The Dictionary of Human Geography

Comp. by: VPugazhenthi Stage : Revises1 ChapterID: 9781405132879_4_S Date:1/4/
09 Time:15:23:36 Filepath:H:/00_Blackwell/00_3B2/Gregory-9781405132879/appln/3B2/
revises/9781405132879_4_S.3d

slash and burn, and swiddening), shifting

cultivation is widespread throughout the

humid tropics, but was also practiced in tem-

perateeurope until the nineteenth century

(and sometimes later) (Conklin, 1962). It is

estimated that there are over 250 million shift-

ing cultivators world-wide, with 100 million

in South-East Asia alone. Shifting cultivation

is enormously heterogeneous and subtypes

can be distinguished according to crops rai-

sed, crop associations and successions, fallow

lengths, climatic and soil conditions, field

technologies, soil treatment and themobility

of settlement. Many distinguish betweeninte-

gral(shifting cultivation as an integral part of

subsistence) andpartial(shifting cultivation

as a technological expedient for cash cropping;

see peasant) forms of shifting cultivation

(Conklin, 1962). In all shifting cultivation

systems, the burning of cleared vegetation is

critical to the release of nutrients, which

ensures field productivity. Shifting cultivation

by definition is land-extensive, and is threa-

tened by population growth andmigration,

and expanding land occupation (seecarrying

capacity;intensive agriculture). Shifting

cultivation is often stigmatized as primitive or

unproductive, but it is typically predicated

upon a sophisticated understanding of local

environmental conditions – so-calledindigen-

ous knowledge– and exhibits considerable

flexibility and adaptiveness. mw

significance test A statistical procedure

for identifying the probability that an observed

event could have occurred by chance. Most

statistics – such as acorrelationcoefficient –

have an associated sampling distribution

identifying all possible values of the coefficient

for a given set of empirical observations. For

example, three observations can be rank-

ordered in six different ways, so that if one is

correlating the rank-order of neighbourhood

death rates with the local levels of lead pollu-

tion, there would be 36 different ways in which

the two sets of three observations could be

organized, each producing a different correl-

ation coefficient. Those coefficients form a

frequency distributionin which some val-

ues would appear quite frequently, others

more rarely. If the observed coefficient’s value

would occur only rarely, this outcome was

unlikely to occur by chance: it is then pre-

sented as a significant result, as a correlation

with interpretative value.

In studies involving large numbers of obser-

vations – 80 neighbourhoods instead of three,

say – the frequency distribution can be

predetermined and the observed statistic

located within it. Most analysts operate the

rule of thumb that if an observed outcome is

likely to occur less frequently than five times in

every 100 ways in which the data can be

organized, that is a statistically significant

result. Statistical significance should not be

confused with substantive importance, how-

ever; it says nothing about the importance of

findings and how they should be interpreted.

Statistical significance tests are used in two

ways. Inconfirmatory data analysis, they

assisthypothesistesting, indicating whether

an observed outcome in a sample (seesam-

pling) is likely to characterize the population

from which it was drawn. (If the 80 neigh-

bourhoods were a sample from 240 within

a city, a correlation coefficient statistically

significant at the 0.05 level would indicate

that the relationship between the two variables

in the sampled areas would almost certainly

be found if the whole population had been

studied.) In exploratory data analysis,

there is no sample involved and the test is

used to indicate the likely importance of an

observed relationship occurring: one that

would occur three times in every ten ways in

which the data are organized is probably less

important than one that would occur only

once in every 500 ways. rj

Suggested reading

Blalock (1970).Seealsohttp://bowland-files.lancs.

ac.uk/monkey/ihe/linguistics/corpus3/3sig.htm.

simulacrum A term originally derived from

Platonic theory, but popularized by Jean

Baudrillard. Literally, the simulacrum is a copy

without an original. Baudrillard extended the

argument about the modern dominance of

imagesandideologyby suggesting an histor-

icalprocessionof the simulacra. First, there was

presence – people met and spoke – then tech-

nologies of writing allowed re-presentation of

those who were not present. Now media allow

the simulation of those who never existed. The

term gained wide usage in discussions of

themed environments such as Disneyland or

shopping malls. ‘Main street’, Disneyland

thus simulates a mythic idea of a small town

community hub that has no counterpoint in

reality (seedisneyfication). mc

Suggested reading

Baudrillard (1983).

simulation Aheuristicdevice for solving

otherwise intractable mathematical and

Gregory / The Dictionary of Human Geography 9781405132879_4_S Final Proof page 682 1.4.2009 3:23pm

SIGNIFICANCE TEST