Simpson's paradox

Simpson's paradox is a phenomenon in probability and statistics in

which a trend appears in several groups of data but disappears or
reverses when the groups are combined. This result is often
encountered in social-science and medical-science statistics,[1][2][3]
and is particularly problematic when frequency data are unduly
given causal interpretations.[4] The paradox can be resolved when
confounding variables and causal relations are appropriately
addressed in the statistical modeling[4][5] (e.g., through cluster
analysis[6]). Simpson's paradox for quantitative
data: a positive trend ( , )
Simpson's paradox has been used to illustrate the kind of misleading    
appears for two separate groups,
results that the misuse of statistics can generate.[7][8] whereas a negative trend ( )
 
appears when the groups are
Edward H. Simpson first described this phenomenon in a technical combined.
paper in 1951,[9] but the statisticians Karl Pearson (in 1899[10]) and
Udny Yule (in 1903[11]) had mentioned similar effects earlier. The
name Simpson's paradox was introduced by Colin R. Blyth in
1972.[12] It is also referred to as Simpson's reversal, the Yule–
Simpson effect, the amalgamation paradox, or the reversal
paradox.[13]

Mathematician Jordan Ellenberg argues that Simpson's paradox is

misnamed as "there's no contradiction involved, just two different
ways to think about the same data" and suggests that its lesson "isn't
really to tell us which viewpoint to take but to insist that we keep Visualization of Simpson's paradox
both the parts and the whole in mind at once."[14] on data resembling real-world
variability indicates that risk of
Examples misjudgment of true causal
relationship can be hard to spot.

UC Berkeley gender bias

One of the best-known examples of Simpson's paradox comes from a study of gender bias among graduate
school admissions to University of California, Berkeley. The admission figures for the fall of 1973 showed
that men applying were more likely than women to be admitted, and the difference was so large that it was
unlikely to be due to chance.[15][16]

All Men Women

Applicants Admitted Applicants Admitted Applicants Admitted

Total 12,763 41% 8,442 44% 4,321 35%

However, when taking into account the information about departments being applied to, the different
rejection percentages reveal the different difficulty of getting into the department, and at the same time it
showed that women tended to apply to more competitive departments with lower rates of admission, even
among qualified applicants (such as in the English department), whereas men tended to apply to less
competitive departments with higher rates of admission (such as in the engineering department). The pooled
and corrected data showed a "small but statistically significant bias in favor of women".[16]

The data from the six largest departments are listed below:

All Men Women

Department
Applicants Admitted Applicants Admitted Applicants Admitted

A 933 64% 825 62% 108 82%

B 585 63% 560 63% 25 68%

C 918 35% 325 37% 593 34%

D 792 34% 417 33% 375 35%

E 584 25% 191 28% 393 24%

F 714 6% 373 6% 341 7%

Total 4526 39% 2691 45% 1835 30%

Legend:

   greater percentage of successful applicants than the other gender

   greater number of applicants than the other gender

bold - the two 'most applied for' departments for each gender

The entire data showed total of 4 out of 85 departments to be significantly biased against women, while 6
to be significantly biased against men (not all present in the 'six largest departments' table above). Notably,
the numbers of biased departments were not the basis for the conclusion, but rather it was the gender
admissions pooled across all departments, while weighing by each department's rejection rate across all of
its applicants. Whether the data show a definite women-favoring bias or just a minority-favoring bias (or a
combination thereof) could be a different aspect for analysis: the data possibly show a bias in favor of the
minority gender, as is visible in occurrence of 'more applicants' (orange) in the exact opposite gender than
the 'more successful applicants' (green), and women were the minority in the entire population of applicants
(see totals), thus are more probable to be the minority in a greater number of departments (would only not
be so if men excess of 856 from the totals was accumulated in the top men departments, which is not the
case). The paper does not explore this detail however (although it does recognize "drive to recruit minority
group members" as explanation for some women-only data phenomena).[16]

Kidney stone treatment

Another example comes from a real-life medical study[17] comparing the success rates of two treatments for
kidney stones.[18] The table below shows the success rates (the term success rate here actually means the
success proportion) and numbers of treatments for treatments involving both small and large kidney stones,
where Treatment A includes open surgical procedures and Treatment B includes closed surgical procedures.
The numbers in parentheses indicate the number of success cases over the total size of the group.
 Treatment
Treatment A Treatment B
Stone size    
Group 1 Group 2
Small stones
93% (81/87) 87% (234/270)

Group 3 Group 4
Large stones
73% (192/263) 69% (55/80)
Both 78% (273/350) 83% (289/350)

The paradoxical conclusion is that treatment A is more effective when used on small stones, and also when
used on large stones, yet treatment B appears to be more effective when considering both sizes at the same
time. In this example, the "lurking" variable (or confounding variable) causing the paradox is the size of the
stones, which was not previously known to researchers to be important until its effects were included.

Which treatment is considered better is determined by which success ratio (successes/total) is larger. The
reversal of the inequality between the two ratios when considering the combined data, which creates
Simpson's paradox, happens because two effects occur together:

1. The sizes of the groups, which are combined when the lurking variable is ignored, are very
different. Doctors tend to give cases with large stones the better treatment A, and the cases
with small stones the inferior treatment B. Therefore, the totals are dominated by groups 3
and 2, and not by the two much smaller groups 1 and 4.
2. The lurking variable, stone size, has a large effect on the ratios; i.e., the success rate is more
strongly influenced by the severity of the case than by the choice of treatment. Therefore, the
group of patients with large stones using treatment A (group 3) does worse than the group
with small stones, even if the latter used the inferior treatment B (group 2).

Based on these effects, the paradoxical result is seen to arise because the effect of the size of the stones
overwhelms the benefits of the better treatment (A). In short, the less effective treatment B appeared to be
more effective because it was applied more frequently to the small stones cases, which were easier to
treat.[18]

Batting averages

A common example of Simpson's paradox involves the batting averages of players in professional baseball.
It is possible for one player to have a higher batting average than another player each year for a number of
years, but to have a lower batting average across all of those years. This phenomenon can occur when there
are large differences in the number of at bats between the years. Mathematician Ken Ross demonstrated this
using the batting average of two baseball players, Derek Jeter and David Justice, during the years 1995 and
1996:[19][20]

Year
1995 1996 Combined
Batter  
Derek Jeter 12/48 .250 183/582 .314 195/630 .310

David Justice 104/411 .253 45/140 .321 149/551 .270

In both 1995 and 1996, Justice had a higher batting average (in bold type) than Jeter did. However, when
the two baseball seasons are combined, Jeter shows a higher batting average than Justice. According to
Ross, this phenomenon would be observed about once per year among the possible pairs of players.[19]

Vector interpretation
Simpson's paradox can also be illustrated using a 2-dimensional
vector space.[21] A success rate of (i.e., successes/attempts) can
be represented by a vector , with a slope of . A steeper
vector then represents a greater success rate. If two rates and
are combined, as in the examples given above, the result can be
represented by the sum of the vectors and , which Vector interpretation of Simpson's
according to the parallelogram rule is the vector paradox
, with slope .

Simpson's paradox says that even if a vector (in orange in figure) has a smaller slope than another
vector (in blue), and has a smaller slope than , the sum of the two vectors can
potentially still have a larger slope than the sum of the two vectors , as shown in the example. For
this to occur one of the orange vectors must have a greater slope than one of the blue vectors (here and
), and these will generally be longer than the alternatively subscripted vectors – thereby dominating the
overall comparison.

Correlation between variables

Simpson's reversal can also arise in correlations, in which two variables appear to have (say) a positive
correlation towards one another, when in fact they have a negative correlation, the reversal having been
brought about by a "lurking" confounder. Berman et al.[22] give an example from economics, where a
dataset suggests overall demand is positively correlated with price (that is, higher prices lead to more
demand), in contradiction of expectation. Analysis reveals time to be the confounding variable: plotting
both price and demand against time reveals the expected negative correlation over various periods, which
then reverses to become positive if the influence of time is ignored by simply plotting demand against price.

Psychology
Psychological interest in Simpson's paradox seeks to explain why people deem sign reversal to be
impossible at first, offended by the idea that an action preferred both under one condition and under its
negation should be rejected when the condition is unknown. The question is where people get this strong
intuition from, and how it is encoded in the mind.

Simpson's paradox demonstrates that this intuition cannot be derived from either classical logic or
probability calculus alone, and thus led philosophers to speculate that it is supported by an innate causal
logic that guides people in reasoning about actions and their consequences.[4] Savage's sure-thing
principle[12] is an example of what such logic may entail. A qualified version of Savage's sure thing
principle can indeed be derived from Pearl's do-calculus[4] and reads: "An action A that increases the
probability of an event B in each subpopulation Ci of C must also increase the probability of B in the
population as a whole, provided that the action does not change the distribution of the subpopulations."
This suggests that knowledge about actions and consequences is stored in a form resembling Causal
Bayesian Networks.

Probability
A paper by Pavlides and Perlman presents a proof, due to Hadjicostas, that in a random 2 × 2 × 2 table with
uniform distribution, Simpson's paradox will occur with a probability of exactly 1 ⁄60 .[23] A study by Kock
suggests that the probability that Simpson's paradox would occur at random in path models (i.e., models
generated by path analysis) with two predictors and one criterion variable is approximately 12.8 percent;
slightly higher than 1 occurrence per 8 path models.[24]

Simpson's second paradox

A second, less well-known paradox was also discussed in Simpson's 1951 paper. It can occur when the
"sensible interpretation" is not necessarily found in the separated data, like in the Kidney Stone example,
but can instead reside in the combined data. Whether the partitioned or combined form of the data should
be used hinges on the process giving rise to the data, meaning the correct interpretation of the data cannot
always be determined by simply observing the tables.[25]

Judea Pearl has shown that, in order for the partitioned data to represent the correct causal relationships
between any two variables, and , the partitioning variables must satisfy a graphical condition called
"back-door criterion":[26][27]

1. They must block all spurious paths between and

2. No variable can be affected by

This criterion provides an algorithmic solution to Simpson's second paradox, and explains why the correct
interpretation cannot be determined by data alone; two different graphs, both compatible with the data, may
dictate two different back-door criteria.

When the back-door criterion is satisfied by a set Z of covariates, the adjustment formula (see Confounding)
gives the correct causal effect of X on Y. If no such set exists, Pearl's do-calculus can be invoked to
discover other ways of estimating the causal effect.[4][28] The completeness of do-calculus [29][28] can be
viewed as offering a complete resolution of the Simpson's paradox.

Criticism
One criticism is that the paradox is not really a paradox at all, but rather a failure to properly account for
confounding variables or to consider causal relationships between variables.[30]

Another criticism of the apparent Simpson's paradox is that it may be a result of the specific way that data is
stratified or grouped. The phenomenon may disappear or even reverse if the data is stratified differently or
if different confounding variables are considered. Simpson's example actually highlighted a phenomenon
called noncollapsibility,[31] which occurs when subgroups with high proportions do not make simple
averages when combined. This suggests that the paradox may not be a universal phenomenon, but rather a
specific instance of a more general statistical issue.
Critics of the apparent Simpson's paradox also argue that the focus on the paradox may distract from more
important statistical issues, such as the need for careful consideration of confounding variables and causal
relationships when interpreting data.[32]

Despite these criticisms, the apparent Simpson's paradox remains a popular and intriguing topic in statistics
and data analysis. It continues to be studied and debated by researchers and practitioners in a wide range of
fields, and it serves as a valuable reminder of the importance of careful statistical analysis and the potential
pitfalls of simplistic interpretations of data.

See also
Aliasing – Signal processing effect
Anscombe's quartet – Four data sets with the same descriptive statistics, yet very different
distributions
Berkson's paradox – Tendency to misinterpret statistical experiments involving conditional
probabilities
Cherry picking – Fallacy of incomplete evidence
Condorcet paradox – Situation in social choice theory where collective preferences are
cyclic
Ecological fallacy – Logical fallacy that occurs when group characteristics are applied to
individuals
Low birth-weight paradox – Statistical quirk of babies' birth weights
Modifiable areal unit problem – Source of statistical bias
Prosecutor's fallacy – Error in thinking which involves under-valuing base rate information
Will Rogers phenomenon – phenomenon in which moving an element from one set to
another set raises the average values of both sets
Spurious correlation
Omitted-variable bias

References
1. Clifford H. Wagner (February 1982). "Simpson's Paradox in Real Life". The American
Statistician. 36 (1): 46–48. doi:10.2307/2684093 (https://doi.org/10.2307%2F2684093).
JSTOR 2684093 (https://www.jstor.org/stable/2684093).
2. Holt, G. B. (2016). Potential Simpson's paradox in multicenter study of intraperitoneal
chemotherapy for ovarian cancer. (http://jco.ascopubs.org/content/34/9/1016.1.full) Journal of
Clinical Oncology, 34(9), 1016–1016.
3. Franks, Alexander; Airoldi, Edoardo; Slavov, Nikolai (2017). "Post-transcriptional regulation
across human tissues" (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5440056). PLOS
Computational Biology. 13 (5): e1005535. arXiv:1506.00219 (https://arxiv.org/abs/1506.0021
9). Bibcode:2017PLSCB..13E5535F (https://ui.adsabs.harvard.edu/abs/2017PLSCB..13E55
35F). doi:10.1371/journal.pcbi.1005535 (https://doi.org/10.1371%2Fjournal.pcbi.1005535).
ISSN 1553-7358 (https://www.worldcat.org/issn/1553-7358). PMC 5440056 (https://www.ncb
i.nlm.nih.gov/pmc/articles/PMC5440056). PMID 28481885 (https://pubmed.ncbi.nlm.nih.gov/
28481885).
4. Judea Pearl. Causality: Models, Reasoning, and Inference, Cambridge University Press
(2000, 2nd edition 2009). ISBN 0-521-77362-8.
 5. Kock, N., & Gaskins, L. (2016). Simpson's paradox, moderation and the emergence of
quadratic relationships in path models: An information systems illustration. (http://cits.tamiu.e
du/kock/pubs/journals/2016JournalIJANS_ModJCveNetCorrp/Kock_Gaskins_2016_IJANS
_SimpPdox.pdf) International Journal of Applied Nonlinear Science, 2(3), 200–234.
6. Rogier A. Kievit, Willem E. Frankenhuis, Lourens J. Waldorp and Denny Borsboom,
Simpson's paradox in psychological science: a practical guide
https://doi.org/10.3389/fpsyg.2013.00513
7. Robert L. Wardrop (February 1995). "Simpson's Paradox and the Hot Hand in Basketball".
The American Statistician, 49 (1): pp. 24–28.
8. Alan Agresti (2002). "Categorical Data Analysis" (Second edition). John Wiley and Sons
ISBN 0-471-36093-7
9. Simpson, Edward H. (1951). "The Interpretation of Interaction in Contingency Tables".
Journal of the Royal Statistical Society, Series B. 13: 238–241.
10. Pearson, Karl; Lee, Alice; Bramley-Moore, Lesley (1899). "Genetic (reproductive) selection:
Inheritance of fertility in man, and of fecundity in thoroughbred racehorses" (https://doi.org/1
0.1098%2Frsta.1899.0006). Philosophical Transactions of the Royal Society A. 192: 257–
330. doi:10.1098/rsta.1899.0006 (https://doi.org/10.1098%2Frsta.1899.0006).
11. G. U. Yule (1903). "Notes on the Theory of Association of Attributes in Statistics" (https://zeno
do.org/record/1431599). Biometrika. 2 (2): 121–134. doi:10.1093/biomet/2.2.121 (https://doi.
org/10.1093%2Fbiomet%2F2.2.121).
12. Colin R. Blyth (June 1972). "On Simpson's Paradox and the Sure-Thing Principle". Journal
of the American Statistical Association. 67 (338): 364–366. doi:10.2307/2284382 (https://doi.
org/10.2307%2F2284382). JSTOR 2284382 (https://www.jstor.org/stable/2284382).
13. I. J. Good, Y. Mittal (June 1987). "The Amalgamation and Geometry of Two-by-Two
Contingency Tables" (https://doi.org/10.1214%2Faos%2F1176350369). The Annals of
Statistics. 15 (2): 694–711. doi:10.1214/aos/1176350369 (https://doi.org/10.1214%2Faos%2
F1176350369). ISSN 0090-5364 (https://www.worldcat.org/issn/0090-5364).
JSTOR 2241334 (https://www.jstor.org/stable/2241334).
14. Ellenberg, Jordan (May 25, 2021). Shape: The Hidden Geometry of Information, Biology,
Strategy, Democracy and Everything Else (https://www.worldcat.org/oclc/1226171979). New
York: Penguin Press. p. 228. ISBN 978-1-9848-7905-9. OCLC 1226171979 (https://www.wor
ldcat.org/oclc/1226171979).
15. David Freedman, Robert Pisani, and Roger Purves (2007), Statistics (4th edition), W. W.
Norton. ISBN 0-393-92972-8.
16. P.J. Bickel, E.A. Hammel and J.W. O'Connell (1975). "Sex Bias in Graduate Admissions:
Data From Berkeley" (http://homepage.stat.uiowa.edu/~mbognar/1030/Bickel-Berkeley.pdf)
(PDF). Science. 187 (4175): 398–404. Bibcode:1975Sci...187..398B (https://ui.adsabs.harva
rd.edu/abs/1975Sci...187..398B). doi:10.1126/science.187.4175.398 (https://doi.org/10.112
6%2Fscience.187.4175.398). PMID 17835295 (https://pubmed.ncbi.nlm.nih.gov/17835295).
S2CID 15278703 (https://api.semanticscholar.org/CorpusID:15278703). Archived (https://we
b.archive.org/web/20160604220121/http://homepage.stat.uiowa.edu/~mbognar/1030/Bickel-
Berkeley.pdf) (PDF) from the original on 2016-06-04.
17. C. R. Charig; D. R. Webb; S. R. Payne; J. E. Wickham (29 March 1986). "Comparison of
treatment of renal calculi by open surgery, percutaneous nephrolithotomy, and
extracorporeal shockwave lithotripsy" (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13399
81). Br Med J (Clin Res Ed). 292 (6524): 879–882. doi:10.1136/bmj.292.6524.879 (https://do
i.org/10.1136%2Fbmj.292.6524.879). PMC 1339981 (https://www.ncbi.nlm.nih.gov/pmc/artic
les/PMC1339981). PMID 3083922 (https://pubmed.ncbi.nlm.nih.gov/3083922).
18. Steven A. Julious; Mark A. Mullee (3 December 1994). "Confounding and Simpson's
paradox" (http://bmj.bmjjournals.com/cgi/content/full/309/6967/1480). BMJ. 309 (6967):
1480–1481. doi:10.1136/bmj.309.6967.1480 (https://doi.org/10.1136%2Fbmj.309.6967.148
0). PMC 2541623 (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2541623). PMID 7804052
(https://pubmed.ncbi.nlm.nih.gov/7804052).
19. Ken Ross. "A Mathematician at the Ballpark: Odds and Probabilities for Baseball Fans
(Paperback)" Pi Press, 2004. ISBN 0-13-147990-3. 12–13
20. Statistics available from Baseball-Reference.com: Data for Derek Jeter (https://www.basebal
l-reference.com/j/jeterde01.shtml); Data for David Justice (https://www.baseball-reference.co
m/j/justida01.shtml).
21. Kocik Jerzy (2001). "Proofs without Words: Simpson's Paradox" (http://www.math.siu.edu/ko
cik/papers/simpson2.pdf) (PDF). Mathematics Magazine. 74 (5): 399. doi:10.2307/2691038
(https://doi.org/10.2307%2F2691038). JSTOR 2691038 (https://www.jstor.org/stable/269103
8). Archived (https://web.archive.org/web/20100612220747/http://www.math.siu.edu/kocik/pa
pers/simpson2.pdf) (PDF) from the original on 2010-06-12.
22. Berman, S. DalleMule, L. Greene, M., Lucker, J. (2012), "Simpson's Paradox: A Cautionary
Tale in Advanced Analytics (http://www.statslife.org.uk/the-statistics-dictionary/2012-simpson
-s-paradox-a-cautionary-tale-in-advanced-analytics) Archived (https://web.archive.org/web/2
0200510171740/https://www.statslife.org.uk/the-statistics-dictionary/2012-simpson-s-parado
x-a-cautionary-tale-in-advanced-analytics) 2020-05-10 at the Wayback Machine",
Significance.
23. Marios G. Pavlides & Michael D. Perlman (August 2009). "How Likely is Simpson's
Paradox?" (https://semanticscholar.org/paper/0dc06fe2b58a3b4758c57198609f5c66550ada
e4). The American Statistician. 63 (3): 226–233. doi:10.1198/tast.2009.09007 (https://doi.org/
10.1198%2Ftast.2009.09007). S2CID 17481510 (https://api.semanticscholar.org/CorpusID:1
7481510).
24. Kock, N. (2015). How likely is Simpson's paradox in path models? (http://cits.tamiu.edu/kock/
pubs/journals/2015JournalIJeC/Kock_2015_IJeC_SimpPdox.pdf) International Journal of e-
Collaboration, 11(1), 1–7.
25. Norton, H. James; Divine, George (August 2015). "Simpson's paradox ... and how to avoid it"
(https://doi.org/10.1111%2Fj.1740-9713.2015.00844.x). Significance. 12 (4): 40–43.
doi:10.1111/j.1740-9713.2015.00844.x (https://doi.org/10.1111%2Fj.1740-9713.2015.00844.
x).
26. Pearl, Judea (2014). "Understanding Simpson's Paradox". The American Statistician. 68 (1):
8–13. doi:10.2139/ssrn.2343788 (https://doi.org/10.2139%2Fssrn.2343788).
S2CID 2626833 (https://api.semanticscholar.org/CorpusID:2626833).
27. Pearl, Judea (1993). "Graphical Models, Causality, and Intervention" (https://doi.org/10.121
4%2Fss%2F1177010894). Statistical Science. 8 (3): 266–269. doi:10.1214/ss/1177010894
(https://doi.org/10.1214%2Fss%2F1177010894).
28. Pearl, J.; Mackenzie, D. (2018). The Book of Why: The New Science of Cause and Effect.
New York, NY: Basic Books.
29. Shpitser, I.; Pearl, J. (2006). Dechter, R.; Richardson, T.S. (eds.). "Identification of
Conditional Interventional Distributions". Proceedings of the Twenty-Second Conference on
Uncertainty in Artificial Intelligence. Corvallis, OR: AUAI Press: 437–444.
30. Blyth, Colin R. (June 1972). "On Simpson's Paradox and the Sure-Thing Principle" (http://w
ww.tandfonline.com/doi/abs/10.1080/01621459.1972.10482387). Journal of the American
Statistical Association. 67 (338): 364–366. doi:10.1080/01621459.1972.10482387 (https://do
i.org/10.1080%2F01621459.1972.10482387). ISSN 0162-1459 (https://www.worldcat.org/iss
n/0162-1459).
31. Greenland, Sander (2021-11-01). "Noncollapsibility, confounding, and sparse-data bias.
Part 2: What should researchers make of persistent controversies about the odds ratio?" (htt
ps://www.jclinepi.com/article/S0895-4356(21)00182-7/fulltext). Journal of Clinical
Epidemiology. 139: 264–268. doi:10.1016/j.jclinepi.2021.06.004 (https://doi.org/10.1016%2F
j.jclinepi.2021.06.004). ISSN 0895-4356 (https://www.worldcat.org/issn/0895-4356).
PMID 34119647 (https://pubmed.ncbi.nlm.nih.gov/34119647).
32. Hernán, Miguel A.; Clayton, David; Keiding, Niels (June 2011). "The Simpson's paradox
unraveled" (https://pubmed.ncbi.nlm.nih.gov/21454324). International Journal of
Epidemiology. 40 (3): 780–785. doi:10.1093/ije/dyr041 (https://doi.org/10.1093%2Fije%2Fdy
r041). ISSN 1464-3685 (https://www.worldcat.org/issn/1464-3685). PMC 3147074 (https://w
ww.ncbi.nlm.nih.gov/pmc/articles/PMC3147074). PMID 21454324 (https://pubmed.ncbi.nlm.
nih.gov/21454324).

Bibliography
Leila Schneps and Coralie Colmez, Math on trial. How numbers get used and abused in the
courtroom, Basic Books, 2013. ISBN 978-0-465-03292-1. (Sixth chapter: "Math error number
6: Simpson's paradox. The Berkeley sex bias case: discrimination detection").

External links
Simpson's Paradox (http://plato.stanford.edu/entries/paradox-simpson/) at the Stanford
Encyclopedia of Philosophy, by Jan Sprenger and Naftali Weinberger.
How statistics can be misleading – Mark Liddell (http://ed.ted.com/lessons/how-statistics-can
-be-misleading-mark-liddell) – TED-Ed video and lesson.
Pearl, Judea, "Understanding Simpson’s Paradox" (https://ftp.cs.ucla.edu/pub/stat_ser/r414.
pdf) (PDF)
Simpson's Paradox (http://www.cut-the-knot.org/Curriculum/Algebra/SimpsonParadox.shtm
l), a short article by Alexander Bogomolny on the vector interpretation of Simpson's paradox
The Wall Street Journal column "The Numbers Guy" (https://www.wsj.com/articles/SB12597
0744553071829) for December 2, 2009 dealt with recent instances of Simpson's paradox in
the news. Notably a Simpson's paradox in the comparison of unemployment rates of the
2009 recession with the 1983 recession.
At the Plate, a Statistical Puzzler: Understanding Simpson's Paradox (http://www.stateoftheu
sa.org/content/at-the-plate-a-statistical-puz.php) by Arthur Smith, August 20, 2010
Simpson's Paradox (https://www.youtube.com/watch?v=ebEkn-BiW5k), a video by Henry
Reich of MinutePhysics

Retrieved from "https://en.wikipedia.org/w/index.php?title=Simpson%27s_paradox&oldid=1165543462"