Page 318 - Special Topic Session (STS) - Volume 2

STS496 Alphonse L.M.
tradition. Since its independence in the early 19th century Chile systematically
carried out national population censuses, from1835 and since then roughly
every ten years. The rate of under-enumeration of the Chilean census during
the period 1950 – 1990 in general was below the regional average. But in the
2000 census round the under-enumeration in Chile was 3.8 % while the
regional average 3. 3 %. Up to the 2000 census round the census in Chile was
carried out in one day by a large group of volunteers on the basis of the de
facto concept, using a basic questionnaire. For the 2010 census round this was
changed to a census carried out by paid enumerators, using the de jure
concept, with a complex questionnaire and a multi-week field enumeration
period.
Fieldwork for the 2012 census was carried out from 9 April to the end July
2012; the results of the census were released on 2 April 2013; on 26 April 2013
the Director of the National Statistical Office (INE) resigned, to be replaced on
29 April 2013, and on 2 May 2013 the results of the 2012 were withdrawn. The
resignation of the Director and the withdrawal of the published census results
were caused by the allegations by senior INE staff members that the data had
been manipulated on the instructions of the Director. Bezama y González, M.
(2013), Economist (2013)
Two review commissions, one national and the other international, were
appointed to investigate the allegations and assess the quality of the census.
Their findings were as follows: The census methodology was not correctly
applied due to a dysfunctional atmosphere in INE, causing the marginalisation
of knowledgeable technical staff; lack of understanding of the de jure concept,
insufficient time for the preparation and training of staff, and lack of adequate
funding. The census did not cover about 9.3 % of the population, either by not
contacting the households, or because dwellings were wrongly classified as
unoccupied, or because of refusals. The missing data were imputed. The non-
coverage could be higher in particular geographical areas, administrative
regions, communities, or in specific social categories. Coverage of the census
population could not be properly assessed because the Post Enumeration
Survey (PES) was not properly designed and executed and was only attempted
in urban areas. Training of the field staff was uneven, and inadequate. The
census reference date was not clearly identified or used. The documentation
on the census procedures was defective; there were no methodological or
administrative reports. (INE (2014)

5. Conclusions and way forward
“Statisticians misbehaving” appear to be rare events. Since the second half
of the 20th century the scope of statistics has greatly expanded and currently
it is fashionable to refer to “data and statistics”. The term “statistician” not
only refers to persons with formal degrees in statistics, but also to persons

307 | I S I W S C 2 0 1 9