303:(Gaussian) has variance as a parameter, any data with finite variance (including any finite data) can be modeled with a normal distribution with the exact variance – the normal distribution is a two-parameter model, with mean and variance. Thus, in the absence of an underlying model, there is no notion of data being overdispersed relative to the normal model, though the fit may be poor in other respects (such as the higher moments of
36:
371:, however, meanings have been transposed, so that overdispersion is actually taken to mean more even (lower variance) than expected. This confusion has caused some ecologists to suggest that the terms 'aggregated', or 'contagious', would be better used in ecology for 'overdispersed'. Such preferences are creeping into
252:
distribution is a popular and analytically tractable alternative model to the binomial distribution since it provides a better fit to the observed data. To capture the heterogeneity of the families, one can think of the probability parameter of the binomial model (say, probability of being a boy) is
244:
for one possible explanation) i.e. there are more all-boy families, more all-girl families and not enough families close to the population 51:49 boy-to-girl mean ratio than expected from a binomial distribution, and the resulting empirical variance is larger than specified by a binomial model.
208:. The Poisson distribution has one free parameter and does not allow for the variance to be adjusted independently of the mean. The choice of a distribution from the Poisson family is often dictated by the nature of the empirical data. For example,
330:
of repeated surveys of a fixed population (say with a given sample size, so margin of error is the same), one expects the results to fall on normal distribution with standard deviation equal to the margin of error. However, in the presence of
347:
all with a margin of error of 3%, if they are conducted by different polling organizations, one expects the results to have standard deviation greater than 3%, due to pollster bias from different methodologies.
187:
means that there was less variation in the data than predicted. Overdispersion is a very common feature in applied data analysis because in practice, populations are frequently
311:, etc.). However, in the case that the data is modeled by a normal distribution with an expected variation, it can be over- or under-dispersed relative to that prediction.
291:
With respect to binomial random variables, the concept of overdispersion makes sense only if n>1 (i.e. overdispersion is nonsensical for
Bernoulli random variables).
581:
216:. If overdispersion is a feature, an alternative model with additional free parameters may provide a better fit. In the case of count data, a Poisson
224:
can be proposed instead, in which the mean of the
Poisson distribution can itself be thought of as a random variable drawn – in this case – from the
228:
thereby introducing an additional free parameter (note the resulting negative binomial distribution is completely characterized by two parameters).
171:. However, especially for simple models with few parameters, theoretical predictions may not match empirical observations for higher
559:
532:
400:
340:
217:
119:
53:
241:
364:, the term 'overdispersion' is generally used as defined here – meaning a distribution with a higher than expected variance.
221:
100:
57:
236:
As a more concrete example, it has been observed that the number of boys born to families does not conform faithfully to a
72:
240:
as might be expected. Instead, the sex ratios of families seem to skew toward either boys or girls (see, for example the
79:
46:
591:
284:. In this case, if the variance of the normal variable is zero, the model reduces to the standard (undispersed)
86:
586:
383:
163:
of the chosen model. It is usually possible to choose the model parameters in such a way that the theoretical
204:
Overdispersion is often encountered when fitting very simple parametric models, such as those based on the
68:
269:
141:
382:, overdispersion is often evident in the analysis of death count data, but demographers prefer the term '
262:
237:
254:
205:
172:
395:
332:
300:
285:
277:
273:
249:
503:
357:
315:
225:
209:
343:
and will be overdistributed relative to the predicted distribution. For example, given repeated
191:(non-uniform) contrary to the assumptions implicit within widely used simple parametric models.
555:
528:
495:
454:
446:
375:
too. Generally this suggestion has not been heeded, and confusion persists in the literature.
258:
145:
485:
438:
405:
281:
156:
93:
288:. This model has an additional free parameter, namely the variance of the normal variable.
319:
164:
160:
323:
575:
336:
327:
188:
159:
to fit a given set of empirical observations. This necessitates an assessment of the
507:
372:
361:
344:
522:
168:
35:
356:
Over- and underdispersion are terms which have been adopted in branches of the
17:
549:
379:
268:
Another common model for overdispersion—when some of the observations are not
213:
152:
133:
450:
490:
473:
426:
499:
458:
442:
308:
304:
176:
368:
326:
and hence dispersion of results on repeated surveys. If one performs a
27:
Presence of greater variability in a data set than would be expected
474:"Analysis of the Human Sex Ratio by using Overdispersion Models"
29:
427:"The most widely publicized gender problem in human genetics"
425:
Stansfield, William D.; Carlton, Matthew A. (February 2009).
280:. Software is widely available for fitting this type of
144:) in a data set than would be expected based on a given
179:
is higher than the variance of a theoretical model,
60:. Unsourced material may be challenged and removed.
527:(Third ed.). University of California Press.
478:Journal of the Royal Statistical Society, Series C
265:(beta-binomial) has an additional free parameter.
8:
352:Differences in terminology among disciplines
167:of the model is approximately equal to the
261:as the mixing distribution. The resulting
489:
472:Lindsey, J. K.; Altham, P. M. E. (1998).
322:(determined by sample size) predicts the
120:Learn how and when to remove this message
140:is the presence of greater variability (
417:
7:
58:adding citations to reliable sources
212:analysis is commonly used to model
25:
551:Evolutionary Ecology of Parasites
401:Compound probability distribution
582:Probability distribution fitting
339:, the distribution is instead a
34:
257:) drawn for each family from a
253:itself a random variable (i.e.
45:needs additional citations for
554:. Princeton University Press.
222:negative binomial distribution
1:
335:where studies have different
272:—arises from introducing a
608:
524:Quantitative Plant Ecology
242:Trivers–Willard hypothesis
183:has occurred. Conversely,
151:A common task in applied
521:Greig-Smith, P. (1983).
384:unobserved heterogeneity
491:10.1111/1467-9876.00103
274:normal random variable
142:statistical dispersion
341:compound distribution
263:compound distribution
238:binomial distribution
443:10.3378/027.081.0101
255:random effects model
206:Poisson distribution
175:. When the observed
54:improve this article
548:Poulin, R. (2006).
396:Index of dispersion
358:biological sciences
333:study heterogeneity
301:normal distribution
295:Normal distribution
286:logistic regression
250:beta-binomial model
316:statistical survey
314:For example, in a
248:In this case, the
226:gamma distribution
210:Poisson regression
367:In some areas of
259:beta distribution
146:statistical model
130:
129:
122:
104:
16:(Redirected from
599:
592:Spatial analysis
566:
565:
545:
539:
538:
518:
512:
511:
493:
469:
463:
462:
422:
406:Quasi-likelihood
282:multilevel model
157:parametric model
125:
118:
114:
111:
105:
103:
69:"Overdispersion"
62:
38:
30:
21:
607:
606:
602:
601:
600:
598:
597:
596:
587:Point processes
572:
571:
570:
569:
562:
547:
546:
542:
535:
520:
519:
515:
471:
470:
466:
424:
423:
419:
414:
392:
378:Furthermore in
354:
320:margin of error
297:
234:
202:
197:
185:underdispersion
165:population mean
126:
115:
109:
106:
63:
61:
51:
39:
28:
23:
22:
18:Underdispersion
15:
12:
11:
5:
605:
603:
595:
594:
589:
584:
574:
573:
568:
567:
560:
540:
533:
513:
484:(1): 149–157.
464:
416:
415:
413:
410:
409:
408:
403:
398:
391:
388:
353:
350:
324:sampling error
296:
293:
278:logistic model
233:
230:
201:
198:
196:
193:
181:overdispersion
155:is choosing a
138:overdispersion
128:
127:
42:
40:
33:
26:
24:
14:
13:
10:
9:
6:
4:
3:
2:
604:
593:
590:
588:
585:
583:
580:
579:
577:
563:
561:9780691120850
557:
553:
552:
544:
541:
536:
534:0-632-00142-9
530:
526:
525:
517:
514:
509:
505:
501:
497:
492:
487:
483:
479:
475:
468:
465:
460:
456:
452:
448:
444:
440:
436:
432:
431:Human Biology
428:
421:
418:
411:
407:
404:
402:
399:
397:
394:
393:
389:
387:
385:
381:
376:
374:
370:
365:
363:
359:
351:
349:
346:
345:opinion polls
342:
338:
337:sampling bias
334:
329:
328:meta-analysis
325:
321:
317:
312:
310:
306:
302:
294:
292:
289:
287:
283:
279:
275:
271:
266:
264:
260:
256:
251:
246:
243:
239:
231:
229:
227:
223:
219:
218:mixture model
215:
211:
207:
199:
194:
192:
190:
189:heterogeneous
186:
182:
178:
174:
170:
166:
162:
158:
154:
149:
147:
143:
139:
135:
124:
121:
113:
102:
99:
95:
92:
88:
85:
81:
78:
74:
71: –
70:
66:
65:Find sources:
59:
55:
49:
48:
43:This article
41:
37:
32:
31:
19:
550:
543:
523:
516:
481:
477:
467:
434:
430:
420:
377:
373:parasitology
366:
362:parasitology
355:
313:
298:
290:
267:
247:
235:
203:
184:
180:
150:
137:
131:
116:
110:January 2008
107:
97:
90:
83:
76:
64:
52:Please help
47:verification
44:
437:(1): 3–11.
169:sample mean
576:Categories
412:References
380:demography
214:count data
153:statistics
134:statistics
80:newspapers
451:1534-6617
270:Bernoulli
220:like the
508:22354905
500:12293397
459:19589015
390:See also
309:kurtosis
232:Binomial
195:Examples
177:variance
369:ecology
299:As the
276:into a
200:Poisson
173:moments
94:scholar
558:
531:
506:
498:
457:
449:
318:, the
96:
89:
82:
75:
67:
504:S2CID
360:. In
101:JSTOR
87:books
556:ISBN
529:ISBN
496:PMID
455:PMID
447:ISSN
305:skew
73:news
486:doi
439:doi
386:'.
161:fit
132:In
56:by
578::
502:.
494:.
482:47
480:.
476:.
453:.
445:.
435:81
433:.
429:.
307:,
148:.
136:,
564:.
537:.
510:.
488::
461:.
441::
123:)
117:(
112:)
108:(
98:·
91:·
84:·
77:·
50:.
20:)
Text is available under the Creative Commons Attribution-ShareAlike License. Additional terms may apply.
