Page 24 - 1960 (3) Pakistan Census Of Agriculture Vol 2 West Pakistan

4 tabulators, 2 sorters, and 6 summary punches, employ- ships and patterns, in the sphere of agriculture, and
ing 350 persons. The Processing of data included the the derivation of the Census Totals of different items
following stages of work :- was only incidental. The importance of Totals was
further reduced by the fact that the Census was based
(i) EDITING AND CODING: The question- on a sample. In the matter of inter-relationship of"
naires were thoroughly checked to make various factors, the values produced in a survey as
sure that these were complete in all respects; big as the present Census could be safely considered
that the entries were consistent and legible as stable in spite of the errors that might show up in
and the totals were correct. After editing, the Totals.
some of the entries were given code numbers
to represent the qualitative answers. 43. The Census Organization has endeavoured_
to form an objective evaluation of the quality of the-
(ii) CARD PUNCHING AND VERIFICA- Census data, which appeared to have been somewhat
TION: The information contained in the affected by some significant factors such as:-
Census questionnaires was transferred to
punch cards of which three were used for (i) The errors existing in the Revenue records;
each respondent. The cards were duly which formed the primary source of in for-
verified to see that the information had mation in case of several items of the Census;·
been correctly punched.
(ii) Errors of transcription introduced at the,
(iii) MACHINE EDITING: The punched cards stage of extraction of information from
were edited with the help of an electronic Revenue records;
machine which carried out final checks about
the consistency of the information in the (iii) Errors of enumeration at the time of inter-
carcls, the correctness of the totals and the . viewing the farmers for the sake of Part R
completeness of the entries. Any incorrect of the questionnaire; and
cards found at this stage were substituted
by correct cards. (iv) Sampling error.
(iv) MACHINE TABULATION: The cards 44. A quality check of the revenue records was.
were passed through the tabulating machines made in a number of check surveys conducted after
and the information was cross-related and the Census enumeration. The information collected
summarized in the prescribed tables.
in these surveys was not considered adequate for
obtaining quantitative estimates of different types of"
(v) QUALITY CONTROL: The tabulations errors existing in the revenue records. Therefore, a
obtained from the machines were subjected further intensive investigation in that direction will
to a final quality check.
have to be conducted if the same records were to be·
used in the next Census.
(vi) RAISING OF THE SAMPLE ESTI-
MATES: The sample estimates of tehsils/ 45. The second factor affecting the quality or·
talukas were raised by the Ratio Estimates
method using appropriate ratios. Multipli- the Census data was the incidence of errors of enumera-
cation by the raising factors was done with tion inherent · in every census, whether complete or·
the help of the IBM Calculating Punch. sample. But in a sample Census such errors could_
be controlled to some extent by good planning, good'.
training and intensive control and supervision of the,
(vii) PREPARATION OF FINAL TABLES :
The final tables were prepared on the operations. Nevertheless, errors may be caused
tabulators from the summary cards. either by faulty wording of questions or the personal
bias of the enumerators or misunderstanding of the
Quality of the Census Data instructions. It may also- happen that the farmer
respondents, due to illiteracy might not be able to-
41. The Census data are to form the basis of supply proper answers, or be unwilling to give correct.
important decisions in different fields of national answers on account of misapprehension about the
activity such as agricultural development plans, land objectives of the Census. Along with that the pos-
reforms, farm price policies, farm mechanization pro- sibilities of omission, double counting and incorrect.
gramme, regulation of animal slaughter, calculation of recording are always there. Besides, when information
national income, etc. The Census Organization, there- is extracted from the records some transcription errors.
fore, had to take a number of steps to ensure high are unavoidable.
quality of the data, but inspite of all possible controls,
some inaccuracies were unavoidable on account of 46. Part A of the Census questionnaire pertaining;
the vast dimensions of the Census operations and its to the number and size of farms, tenure, land use and
multifarious problems of complicated nature. cropping was filled in by extracting relevant information.
from the Revenue records. Checking done in the Post
42. While assessing the quality of the Census Enumeration Survey had established that the Census
data it may be emphasized that the main objective enumerators hg_d omitted good many farms that should
of the Census was to discover important inter-relation- have been enumerated. That factor had influenced in a.