Absence of evidence IS evidence of absence

quote:

---

A problem may be that you are trying to be black and white. For example, in the football analogy. No one would conclude there is no football after looking at one blade of grass. But after some number of random searches (short of the whole field) more and more of us will conclude there is no football.

---

Two questions in response:1) What percent of possible universes have we examined?
2) How much of this universe have we examined?In short, the case that /*Pretentious*/ is arguing for is looking at a blade of grass for the football, not the entire pitch. I would have absolutely no problem if they argued "There's not enough data gathered yet, so we can't come to a conclusion yet". However, he's arguing that there's enough data. Can you, in good conscience, take the answers to questions #1 and #2, and conclude that there's even remotely enough data?If a statistician tried to argue absence from such a miniscule percentage of total data, they'd be fired on the spot.------------------
"Illuminant light,
illuminate me."

quote:

---

Uhm...no I am not.

---

Then what is your argument that this universe is fine-tuned based on?------------------
"Illuminant light,
illuminate me."

quote:

---

Based on the supported (but not proven) assumption that life is restricted to LAWKI, the Universe is fine-tuned for life. For example, make the strong force a few percent stronger or weaker and the Universe is no longer compatible with life. Same goes for similar changes in the strengths of the weak force, the gravitational force, or the electromagnetic force. Even though the four forces range over approximately 40 orders of magnitude, a small (< 5%?) change in any one of them means no life.

​

I will also point out that this thread has left its original topic.

---

You assume your conclusion: "Based on the supported (but not proven) assumption that life is restricted to LAWKI". Another fundamental logical flaw. If life was restricted to LAWKI, of course life wouldn't exist in other universes. That wasn't what the debate was about, everyone accepted that. The debate was whether non-LAWKI can exist in other universes - for which the key point of contention is what the requisites for any sort of life are, and whether other universes readily would attain them.However, you're right, this may well be a side topic; I won't post on it any more if you don't think we should.------------------
"Illuminant light,
illuminate me."[This message has been edited by Rei, 11-06-2003]

Ok. You provided 4; I'll provide 8. Provide 8, I'll provide 16. See how this works? This is the exact reason why anecdotal - i.e., statistically insignificant - cases of anything are largely irrelevant except for spurring interest. That's the chief problem of your unverse analysis using this logic - we've only examined a tiny fraction of a single case.

quote:

---

Absence of evidence is not evidence of absence

​

Douglas G Altman, head,a J Martin Bland, reader in medical statistics

​

The non-equivalence of statistical significance and clinical importance has long been recognised, but this error of interpretation remains common. Although a significant result in a large study may sometimes not be clinically important, a far greater problem arises from misinterpretation of non-significant findings. By convention a P value greater than 5% (P>0.05) is called "not significant." Randomised controlled clinical trials that do not show a significant difference between the treatments being compared are often called "negative." This term wrongly implies that the study has shown that there is no difference, whereas usually all that has been shown is an absence of evidence of a difference. These are quite different statements.

​

...

​

When we are told that "there is no evidence that A causes B" we should first ask whether absence of evidence means simply that there is no information at all. If there are data we should look for quantification of the association rather than just a P value. Where risks are small P values may well mislead: confidence intervals are likely to be wide, indicating considerable uncertainty. While we can never prove the absence of a relation, when necessary we should seek evidence against the link between A and B--for example, from case-control studies. The importance of carrying out such studies will relate to the seriousness of the postulated effect and how widespread is the exposure in the population.

​

Statistics notes: Absence of evidence is not evidence of absence | The BMJ

---

(in reference to the faulty "absence of evidence is evidence of absence" logic that led to the spread of BSE)

quote:

---

Absence of evidence for your hypothesis does not mean that the opposite of the hypothesis is true, nor does it mean that the hypothesis must be false. Likewise, absence of evidence against your hypothesis does not mean that the hypothesis is true, or that the opposite of your hypothesis is false. This seems clear enough, but it happens often that people make mistakes this way.

​

http://www.sunspot.noao.edu/sunspot/pr/science-11.html

---

quote:

---

Professor Terje Traavik, the Scientific Director of the Norwegian Institute of Gene Ecology, Norway, will present Risks of genetic engineering: absence of evidence is not evidence of absence, in C1 at 7pm.

​

http://www.newsroom.canterbury.ac.nz/stories/03100801.htm

---

quote:

---

Absence of evidence is not evidence of absence is something every investigator should know, but too few do. Along with Interaction 2: compare effect sizes not P values, these articles describe two of the most common fatal mistakes in manuscripts submitted to research journals. The faulty reasoning leading to these errors is so seductive that papers containing these errors sometimes slip through the reviewing process and misinterpretations of data are published as fact.

​

Correlation, regression, and repeated data, Calculating correlation coefficients with repeated observations: Part 1--correlation within subjects, and Calculating correlation coefficients with repeated observations: Part 2--correlation between subjects provide an excellent introduction to the subtleties of analyzing repeated measurements on the same subject.

​

Page Not Found | Tufts University

---

quote:

---

disagree with the statement that "there is absolutely no evidence that vCJD can be spread through the blood supply." However, I accept Dr. Giulivi's questioning of my statement that "vCJD can be spread through the blood supply." Both are pronunciamentos and have little (or even negative) value. If Giulivi means that there are no reported cases of CJD or vCJD that have been transmitted through the blood supply, then he is right. But this lack of evidence does not mean that such transmission is not occurring. Giulivi is making an all-too- common error, since "the absence of evidence is not evidence of absence.

​

...

​

Worst-case estimates of the number of people in the UK infected with vCJD range up to 80 000 or about 1 in every 700 blood donors.6 So it is really important to know if vCJD can be spread through the blood supply. Because there is no screening test for vCJD in humans (or animals), our evidence has to come from animal experiments. Waiting for human epidemiologic evidence would be a mistake.

​

http://collection.nlc-bnc.ca/...maj/vol-158/issue-6/0715.htm

---

quote:

---

Guidance on COX 2 inhibitors suggests costings based partly upon a reduction in the use of gastroprotective agents. This is justified by a statement that there is no evidence of benefit from co-prescription of gastroprotective agents. A cardinal and often quoted rule in evidence based analysis is that absence of evidence is not evidence of absence. In light of the increasing numbers of cases of medicolegal litigation, many advisers find themselves unable to recommend that prescribers discontinue gastroprotection.

​

http://www.parliament.the-stationery-office.../...15ap30.htm

---

quote:

---

(concerning life explicitly):
'... or on any other planets, but our capability to collect such evidence is still poorly
developed; thus "absence of evidence is not evidence of absence." '

​

http://nai.arc.nasa.gov/astrobio/astrobio_questions.cfm?q...

---

quote:

---

"Absence of evidence is not evidence of absence"

​

- Carl Sagan, "Dragons of Eden"

---

I'll say this: Absence of evidence is occasionally evidence of absense, but only when a statistically significant percentage of possibilities have been analyzed. This is easy with a small sample size. Our sample size is not merely our entire universe, but all possible universes. It's ridiculous to make claims on "absense of evidence" as to the universe given how tiny a fraction of the sample size ( limit of X as X approaches 0% ) that we've been able to look at so far.P.S. - I'll preempt your next post, which would otherwise predictably re-setup the following straw man (you've done it before): Absence of evidence is not evidence of presence, either. With a small sample size, absence of evidence means merely "utterly inconclusive".------------------
"Illuminant light,
illuminate me."[This message has been edited by Rei, 11-06-2003]

DNA, when you keep on splitting a single topic into many topics, it becomes frustrating to attempt to bother to reply. You've done it so much in this thread that I'm not even going to bother attempting to piece it back together.Although, I must thank you for stopping the use of the annoying code brackets. ------------------
"Illuminant light,
illuminate me."

quote:

---

I disagree. According to your above reasoning, we could search the table, and the drawer, and the microwave, and the shelf, all without finding the keys, yet we would still have NO EVIDENCE WHATSOEVER that they werent in the kitchen. In fact, by your reasoning, we would have NO EVIDENCE WHATSOEVER that the keys werent in the kitchen until were performed a COMPLETE AND EXHAUSTIVE search and failed. But by that point wed have PROOF of absence in the kitchen. So you too are equating mere EVIDENCE with much stronger terms like PROOF or INDISPUTABLE FACT, you're just doing so implicitly. Thus, you are not addressing the position I put forward.

---

So, we've done a complete and exhaustive search of not only this universe, but all possible universes?I'll put it this way: two weeks ago, I lost my keys. I knew they had to be somewhere in my place, but I couldn't find them. I checked the kitchen table. I checked the ironing board, where I occasionally set stuff. I scoured my purse in and out. I checked the dresser. I checked all of the end tables in the living room. I pretty much gave up, and was going to get a new set. Should I have concluded that they were not in the place (I didn't just search a tiny fraction, but most of the apartment)?A week later, they turned up in a coat pocket.It was an early prediction of Conway's Game of Life that there would be no infinitely expansive patterns ("puffers"). None turned up during Conway's work. Should he have concluded that there were none?There were.The key here is the percentage of the possibilities that you look at. If your sample size is "all possible universes" or even just "this universe", the percentage you're going to look at is virtually zero; thus, an absence of evidence argument is effectively worthless.------------------
"Illuminant light,
illuminate me."

I agree completely, minnemooseus. It depends on the sample size; the basic logical fallacy exists when someone attempts to apply it to a situation in which only a miniscule percent of the total set has been examined.BTW, I like your avatar. ------------------
"Illuminant light,
illuminate me."