phenomenon · claim-type · cross-domain-transfer · mechanism · claim-relation · replication-status · limitation — grounded in the full text| phenomenon | power-law distributions in empirical data— “Power-law distributions occur in many situations of scientific interest and have significant consequences for our understanding of natural and man-made phenomena.” | |
| claim-type | method— “Here we present a principled statistical framework for discerning and quantifying power-law behavior in empirical data.” | |
| method — evidence | — “We evaluate the effectiveness of the approach with tests on synthetic data and give critical comparisons to previous approaches.” | |
| transfer | from twenty-four fields at once (physics, earth sciences, biology, ecology, paleontology, computer and information science): the fitting procedure is deliberately applied across twenty-four data sets from unrelated branches of enquiry, so that the method rather than any one field's data is what is tested— “The twenty-four data sets we study are drawn from a broad variety of different branches of human endeavor, including physics, earth sciences, biology, ecology, paleontology, computer and information sciences, engineering, and the social sciences.” | transfers[0] |
| mechanism | the lower bound of the power-law range is chosen by minimising the Kolmogorov-Smirnov distance between the empirical distribution and the best-fit model above that bound - too high a bound shrinks the data and too low a one includes non-power-law points, so the minimum lies between— “The fundamental idea behind this method is simple: we choose the value of ˆxmin that makes the probability distributions of the measured data and the best-fit power-law model as similar as possible above ˆxmin. In general, if we choose ˆxmin higher than the true value xmin, then we are effectively reducing the size of our data set, which will… | mechanisms[0] |
| mechanism | each synthetic data set is scored against ITS OWN best-fit power law rather than against the distribution it was drawn from, which is what makes the resulting p-value unbiased: the synthetic sets undergo exactly the calculation the real data underwent, fitting included— “Note crucially that for each synthetic data set we compute the KS statistic relative to the best-fit power law for that data set, not relative to the original distribution from which the data set was drawn. In this way we ensure that we are performing for each synthetic data set the same calculation that we performed for the real data set, a… | mechanisms[1] |
| mechanism | synthetic data are generated semiparametrically - drawn from the fitted power law with probability equal to the observed tail fraction, and otherwise resampled from the observed points below the cut-off - so the synthetic sets share the empirical body while following the power law in the tail— “To obtain accurate estimates of p we need synthetic data that have a distribution similar to the empirical data below xmin but that follow the fitted power law above xmin. To generate such data we make use of a semiparametric approach. Suppose that our observed data set has ntail observations x xmin and n observations in total. We generate a… | mechanisms[2] |
| mechanism | the test distribution is constructed to be hard rather than convenient: it is exponential below the cut-off and a power law above, and its slope is CONTINUOUS at the cut-off, so it deviates only gently from the power law below that point - the authors' word is challenging— “This distribution follows a power law at xmin and above but an exponential below. Furthermore, it has a continuous slope at xmin and thus deviates only gently from the power law as we pass below this point, making for a challenging test.” | mechanisms[3] |
| mechanism | uncertainty on the fitted parameters is obtained by a nonparametric bootstrap - drawing a fresh sequence of n points uniformly at random from the original data and re-estimating both the lower bound and the scaling parameter on each surrogate set— “One way to do this is to make use of a nonparametric “bootstrap” method [16]. Given our n measurements, we generate a synthetic data set with a similar distribution to the original by drawing a new sequence of points xi, i = 1 . . . n uniformly at random from the original data. Using either method described above, we then estimate xmin and α… | mechanisms[4] |
| mechanism | the alternative goodness-of-fit statistics are tested rather than assumed: reweighted KS and Kuiper behave like plain KS, while Anderson-Darling returns cut-offs an order of magnitude too large, and the cost of that conservatism is stated as fewer tail observations, larger error on the scaling parameter and less ability to validate the model— “We have performed tests with each of these alternative statistics and find that results for the reweighted KS and Kuiper statistics are very similar to those for the standard KS statistic. The Anderson-Darling statistic, on the other hand, we find to be highly conservative in this application, giving estimates ˆxmin that are too large by an… | mechanisms[5] |
| mechanism | the number of synthetic data sets is fixed by the precision wanted in the p-value rather than by convention, derived from the test's expected worst-case performance— “We also need to decide how many synthetic data sets to generate. Based on an analysis of the expected worst-case performance of the test, a good rule of thumb turns out to be the following: if we wish our p-values to be accurate to within about” | mechanisms[6] |
| mechanism | the direction of the p-value is inverted relative to its ordinary use: the quantity being tested is the model the authors want to assess rather than a null, so a HIGH p-value is the favourable one and a low value rules the power law out— “Here, by contrast, we use the p-value as a measure of the hypothesis we are trying to verify, and hence high values, not low, are “good.”” | mechanisms[7] |
| contradicts | previously claimed power laws in the literature— “In some cases we find these conjectures to be consistent with the data while in others the power law is ruled out.” | claim_relations[0] |
| contradicts | previously reported scaling parameters for the citation distribution of [48], namely 2.9 in [62] and 2.5 in [33]— “Similarly, the citation distribution data of [48] have been reported to have a scaling parameter of either 2.9 [62] or 2.5 [33], neither of which are compatible with our maximum likelihood figure of 3.16 0.06.” | claim_relations[1] |
| builds-on | Ijiri and Simon [28] and Mitzenmacher [42] on the limits of characterising empirical distributions— “In closing, we echo comments made by Ijiri and Simon [28] more than thirty years ago and similar thoughts expressed more recently by Mitzenmacher [42].” | claim_relations[2] |
| sense | power law - a hypothesis one can only fail to reject; the paper says an observed quantity can rarely if ever be certified as power-law distributed— “In practice, we can rarely, if ever, be certain that an observed quantity is drawn from a power-law distribution. The most we can say is that our observations are consistent with the hypothesis that x is drawn from a distribution of the form” | senses[0] |
| sense | power law - a property of the tail only, holding above a chosen minimum rather than across the whole range— “In practice, few empirical phenomena obey power laws for all values of x. More often the power law applies only for values greater than some minimum xmin. In such cases we say that the tail of the distribution follows a power law.” | senses[1] |
| sense | scaling parameter - a quantity whose published values for one data set disagree, because they depend on the estimator used— “the citation distribution data of [48] have been reported to have a scaling parameter of either 2.9 [62] or 2.5 [33], neither of which are compatible with our maximum likelihood figure of 3.16” | senses[2] |
| control | a goodness-of-fit p-value computed for the power law AND for each plausible competing distribution, and the two compared — held — rules out that the power law fits because heavy-tailed data fits almost any heavy-tailed form - a log-normal or a stretched exponential may fit as well or better, and only comparing them shows which— “We can eliminate this possibility by using a goodness-of-fit test again—we can simply calculate a p-value for a fit to the competing distribution and compare it to the p-value for the power law.” | controls[0] |
| control | an explicit rejection rule fixed in advance: p greater than 0.1 keeps the power-law hypothesis, otherwise it is rejected — held — rules out that the decision to accept a power law is made after seeing how good the fit looks— “If the resulting p-value is greater than 0.1 the power law is a plausible hypothesis for the data, otherwise it is rejected.” | controls[1] |
| control | the whole procedure was run on SYNTHETIC data drawn from known distributions — held — rules out that the method itself is wrong - synthetic data has a known answer, so the procedure can be checked against ground truth before it is trusted on real data— “We evaluate the effectiveness of the approach with tests on synthetic data and give critical comparisons to previous approaches.” | controls[2] |
| audit | power law —contrasted-with→ log-normal — Audit 2026-09-28: DOUBTFUL - read against its full span and kept, but the p-value supports a classification as power-law, log-normal or undecided; contrasted-with is loose.— “If we take the p-value into account, however, using its value to perform a more nuanced classification as power-law, log-normal, or undecided” | concept_edges[0] |
| audit | scaling parameter —operationalised-as→ maximum likelihood — Audit 2026-09-28: DOUBTFUL - read against its full span and kept, but the span compares reported scaling parameters with 'our maximum likelihood figure'; operationalised-as is loose.— “the citation distribution data of [48] have been reported to have a scaling parameter of either 2.9 [62] or 2.5 [33], neither of which are compatible with our maximum likelihood figure of 3.16” | concept_edges[1] |
| edge | xmin —part-of→ power law— “More often the power law applies only for values greater than some minimum xmin.” | concept_edges[2] |
| outcome (stated by the tagger, no quotation) | control — held | controls[0] |
| outcome (stated by the tagger, no quotation) | control — held | controls[1] |
| outcome (stated by the tagger, no quotation) | control — held | controls[2] |
| for which use | limitation — every power-law claim anywhere: THE MOST WE CAN SAY IS THAT OUR OBSERVATIONS ARE CONSISTENT WITH THE HYPOTHESIS - the hypothesis cannot be confirmed— “The most we can say is that our observations are consistent with the hypothesis” | limitations[0] |
| for which use | limitation — the range over which any fitted power law holds: FEW EMPIRICAL PHENOMENA OBEY POWER LAWS FOR ALL VALUES OF x— “few empirical phenomena obey power laws for all values of x” | limitations[1] |
| for which use | limitation — the reanalyses in this paper: WE HAVE NOT ATTEMPTED TO CORRECT ANY BIASES IN OUR ANALYSIS— “We have not attempted to correct any biases in our analysis.” | limitations[2] |
| for which use | limitation — the datasets reanalysed: MANY OF THESE DATA SETS ARE ONLY SUBSETS OF MUCH LARGER ENTITIES (SUCH AS THE WEB SITES, WHICH ARE ONLY A SMALL FRACTION OF THE ENTIRE WEB)— “Many of these data sets are only subsets of much larger entities (such as the web sites, which are only a small fraction of the entire web).” | limitations[3] |
| for which use | limitation — the applicability of the framework: IT COULD ALSO BE EXTENDED TO OTHER, NON-POWER-LAW DISTRIBUTIONS ... ALTHOUGH WE HAVE NOT GIVEN SUCH AN EXTENSION HERE— “In principle, they could also be extended to other, non-power-law distributions as well, although we have not given such an extension here.” | limitations[4] |
| for which use | limitation — any claim that a power law rather than a log-normal has been found: IT IS EXTREMELY DIFFICULT TO TELL THE DIFFERENCE BETWEEN LOG-NORMAL AND POWER-LAW BEHAVIOR— “In general, we find that it is extremely difficult to tell the difference between log-normal and power-law behavior.” | limitations[5] |
| for which use | limitation — extrapolating any of these fits to large rare events: the paper states conclusions from different fitted forms can differ enormously while being indistinguishable on the data— “if we wish to extrapolate a fitted distribution far into its tail, to predict, for example, the frequencies of large but rare events like major earthquakes or meteor impacts, then conclusions based on different fitted forms can differ enormously even if the forms are indistinguishable in the domain covered by the actual data.” | limitations[6] |
| for which use | limitation — any of the seven named data sets as power-law distributed: HTTP connections, earthquakes, web links, fires, wealth, web hits and the metabolic network are ruled out here— “The p-values in Table 6.1 indicate that 17 of the 24 data sets are consistent with a power-law distribution. The remaining seven data sets all have p-values small enough that the power-law model can be firmly ruled out.” | limitations[7] |
| for which use | limitation — a published scaling exponent for a data set: for these citation data three different values are in print, and the paper says the earlier two are incompatible with the maximum-likelihood estimate— “the citation distribution data of [48] have been reported to have a scaling parameter of either 2.9 [62] or 2.5 [33], neither of which are compatible with our maximum likelihood figure of 3.16” | limitations[8] |
| for which use | limitation — the power-law verdicts in this paper: they follow a p-value threshold of 0.1, and cases near it are decided by that choice— “If the resulting p-value is greater than 0.1 the power law is a plausible hypothesis for the data, otherwise it is rejected.” | limitations[9] |
| for which use | limitation — the sample size behind any of these fits: xmin is estimated, and it determines how many points enter the test— “since we fit the power-law form to only the part of the distribution above xmin, the value of xmin effectively controls how many data points we have to work with.” | limitations[10] |
| for which use | limitation — a power-law fit on a small sample in either direction: the paper states both false acceptance and failure to discriminate occur there— “it is difficult to rule out alternative fits to such data, even when they are truly power-law distributed, and conversely the power-law form may appear to be a good fit even when the data are drawn from a non-power-law distribution.” | limitations[11] |
| evidence status note | mechanism — 2026-08-16: THIS QUOTE IS TRUNCATED BECAUSE THE FORMULA IT LEADS TO IS UNREADABLE IN THE STAGED TEXT. The pdftext extraction flattens the expression for the required number of synthetic data sets out of order - the staged body reads 'at least 1 2 synthetic data sets' with '4 e-' appearing separately - so the rule the sentence states cannot be quoted verbatim without quoting nonsense. The span stops at the last readable words. The paper's own worked example (accuracy to two decimal digits requiring about 2500 sets) is likewise in the damaged region and is deliberately NOT reproduced here, because I would be typing it rather than cutting it.— “We also need to decide how many synthetic data sets to generate. Based on an analysis of the expected worst-case performance of the test, a good rule of thumb turns out to be the following: if we wish our p-values to be accurate to within about” | mechanisms[6] |
| replication | untested | |
| limitation | one can never confirm a power law, only find data consistent with it— “The most we can say is that our observations are consistent with the hypothesis” | limitations[0] |
| limitation | few empirical phenomena obey power laws across the whole range of x— “few empirical phenomena obey power laws for all values of x” | limitations[1] |
| limitation | Known sampling biases in several of the reanalysed data sets were left uncorrected.— “We have not attempted to correct any biases in our analysis.” | limitations[2] |
| limitation | Many of the reanalysed data sets are partial samples of larger populations.— “Many of these data sets are only subsets of much larger entities (such as the web sites, which are only a small fraction of the entire web).” | limitations[3] |
| limitation | The framework is presented only for the power law; no extension to other distributional forms is given.— “In principle, they could also be extended to other, non-power-law distributions as well, although we have not given such an extension here.” | limitations[4] |
| limitation | The procedure often cannot separate power-law from log-normal behaviour.— “In general, we find that it is extremely difficult to tell the difference between log-normal and power-law behavior.” | limitations[5] |
| limitation | Two model forms indistinguishable on the observed data can give enormously different extrapolations into the tail, which is where such fits are usually applied.— “if we wish to extrapolate a fitted distribution far into its tail, to predict, for example, the frequencies of large but rare events like major earthquakes or meteor impacts, then conclusions based on different fitted forms can differ enormously even if the forms are indistinguishable in the domain covered by the actual data.” | limitations[6] |
| limitation | Seven of the twenty-four data sets previously conjectured to be power laws are firmly ruled out.— “The p-values in Table 6.1 indicate that 17 of the 24 data sets are consistent with a power-law distribution. The remaining seven data sets all have p-values small enough that the power-law model can be firmly ruled out.” | limitations[7] |
| limitation | Two published scaling parameters for the same citation data disagree with each other and with this paper's maximum-likelihood estimate.— “the citation distribution data of [48] have been reported to have a scaling parameter of either 2.9 [62] or 2.5 [33], neither of which are compatible with our maximum likelihood figure of 3.16” | limitations[8] |
| limitation | The accept/reject verdict rests on a stated p-value threshold of 0.1.— “If the resulting p-value is greater than 0.1 the power law is a plausible hypothesis for the data, otherwise it is rejected.” | limitations[9] |
| limitation | The range fitted is chosen, and that choice controls how much data the test sees.— “since we fit the power-law form to only the part of the distribution above xmin, the value of xmin effectively controls how many data points we have to work with.” | limitations[10] |
| limitation | Errors run in both directions on small samples: a power law can look like a good fit to non-power-law data, and true power laws cannot be separated from alternatives.— “it is difficult to rule out alternative fits to such data, even when they are truly power-law distributed, and conversely the power-law form may appear to be a good fit even when the data are drawn from a non-power-law distribution.” | limitations[11] |
finding · quantity · variable · definition · key-term — grounded in the full text| finding | least-squares fitting yields substantially inaccurate power-law parameters— “such as least-squares fitting, can produce substantially inaccurate estimates of parameters for power-law distributions” | content_tags.findings[0] |
| finding | their framework combines maximum likelihood with KS goodness-of-fit— “Our approach combines maximum-likelihood fitting methods with goodness-of-fit tests based on the Kolmogorov-Smirnov statistic and likelihood ratios.” | content_tags.findings[1] |
| finding | some conjectured power laws are ruled out by the data— “others the power law is ruled out” | content_tags.findings[2] |
| finding | power laws usually hold only above a minimum value xmin— “the power law applies only for values greater than some minimum xmin” | content_tags.findings[3] |
| quantity | 24 data sets — real-world data sets tested for power-law fit— “We also apply the proposed methods to twenty-four real-world” | content_tags.quantities[0] |
| quantity | 2 < α < 3 range — typical range of the scaling parameter— “The scaling parameter typically lies in the range 2 < α < 3” | content_tags.quantities[1] |
| quantity | 8226 people — average US city/town/village population (2000 Census)— “village in the United States is 8226” | content_tags.quantities[2] |
| quantity | 180 cm — typical adult male American height (near-Gaussian example)— “an adult male American is about 180cm” | content_tags.quantities[3] |
| quantity | 2 factor from mean — largest deviations in a well-behaved distribution— “only about a factor of two from the mean” | content_tags.quantities[4] |
| variable | scaling parameter (characterizes the power law)— “a constant parameter of the distribution known as the exponent or scaling parameter” | content_tags.variables[0] |
| variable | lower bound xmin (value above which the power-law tail begins)— “few empirical phenomena obey power laws for all values of x” | content_tags.variables[1] |
| variable | goodness-of-fit statistic (tests whether data follow a power law)— “goodness-of-fit tests based on the Kolmogorov-Smirnov statistic” | content_tags.variables[2] |
| variable | quantity x (observable tested for power-law behavior)— “a quantity x obeys a power law if it is drawn from a probability” | content_tags.variables[3] |
| definition | power law— “Mathematically, a quantity x obeys a power law if it is drawn from a probability” | content_tags.definitions[0] |
| definition | scaling parameter— “known as the exponent or scaling parameter” | content_tags.definitions[1] |
| definition | tail of the distribution— “we say that the tail of the distribution follows a power law” | content_tags.definitions[2] |
| key-terms | power law, power-law distribution, maximum likelihood, scaling parameter, exponent, goodness-of-fit, Kolmogorov-Smirnov, heavy-tailed, tail, empirical data |
| concept | power law— “we address a recurring issue in the scientific literature, the question of how to recognize a power law when we see one” | concepts[0] |
| concept | maximum likelihood— “Our approach combines maximum-likelihood fitting methods with goodness-of-fit tests based on the Kolmogorov-Smirnov statistic and likelihood ratios.” | concepts[1] |
| concept | scaling parameter— “where α is a constant parameter of the distribution known as the exponent or scaling parameter.” | concepts[5] |
| concept | xmin | concepts[6] |
| concept | likelihood ratio test | concepts[7] |
| concept | log-normal | concepts[8] |
| Bayesian | content_tags.statistics[0] | |
| confidence-interval CI | content_tags.statistics[1] | |
| hypothesis-testing | as written: hypothesis test | content_tags.statistics[3] |
| mean μ | as written: Mean of | content_tags.statistics[4] |
| null-hypothesis H0 | content_tags.statistics[5] | |
| p-value p | content_tags.statistics[6] | |
| regression β | as written: linear regression | content_tags.statistics[7] |
| sample-size n/N | Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | content_tags.statistics[8] |
| evidence status note | sample-size — QUOTE REPAIRED: the extractor split a hyphen-space and the quote was recorded faithfully with the damage in it, which made it unanchorable. The repaired string is an EXACT SUBSTRING of the staged text - that is why it was accepted, not because a rule produced it. The text as originally recorded is kept in evidence as extracted. | content_tags.statistics[8] |
| standard-deviation SD/σ | Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | content_tags.statistics[9] |
| standard-error SE | Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | content_tags.statistics[10] |
| evidence status note | standard-error — QUOTE RECOVERED FROM THE PAPER'S OWN CHARACTERS. The recorded quote was LETTERS-ONLY IDENTICAL to this stretch of the staged text but not an exact substring of it - spaces, hyphens, ligature glyphs or mathematical operators differed, so it could never anchor. This is the file's own text between the same two letters: not produced by a rule about the damage, copied out of the file. Round-trip checked (the recovered span is letters-only equal to the original) and length-checked. The text as originally recorded is kept in evidence as extracted. | content_tags.statistics[10] |
| statistical-significance | as written: statistically significant Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | content_tags.statistics[11] |
| variance σ² | Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | content_tags.statistics[12] |
When we have changed our mind. Some entries in the table begin with a word in capitals and the reason for it. EVIDENCE CORRECTED — the finding still stands, but we had quoted the wrong sentence from the paper, and the quotation has been replaced with the right one. RETRACTED — the entry itself was wrong; it stays on the page, marked, with the reason, rather than being deleted. WITHDRAWN AND REPLACED — the entry was wrong, and a corrected version of it appears in the entries below. Each mark carries its own reason, in brackets, because the reasons differ from one entry to the next. An entry we have WITHDRAWN is removed from this page and kept in the record, with its reason. A line reading Audit and a date is a later re-reading of the entry against its whole sentence: DOUBTFUL means it was kept, and the reason says what is loose about it. Quotation marks inside the tagger's sentence. Where a tag is the tagger's own sentence, a phrase in quotation marks inside it is borrowed from the paper — the tagger's wording around it is not. The paper's own sentence is the evidence quote for that row; on this page it is withheld with the article, and the article is at its publisher. Nothing inside quotation marks anywhere on this page is ours: if it is quoted, it is the paper's. Whose words a concept carries. A concept entry records a term the paper itself uses, so the term, the sense given for it, and its evidence are the paper's own words, quoted where the paper fixes them. This is the opposite of a sense entry in the tag table, which is the tagger's sentence about how the paper uses a term. Where a concept entry could not be grounded in the paper's words, the entry says so on its own row.
Tag layer © the Meta-Tagging Project, CC BY 4.0. The paper text is © its authors/publisher
and is linked, not reproduced. Excerpts shown are short quotations for identification and verification.
· back to index
· concepts · authors
· עברית
© 2026 Shir Sivroni · ORCID 0009-0000-2597-2824