 views
Skip to first unread message
Don Macnaughton
Apr 25, 2021, 8:20:25 PM4/25/21
to MedStats
I favor the continued use of statistical significance (e.g., p-values with thresholds) as a necessary (though not sufficient) gateway to the publication of empirical-research papers in scientific journals. This gateway was invented by Karl Pearson, Ronald Fisher, Jerzy Neyman, and Egon Pearson between 1900 and 1935. These statisticians focused on the needs of scientific researchers, though the modern focus of the gateway is on the needs of society.
The threshold-p-value gateway to publication is sensible because it helps journal editors and researchers maximize the overall scientific and social benefit of scientific research. The gateway does this by producing a roughly optimal trade-off between the rates of costly false-positive and costly false-negative errors that a journal makes in accepting or rejecting papers for consideration for publication.
I have recently published a book that explains the logic behind the preceding ideas. The book's title is
The threshold-p-value gateway to publication is sensible because it helps journal editors and researchers maximize the overall scientific and social benefit of scientific research. The gateway does this by producing a roughly optimal trade-off between the rates of costly false-positive and costly false-negative errors that a journal makes in accepting or rejecting papers for consideration for publication.
I have recently published a book that explains the logic behind the preceding ideas. The book's title is
The War on Statistical Significance:
The American Statistician vs. the New England Journal of Medicine
The book is aimed at scientific researchers, journal editors, science teachers, and science students. It will also be of interest to some statisticians for its carefully reasoned arguments about the fundamental practical role of the field of statistics in scientific research.
More information is available at war-stat-sig.com.
Donald B. Macnaughton
Basilio de Braganca Pereira
Apr 25, 2021, 9:55:01 PM4/25/21
to meds...@googlegroups.com
I suggest reading
Gert Gigerenzer - Mindless Statistics ( available in the internet)
Basilio
Enviado do meu iPhone
--
--
To post a new thread to MedStats, send email to MedS...@googlegroups.com .
MedStats' home page is http://groups.google.com/group/MedStats .
Rules: http://groups.google.com/group/MedStats/web/medstats-rules
---
You received this message because you are subscribed to the Google Groups "MedStats" group.
To unsubscribe from this group and stop receiving emails from it, send an email to medstats+u...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/medstats/99ca0a58-9a9e-4664-b752-fe58bf338679n%40googlegroups.com.
A Indrayan
Apr 25, 2021, 11:31:56 PM4/25/21
to meds...@googlegroups.com
Is this allowed for this group?
Sent from my iPhone
On 26-Apr-2021, at 5:50 AM, Don Macnaughton <don...@matstat.com> wrote:
John Whittington
Apr 26, 2021, 9:01:41 AM4/26/21
to meds...@googlegroups.com
At 04:31 26/04/2021, A Indrayan wrote:
Is this allowed for this group?
Abhaya, I would be happy to consider alternative views, if anyone has
them, but I decided that it was probably OK to allow this. Although
in some senses 'advertisements' (since books cost money), I think that
announcements relating to newly published books (in relation to subjects
relevant to the group) are probably acceptable in groups like
this.
One of the things which swayed me in my decision is the fact that, for at least couple of years, Martin Holt (the owner of this group/list) included in the signature of every one of his posts to the group an 'announcement' relating to a book (costing some 8 times the price of Don's recently announced one) which he co-authored with yourself. I took this to mean that, as list owner, he considered this to be an acceptable practice in this group.
However, as above, I would be interesting and pleased to hear of any other views (in any direction) about this.
I'm very sorry to hear what is currently happening in India as regards the pandemic. I hope that you and yours are managing to keep well, and certainly hope that something can be done soon to get things at least vaguely 'under control' in your country - although it is clearly going to be an uphill struggle, even with the international help and support which I imagine (hope) you will be getting.
Kind Regards,
John
One of the things which swayed me in my decision is the fact that, for at least couple of years, Martin Holt (the owner of this group/list) included in the signature of every one of his posts to the group an 'announcement' relating to a book (costing some 8 times the price of Don's recently announced one) which he co-authored with yourself. I took this to mean that, as list owner, he considered this to be an acceptable practice in this group.
However, as above, I would be interesting and pleased to hear of any other views (in any direction) about this.
I'm very sorry to hear what is currently happening in India as regards the pandemic. I hope that you and yours are managing to keep well, and certainly hope that something can be done soon to get things at least vaguely 'under control' in your country - although it is clearly going to be an uphill struggle, even with the international help and support which I imagine (hope) you will be getting.
Kind Regards,
John
On 26-Apr-2021, at 5:50 AM, Don Macnaughton <don...@matstat.com> wrote:
I favor the continued use of statistical significance (e.g., p-values with thresholds) as a necessary (though not sufficient) gateway to the publication of empirical-research papers in scientific journals. This gateway was invented by Karl Pearson, Ronald Fisher, Jerzy Neyman, and Egon Pearson between 1900 and 1935. These statisticians focused on the needs of scientific researchers, though the modern focus of the gateway is on the needs of society.
The threshold-p-value gateway to publication is sensible because it helps journal editors and researchers maximize the overall scientific and social benefit of scientific research. The gateway does this by producing a roughly optimal trade-off between the rates of costly false-positive and costly false-negative errors that a journal makes in accepting or rejecting papers for consideration for publication.
I have recently published a book that explains the logic behind the preceding ideas. The book's title is
The War on Statistical Significance:
The American Statistician vs. the New England Journal of Medicine
The book is aimed at scientific researchers, journal editors, science teachers, and science students. It will also be of interest to some statisticians for its carefully reasoned arguments about the fundamental practical role of the field of statistics in scientific research.
More information is available at war-stat-sig.com.
Donald B. Macnaughton
John
----------------------------------------------------------------
Dr John Whittington,
Voice: +44 (0) 1296 730225
Mediscience Services
Fax: +44 (0) 1296 738893
Twyford Manor, Twyford, E-mail:
Joh...@mediscience.co.uk
Buckingham MK18 4EL,
UK
----------------------------------------------------------------
Abhaya Indrayan
Apr 26, 2021, 10:27:09 AM4/26/21
to meds...@googlegroups.com
John:
I do not want to defend what Martin Holt did regarding his book in his signature but that onus is not on me although I am a co-author. That was not in the body of the post. If this kind of post is allowed in this group, absolutely fine with me. I just wanted to know.
Regarding covid in India, we are indeed struggling but I do not think the onslaught is going to be as bad as in some other countries per million population despite our limited resources and illiterate population. Help is coming but that is nothing relative to our population and we have to depend on our own capacities. The surge was sudden and steep and nobody was prepared.
Warm regards.
~Abhaya
--
Dr Abhaya Indrayan
Personal website: http://indrayan.weebly.com
--
--
To post a new thread to MedStats, send email to MedS...@googlegroups.com .
MedStats' home page is http://groups.google.com/group/MedStats .
Rules: http://groups.google.com/group/MedStats/web/medstats-rules
---
You received this message because you are subscribed to the Google Groups "MedStats" group.
To unsubscribe from this group and stop receiving emails from it, send an email to medstats+u...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/medstats/202104261301.13QD1aYL029678%40mail102c50.megamailservers.eu.
John Whittington
Apr 26, 2021, 11:06:39 AM4/26/21
to meds...@googlegroups.com
Abhaya,
I wasn't implying that you had anything to do with, or necessarily even approved of, what Martin did.
I was merely explaining that, in deciding what view to take of the recent post, I was strongly influenced by the fact that Martin (as 'list owner', hence 'the creator of the rules') repeatedly posted his announcement of the book (albeit not in the body of his posts) - seemingly implying that he regarded such 'announcements' as acceptable in this group.
Pragmatically, I cannot undo the fact that the recent post was distributed to most MedStats members, and I certainly don't intend to 'ban' Don for posting it - so my interest in people's views relates to the future. In fact, since it is not a 'moderated' list, I'll never be able to prevent such messages being posted (and distributed) in the future, but if the membership felt that they did not want to see 'new book announcements' I could at least warn any perpetrators 'not to do it again' (or risk being ';banned'!). Hence, any other views would be welcome.
As for Covid in India, I agree that, even at present, the figures 'per million of population' are fairly modest in comparison with most countries (including the UK) - and that remains the case even if one makes allowance for published figures being (I fear) considerable under-0estimates of the true figures. I have to say that I'm a little surprised by that, since I would have thought that the nature of the country/population was such that 'non-pharmaceutical interventions' ('social distancing', 'lockdowns' etc. and education etc.) would be much more difficult to implement than in, say, 'Western' countries - but I am certainly glad that my 'surprise' is perhaps not all that justiified.
Kind Regards,
John
I wasn't implying that you had anything to do with, or necessarily even approved of, what Martin did.
I was merely explaining that, in deciding what view to take of the recent post, I was strongly influenced by the fact that Martin (as 'list owner', hence 'the creator of the rules') repeatedly posted his announcement of the book (albeit not in the body of his posts) - seemingly implying that he regarded such 'announcements' as acceptable in this group.
Pragmatically, I cannot undo the fact that the recent post was distributed to most MedStats members, and I certainly don't intend to 'ban' Don for posting it - so my interest in people's views relates to the future. In fact, since it is not a 'moderated' list, I'll never be able to prevent such messages being posted (and distributed) in the future, but if the membership felt that they did not want to see 'new book announcements' I could at least warn any perpetrators 'not to do it again' (or risk being ';banned'!). Hence, any other views would be welcome.
As for Covid in India, I agree that, even at present, the figures 'per million of population' are fairly modest in comparison with most countries (including the UK) - and that remains the case even if one makes allowance for published figures being (I fear) considerable under-0estimates of the true figures. I have to say that I'm a little surprised by that, since I would have thought that the nature of the country/population was such that 'non-pharmaceutical interventions' ('social distancing', 'lockdowns' etc. and education etc.) would be much more difficult to implement than in, say, 'Western' countries - but I am certainly glad that my 'surprise' is perhaps not all that justiified.
Kind Regards,
John
At 15:26 26/04/2021, Abhaya Indrayan wrote:
John:
I do not want to defend what Martin Holt did regarding his book in his signature but that onus is not on me although I am a co-author. That was not in the body of the post. If this kind of post is allowed in this group, absolutely fine with me. I just wanted to know.
Regarding covid in India, we are indeed struggling but I do not think the onslaught is going to be as bad as in some other countries per million population despite our limited resources and illiterate population. Help is coming but that is nothing relative to our population and we have to depend on our own capacities. The surge was sudden and steep and nobody was prepared.
Warm regards.
~Abhaya
--
Dr Abhaya Indrayan
Personal website:Â http://indrayan.weebly.com
On Mon, Apr 26, 2021 at 6:31 PM John Whittington <Joh...@mediscience.co.uk > wrote:
- At 04:31 26/04/2021, A Indrayan wrote:
- Is this allowed for this group?
- Abhaya, I would be happy to consider alternative views, if anyone has them, but I decided that it was probably OK to allow this. Although in some senses 'advertisements' (since books cost money), I think that announcements relating to newly published books (in relation to subjects relevant to the group) are probably acceptable in groups like this.
- One of the things which swayed me in my decision is the fact that, for at least couple of years, Martin Holt (the owner of this group/list) included in the signature of every one of his posts to the group an 'announcement' relating to a book (costing some 8 times the price of Don's recently announced one) which he co-authored with yourself. I took this to mean that, as list owner, he considered this to be an acceptable practice in this group.
- However, as above, I would be interesting and pleased to hear of any other views (in any direction) about this.
- I'm very sorry to hear what is currently happening in India as regards the pandemic. I hope that you and yours are managing to keep well, and certainly hope that something can be done soon to get things at least vaguely 'under control' in your country - although it is clearly going to be an uphill struggle, even with the international help and support which I imagine (hope) you will be getting.
- Kind Regards,
- John
- On 26-Apr-2021, at 5:50 AM, Don Macnaughton <don...@matstat.com> wrote:
- I favor the continued use of statistical significance (e.g., p-values with thresholds) as a necessary (though not sufficient) gateway to the publication of empirical-research papers in scientific journals. This gateway was invented by Karl Pearson, Ronald Fisher, Jerzy Neyman, and Egon Pearson between 1900 and 1935. These statisticians focused on the needs of scientific researchers, though the modern focus of the gateway is on the needs of society.
- The threshold-p-value gateway to publication is sensible because it helps journal editors and researchers maximize the overall scientific and social benefit of scientific research. The gateway does this by producing a roughly optimal trade-off between the rates of costly false-positive and costly false-negative errors that a journal makes in accepting or rejecting papers for consideration for publication.
- I have recently published a book that explains the logic behind the preceding ideas. The book's title is
- The War on Statistical Significance:
- The American Statistician vs. the New England Journal of Medicine
- The book is aimed at scientific researchers, journal editors, science teachers, and science students. It will also be of interest to some statisticians for its carefully reasoned arguments about the fundamental practical role of the field of statistics in scientific research.
- More information is available at war-stat-sig.com.
- Donald B. Macnaughton
- John
- ----------------------------------------------------------------
- Dr John Whittington,      Voice:   +44 (0) 1296 730225
- Mediscience Services      Fax:     +44 (0) 1296 738893
- Twyford Manor, Twyford,   E-mail:  Joh...@mediscience.co.uk
- Buckingham MK18 4EL, UK           Â
- ----------------------------------------------------------------
- --
- --
- To post a new thread to MedStats, send email to MedS...@googlegroups.com .
- MedStats' home page is http://groups.google.com/group/MedStats .
- Rules: http://groups.google.com/group/MedStats/web/medstats-rules
- ---
- You received this message because you are subscribed to the Google Groups "MedStats" group.
- To unsubscribe from this group and stop receiving emails from it, send an email to medstats+u...@googlegroups.com.
- To view this discussion on the web, visit https://groups.google.com/d/msgid/medstats/202104261301.13QD1aYL029678%40mail102c50.megamailservers.eu .
--
--
To post a new thread to MedStats, send email to MedS...@googlegroups.com .
MedStats' home page is http://groups.google.com/group/MedStats .
Rules: http://groups.google.com/group/MedStats/web/medstats-rules
---
You received this message because you are subscribed to the Google Groups "MedStats" group.
To unsubscribe from this group and stop receiving emails from it, send an email to medstats+u...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/medstats/CAP7G4a7E6ZWVm8_WJrxy0-6CWJEg59V8yyKax2esH8kCYjf8tw%40mail.gmail.com .
Reply all
Reply to author
Forward
0 new messages