Everitt B.S. The Cambridge Dictionary of Statistics

Подождите немного. Документ загружается.

Gold standard trials: A term usually retained for those

clinical trials

in which there is random

allocation to treatments, a control group and double-blinding.

Gol ub , Gene H oward ( 1 932^20 07): Golub was born in Chicago in 1932. After taking his PhD

at the University of Illinois he arrived at Stanford University in 1965. Gloub became a

founding member of the Computer Science Department in 1962 and remained at Stanford

throughout his career. Golub made fundamental contributions to computational linear

algebra. In particular, he transformed the ﬁeld of matrix computations, his most famous

achievements being associated with the

singular value decomposition

. Golub served as

president of the Society for Industrial and Applied Mathematics (SIAM) and was founding

editor of two SIAM journals. He died at Stanford on November 16th 2007.

Gompertz curve: A curve used to describe the size of a population (y) as a function of time (t),

where relative growth rate declines at a constant rate. Explicitly given by

y ¼ ae

b

[MV2 Chapter 13.]

G ood, Irving J o hn (191 6^2009 ): Born in London, England, Good was a child prodigy in

mathematics and won a mathematics scholarship to Jesus College, Cambridge from where

he graduated in 1938 with a ﬁrst. He obtained his in Ph.D in 1941 for a thesis on topological

dimension. During the Second World War Good worked at Bletchley Park on deciphering

German military and naval radio trafﬁc. After the war he moved to the University of

Manchester and then to Trinity College, Oxford. In 1967 Good moved to the USA as a

professor of statistics at Virginia Tech where he remained until his retirement in 1994. Good

made fundamental and far reaching contributions to several areas of statistics particularly

Bayesian statistics

. He died on April 5th, 2009.

Goodman and Kruskalmeasures of association: Measures of associations that are useful

in the situation where two categorical variables cannot be assumed to be derived from

perhaps unobservable continuous variables and where there is no natural ordering of interest.

The rationale behind the measures is the question, ‘how much does knowledge of the

classiﬁcation of one of the variables improve the ability to predict the classiﬁcation on the

other variable’.[The Analysis of Contingency Tables, 2nd edition, 1992, B. S. Everitt,

Chapman and Hall/CRC Press, London.]

Goodness-of-fitstatistics: Measures of the agreement between a set of sample observations and

the corresponding values predicted from some model of interest. Many such measures have

been suggested; see

chi-squared statistic

deviance

likelihood ratio

, G

and X

. [SMR

Chapter 8.]

G ood’s method: A procedure for combining independent tests of hypotheses. [Probability and the

Weighing of Evidence, 1950, I. J. Good, Hafner, New York.]

Gossett,William Sealy (1876^1937): Born in Canterbury, England, Gossett obtained a degree

in chemistry at Oxford before joining Guinness breweries in Dublin in 1899. He continued

to work for Guinness for the next three decades. Practical problems in his work led him to

seek exact error probabilities of statistics from small samples, a previously un-researched

area. Spent the academic year 1906–1907 studying under

Karl Pearson

at University

College, London. Writing under the pseudonym of ‘Student’ (as required by Guinness)

his paper essentially introducing the Student’s t-distribution was published in Biometrika in

1908. Gossett died in London on 16 October 1937.

189

Goulden, Cyri l Harold ( 1 897^1 981): Born in Bridgend, Wales, Goulden’s family moved to

Canada and he graduated from the University of Saskatchewan in 1921. Between 1925 and

1930 he taught himself mathematics and statistical methods and began to apply methods

developed by

Gosset

and

Pearson

in agricultural experiments as part of his job at the

Dominion Rust Research Laboratory in Winnipeg. In 1930 Goulden went to work at

Rothamsted Experimental Station and met

Fisher

. His book, Methods of Statistical

Analysis, became a standard textbook in Canada and the USA. In 1958 he acted as

President of the Biometrics Society. Goulden died on February 3rd, 1981 in Ottawa, Canada.

Gower’s similarity coefficient: A

similarity coef ﬁcie nt

particularly suitable when the

measurements con tain b oth continuous variables and cate gorical variables. [MV2

Chapter 10.]

G r ade of membersh ip model: A general

distribution free method

for the clustering of

multivariate data in which only categorical variables are involved. The model assumes

that individuals can exhibit characteristics of more than one cluster, and that the state of an

individual can be represented by a set of numerical quantities, each one corresponding to one

of the clusters, that measure the ‘strength ’ or grade of membership of the individual for the

cluster. Estimation of these quantities and the other parameters in the model is undertaken by

maximum likelihood estimation

. See also latent class analysis and fuzzy set theory.

[Statistical Applications of Fuzzy Sets , 1994, K. G. Manton, M. A. Woodbury and H. D.

Tolley, Wiley, New York.]

Graduation: A term employed most often in the application of

actuarial statistics

to denote procedures

by which a set of observed probabilities is adjusted to provide a suitable basis for inferences

and further practical calculations to be made. [The Analysis of Mortality and Other Actuarial

Statistics, 3rd edition, 1993, B. Benjamin and J. H. Pollard, Institute and Faculty of

Actuaries, Oxford.]

Gra eco -Latin sq ua r e: An extension of a

Latin square

that allows for three extraneous sources of

variation in an experiment. A three-by-three example of such a square is

Aα Bβ Cγ

Bγ Cα Aβ

Cβ Aγ Bα

[The Mathematics of Experimental Designs; Incomplete Block Designs and Latin Squares,

S. Vajda, 1967, Charles Grifﬁn, London.]

Gram^Charlier Type A series: An expansion of a probability distribution, f(x) in terms of

Chebyshev–Hermite polynomials

, H

ðxÞ. Given explicitly by

f ðxÞ¼αðxÞ 1 þ

ð

 1ÞH



ð

 6

þ 3ÞH

þ



where

αðxÞ¼

ﬃﬃﬃﬃﬃﬃ



See also Edgeworth’s form of the Type A series. [KA1 Chapter 6.]

Gramian matrix: A symmetric matrix, A, whose elements are real numbers and for which there

exists a matrix B also consisting of real numbers, such that BB

=A or B

B=A. An example is

correlation matrix

190

Grand mean: Mean of all the values in a grouped data set irrespective of groups.

Granger causality: A

time series

X is said to Granger-cause a time series Y if lagged values of

X predict future values of Y, given the lagged values of Y.[Econometrica, 1969, 37,

424–438.]

Granger, Sir CliveWilliam John (193 4^20 09): Granger was born in Swansea, Wales, in 1934.

He developed a strong interest in mathematics as a child but ambled into a career in statistics

haphazardly. As a grammar school student in 1946, his teacher instructed him and his

classmates to stand up one day and announce the careers they wanted. “I preferred to use

mathematics in some practical fashion and thought that meteorology sounded promising,”

he wrote in his Nobel biography. “In those days I stuttered somewhat and when my turn

came to stand up, I tried to say ‘meteorology’ but found I could not get the word out, so I just

said ‘statistics,’ thereby determining my future path.” Granger went on to the University of

Nottingham, where he earned a bachelor’s degree in mathematics in 1955 and a Ph.D. in

statistics in 1959, publishing his ﬁrst statistical paper on sunspot activity. He worked at

Nottingham until the early 1970s when he moved to the University of California at San

Diego, where he built one of the world’s most prominent econometrics programs. Granger

was awarded the Nobel Prize in economics in 2003 for his contributions to the analysis of

nonstationary time-series

Graphicaldeception:

Statistical graphics

that are not as honest as they should be. It is relatively easy

to mislead the unwary with graphical material. For example, consider the plot of the death rate

per million from cancer of the breast, for several periods over the last three decades, shown in

Figure 68(a). The rate appears to show a rather alarming increase. However, when the data are

replotted with the vertical scale beginning at zero, as shown in Figure 68(b), the increase in the

breast cancer death rate is altogether less startling. This example illustrates that undue

exaggeration or compression of the scales is best avoided when drawing graphs. For another

example of a common graphical distortion see

lie factor

.[Visual Display of Quantitative

Information, 1983, E. R. Tufte, Graphics Press, Cheshire, Connecticut.]

500450400

Deaths per million

350

500300 400100 200

Deaths per million

1955 1965 1975

Year

(a) (b)

1955 1965 1975

Year

Fig. 68 (a) Death rates

from cancer of the breast

where the y-axis does

not include the origin.

(b) Death rates from

cancer of the breast when

the y axis does include

the origin.

191

Graph theory: A branch of mathematics concerned with the properties of sets of points (vertices or

nodes) some of which are connected by lines known as edges. A directed graph is one in

which direction is associated with the edges and an undirected graph is one i n which no

direction is invol ved in the connecti ons between points. A gra ph may be repres ented as an

adjacency matrix

. See also conditional independence graph.[Graphs and Digraphs ,

1979, M. Behzad, G. Chartrand and L. Lesniak-Foster, Prindle, Weber and Schmidt,

Boston.]

G raunt, Joh n ( 1 620^1 674): The son of a city tradesman, Graunt is generally regarded as having

laid the foundations of

demography

as a science with the publication of his Natural and

Political Observations Made Upon the Bills of Mortality published in 1662. His most

important contribution was the introduction of a rudimentary

life table

. Graunt died on 18

April, 1674 in London.

Greatest characteristic root test: Synonym for Roy’s largest root criterion.

Greenberg, Bernard George (1919^198 5): Born in New York City, Greenberg obtained a

degree in mathematics from the City College of New York in 1939. Ten years later after a

period of military service he obtained a Ph.D. from the North Carolina State University

where he studied under

Hotelling

. Greenberg founded the Department of Biostatistics at the

University of North Carolina and was a pioneer in the ﬁeld of public health and medical

research. He died on 24 November 1985 in Chapel Hill.

Greenhouse^Geisser correction: A method of adjusting the degrees of freedom of the within-

subject

F-tests

in the

analysis of variance

longitudinal data

so as to allow for possible

departures of the

variance–covariance matrix

of the measurements from the assumption of

sphericity

. If this condition holds for the data then the correction factor is one and the simple

F-tests are valid. Departures from

sphericity

result in an estimated correction factor less than

one, thus reducing the degrees of freedom of the relevant F-tests. See also Huynh –Feldt

correction and Mauchly test. [MV2 Chapter 13.]

Greenhouse, Samuel (1918^2000): Greenhouse began his career at the National Institutes of

Health in the National Cancer Institute. Later he become Chief of the Theoretical Statistics

and Mathematics Section in the National Institute of Mental Health. After 1974 Greenhouse

undertook a full time academic career at George Washington University. He was inﬂuential

in the early development of the theory and practice of

clinical trials

and his work also

included the evaluation of diagnostic tests and the analysis of

repeated measure designs

Greenhouse died on 28 September 2000 in Rockville, MD, USA.

Greenwood, Major (1880^1949): Born in the East End of London, Greenwood studied medi-

cine at University College, London and the London Hospital, but shortly after qualifying

forsook clinical medicine and following a period of study with Karl Pearson was, in 1910,

appointed statistician to the Lister Institute. Here he carried out statistical investigations into

such diverse topics as the fatality of fractures and pneumonia in hospital practice, the

epidemiology of plague and factors inﬂuencing rates of infant mortality. In 1919 he became

Head of Medical Statistics in the newly created Ministry of Health where he remained until

1928, when he was appointed to the chair of Vital Statistics and Epidemiology at the London

School of Hygiene and Tropical Medicine. Here he remained until his retirement in 1945.

Greenwood was President of the Royal Statistical Society from 1934 to 1936 and was

awarded their Guy medal in gold in 1945. He died on 5 October 1949.

Green w ood’sformula: A formula giving the variance of the

product limit estimator

of a

survival

function

, namely

192

varð

SðtÞÞ¼½

SðtÞ

jjt

ðjÞ

t

ðr

 d

where

SðtÞis the estimated

survival function

at time, t; t

ð1Þ

ð2Þ



ðnÞ

are the ordered,

observed

survival times

, r

is the number of individuals at risk at time t

(j)

, and d

is the number

who experience the event of interest at time t

. (Individuals censored at t

(j)

are included in r

)

[Modelling Survival Data in Medical Research, 2nd edn, 2003, D. Collett, Chapman and

Hall/CRC Press, London.]

Gripenberg estimator: A

distribution free

estimator for the

partial correlation

between two

variables, X and Y, conditional on a third variable, Z.[Journal of the American Statistical

Association, 1992, 87, 546–51.]

G roup average clustering: Synonym for average linkage clustering.

Group divisible design: An arrangement of v ¼ mn treatments in b blocks such that:

each block contains k distinct treatments, k

each treatment is replicated r times;

the treatments can be divided into m groups of n treatments each, any two treatments

occurring together in λ

blocks if they belong to the same group and in λ

blocks if

they belong to different groups. [Biometrika, 1976, 63, 555–8.]

Groupedbinarydata: Observations on a binary variable tabulated in terms of the proportion of one

of the two possible outcomes amongst patients or subjects who are, for example, the same

diagnosis or same sex, etc. [SORT, 2004, 28, 125–60.]

Grouped data: Data recorded as frequencies of observations in particular intervals.

Group matching: See matching.

G rou p sequent ia l desi g n: See sequential design.

Growth char ts: Synonym for centile reference charts.

Growth curve: See growth curve analysis.

G rowth curve analysis: A general term for methods dealing with the development of individuals

over time. A classic example involves recordings made on a group of children, say, of height

or weight at particular ages. A plot of the former against the latter for each child gives the

individual’s growth curve. Traditionally low-degree polynomials are ﬁtted to such curves,

and the resulting parameter estimates used for inferences such as comparisons between boy

and girls. See also multilevel models and random effects. [An Introduction to Latent

Variable Growth Curve Modeling: Concepts, Issues and Applications, 1999, S. C. Duncan

and L. A. Strycker, Lawrence Erlbaum, Mahwah, NJ.]

Growth rate: A measure of population growth calculated as

live births during the year  deaths during the year

midyear population

 100

Grubb’s estimators: Estimators of the measuring precisions when two instruments or techniques

are used to measure the same quantity. For example, if the two measurements are denoted by

and y

for i =1,..., n, we assume that

193

¼ τ

þ 

¼ τ

þ 

where τ

is the correct unknown value of the ith quantity and 

and δ

are measurement errors

assumed to be independent, then Grubb’s estimators are

V ðτ

Þ¼covarianceðx; yÞ

V ð

Þ¼varianceðxÞcovarianceðx; yÞ

V ð

Þ¼varianceðyÞcovarianceðx; yÞ

[Technometrics, 1973, 15,53–66.]

GTdistribution: A probability distribution, f (x), related to

Student’s t-distribution

and given by

f ðxÞ¼

2q

1=p

Bðp

1

; qÞ

1 þ

jxj

ðq



qþp

1

where B is the

beta function

In f(x), σ is a

scale parameter

while p and q control the shape of the distribution. Larger

values of p and q are associated with lighter tails for the distribution. When p = 2 and

 ¼

ﬃﬃﬃ

α, this becomes Student’s t-distribution with 2q degrees of freedom. [Continuous

Univariate Distributions, Volume 2, 2nd edition, 1995, N. L. Johnson, S. Kotz and

N. Balakrishnan, Wiley, New York.]

Guarantee time: See two-parameter exponential distribution.

G umbel d istr i buti o n: Synonym for extreme value distribution.

Gumbel, Emil Julius (1891^1966): Born in Munich, Germany, Gumbel obtained a Ph.D. in

economics and mathematical statistics from the University of Munich in 1914. Between

1933 and 1940 he worked in France, ﬁrst at the Institut Henri Poincaré in Paris, later at the

University of Lyon. Gumbel made important contributions to the theory of

extreme values

Gupta’s selection procedure: A method which, for given samples of size n from each of k

normal populations, selects a subset of the k populations which contains the population with

the largest mean with some minimum probability. [Communications in Statistics: Theory

and Methods, 1997, 26, 1291–1312.]

Guttman, Louis (1916^1987): Born in Brooklyn, New York, Guttman received his Ph.D in

sociology in 1942 from the University of Minnesota with a thesis on the algebra of

factor

analysis

. In 1947 he emigrated to Israel where he founded and directed the Israel Institute of

Applied Social Research, later to be known as the Guttman Institute of Applied Social

Research. From 1955 he also became Professor of Social and Psychological Assessment at

the Hebrew University of Jerusalem. Guttman’s greatest contributions to statistics were in

the area of scaling theory and factor analysis. He died on October 25th, 1987 in Minneapolis,

Minnesota, USA.

Guttman scale: A scale based on a set of

binary variables

which measure a one-dimensional

latent

variable

.Thebasicideaisthatk binary items are ordered such that a subject who answers “Yes”

to item j (< k) will probably answer “Yes” to items j +1,..., k.SeealsoCronbach’salpha.

[Scale Development Theory and Applications, 1991, R. F. Devellis, Sage, Beverly Hills.]

Guy,William Augustus (1810^18 85): Born in Chichester, Guy studied medicine at both Christ’s

Hospital and Guy’s Hospital, London. In 1831 he was awarded the Fothergillian medal of

194

the Medical Society of London for the best paper on asthma. In 1838 Guy was appointed

to the Chair of Forensic Medicine at King’s College, London. Guy, like

Farr

, was strongly of

the opinion that statistics was seriously needed for the study of medical problems, and his

contribution to statistics rests primarily on the compilation of bodies of material relating to

public health. Guy was a very active member of the Statistical Society of London and

because of his work on behalf of the Society, the Royal Statistical Society voted in 1891 to

establish in his honour the Guy medal. He was President of the Royal Statistical Society

from 1873 to 1875. Guy died on 10 September 1885, in London.

195

: Symbol for null hypothesis.

: Symbol for alternative hypothesis.

H aavelmo , Trygve M agnus ( 1 911^1 999): Haavelmo was born close to Oslo, Norway in 1911.

He went to the United States in 1939 as a Fulbright scholar where he remained until 1947

and was then appointed Professor of Economics and Statistics at the University of Oslo

in 1948. During his stay in the United States Haavelmo completed his highly inﬂuential

work on ‘The Probability Approach to Econometrics’ where he argued that economic data

should be envisioned as being samples selected by Nature, randomly drawn from hypo-

thetical distributions which governed reality but which were unobservable. Haavelmo was

awarded the Nobel Memorial Prize in Economics in 1989. He died in Oslo in 1999.

H aar wavelet: See wavelet analysis.

H ada ma rd matrix: An n×nmatrix H

consisting entirely of ±1s and such that

¼ diagðn; n; ...; nÞ. Important in

response surface methodology

.[Annals of

Statistics, 1978, 6, 1184–238.]

Haenszel, William (1910^1998): Born in Rochester, New York, Haenszel received a B.A.

degree in sociology and mathematics from the University of Buffalo in 1931, and an

M.A. degree in statistics in 1932. For the next 20 years he worked ﬁrst as a statistician

at the New York State Department of Health and then as Director of the Bureau of Vital

Statistics at the Connecticut State Department of Health. In 1952 Haenszel became Head

of the Biometry Section of the National Cancer Institute where he stayed until his

retirement in 1976. Haenszel made important contributions to epidemiology including

the

Mantel–Haenszel test

. He died on 13 March 1998, in Wheaton, Illinois.

Hajek, Jaroslav (1926^1974): Born in Podebrady, Czechoslovakia, Hajek studied statistical engi-

neering at the College of Special Sciences at the Czech Technical University and graduated in

1950. In 1963 he was awarded a doctorate for his thesis on statistical problems in

stochastic

processes

. In 1966 Hajek became professor in the Faculty of Mathematics and Physics of

Charles University. His major contributions to statistics were in the ﬁelds of stochastic

processes and rank tests. Hajek died on June 10th, 1974 in Prague, Czechoslovakia.

Hajnal, John (1924^20 08): Born in Darmstadt, Garmany, Hajnal was educated at a Quaker school in

Holland after his family left Germany in 1936. Later they moved to London where Hajnal

attended University College School, Hampsted and at 16 went on to Oxford to read politics,

philosophy and economics gaining a ﬁrst in 1943. In 1948 he was recruited to work on

demography for the UN in New York. In the early 1950s Hajnal left full-time employment to

study academic mathematics and in 1953 returned to England to become a medical statistician

in Manchester. In 1956 he obtained a lectureship at the London School of Economics eventually

becoming professor in 1975. His research interests were a mixture of the practical in demog-

raphy and theoretical on

Markov chains

and

stochastic processes

. Hajnal died on November

30th, 2008.

196

Hald, Anders (191 3^20 07): Born in Jutland, Denmark, Hald studied actuarial science and

statistics at the University of Copenhagen from where he graduated in 1939. In 1948 he

was awarded a D.Phil, his dissertation involving identifying a trend, seasonal variation

and noise in a

time series

. Also in 1948 Hald became professor of statistics in the Social

Science Faculty of the University of Copenhagen and in 1960 took up a new chair of

mathematical statistics at the university from where he developed mathematical statistics

in Denmark. His main contributions to statistics were in the areas of sampling plans and

the history of the discipline. Hald died on November 11th, 2007.

Haldane’s estimator: An estimator of the

odds ratio

given by

ψ ¼

ða þ

Þðd þ

ðb þ

Þðc þ

where a, b, c and d are the cell frequencies in the

two-by-two contingency table

of interest. See

also Jewell’s estimator.[Journal of the Royal Statistical Society, Series D, 2001, 50, 309–19.]

Half-mode: A term sometimes used for the mode of a probability distribution or a frequency

distribution if this occurs at an extreme value of the variable, for example, in the case of a

J-shaped distribution

Half-normal distribution: Synonym for folded normal distribution.

Half-normal plot: A plot for diagnosing model inadequacy or revealing the presence of

outliers

,in

which the absolute values of, for example, the

residuals

from a

multiple regression

are plotted

against the quantiles of the standard normal distribution. Outliers will appear at the top right of

the plot as points that are separated from the others, while systematic departures from a straight

line could indicate that the model is unsatisfactory. [Principles of Multivariate Analysis,2nd

edition, 2000, W. J. Krzanowski, Oxford Science Publications, Oxford.]

Halo effect: The tendency of a subject’s performance on some task to be overrated because of the

observer’s perception of the subject ‘doing well’ gained in an earlier exercise or when

assessed in a different area. [Journal of Applied Psychology, 1920, IV,25–9.]

Halperin, Max (191 7^198 8): Born in Omaha, Nebraska, Halperin graduated from the University of

Omaha in 1940 with a degree in mathematics. In 1950 he obtained a Ph.D. degree from the

University of North Carolina. His career began as a research mathematician at the

RAND corporation and posts at the National Institutes of Health and the Division of

Biologic Standards followed. Later he became Research Professor of Statistics and Director

of the Biostatistics Center of the Department of Statistics at the George Washington University.

Halperin made important contributions in many areas of statistics including

multivariate

analysis

, regression analysis,

multiple comparisons

and the detection of

outliers

.Hediedon

1 February 1988 in Fairfax, Virginia.

Hanging rootogram: A diagram comparing an observed

rootogram

with a ﬁtted curve, in which

differences between the two are displayed in relation to the horizontal axis, rather than to the

curve itself. This makes it easier to spot large differences and to look for patterns. An example is

given in Fig. 69.[Exploratory Data Analysis, 1977, J. W. T ukey, Addison-W esley, Reading, MA.]

Hankel matrix: A

variance-covariance matrix

between past and present values of a

time series

with some future values of the series at some time t.[Journal of the American Statistical

Association, 1998, 93, 770–82.]

Hannan, EdwardJames (1921^1994): Having obtained a commerce degree at the University of

Melbourne, Hannan began his career as an economist in the Reserve Bank of Australia. When

spending a year at the Australian National University in 1953 he was ‘spotted’ by

P. A. P. Moran

and began his research on various aspects of the analysis of time series which eventually

197

brought him international recognition. Hannan received many honours including the L y le

medal for research in mathematics and physics from the Australian Academy of Science and

the Pitman medal from the Statistical Society of Australia. He died on 7 January 1994.

Hansen ^ Hurwitz estimator: An unbiased estimator of the total size of a population given by

^τ ¼

i¼1

where n is the number of

sampling units

(often regions or areas in this context), y

is the number

of individuals, animals or species, observed in the ith sampling unit, and p

is the probability

of selecting the ith unit for i =1,2,..., N where N is the number of population units. (When

sampling is with replacement n=N.) An unbiased estimator of the variance of ^τ is

varð^τÞ¼

nðn 1Þ

i¼1

^τ



See also Horvitz–Thompson estimator.[Experimental and Ecological Statistics, 2003,

10,115–127.]

Hansen, Morris Howard (1910^199 0): Born in Thermopolis, Wyoming, Hansen studied

accounting at the University of Wyoming obtaining a B.S. degree in 1934. His formal

training in statistics consisted of after-hours classes taken at the Graduate School of the

U.S. Department of Agriculture. Hansen received a master’s degree in statistics in 1940.

In 1941 he joined the Census Bureau where he developed the mathematical theory

underlying sampling methods. This resulted in the publication in 1953 of the standard

reference work, Sample Survey Methods and Theory. Elected a member of the National

Academy of Sciences in 1976, Hansen died on 9 October 1990.

Haplotype: A combination of two or more

alleles

that are present in the same gamete. [Statistics in

Human Genetics, 1998, P. Sham, Arnold, London.]

Fig. 69 An example of a hanging rootogram.

198