Computerizing Enumeration:

Advantages and Disadvantages






	Ahmad Hleihel
	Data Processing Division

	Tel. 972-2-655 3482
	FAX 972-2-655 3531

Introduction:


	In each sub-regional office (SRO) a computer network was installed comprising half as many work stations as there were supervisors, i.e. each supervisor and his team of enumerators were allotted half a station. If there were 6-10 supervisors in an SRO, 3-5 work stations linked in a NOVELL network were set up. At times, based on the workload in an SRO, this rule was broken so that enumerators reporting to a supervisor would be allotted more than one- half station.
	The SRO network was linked up via external communications to two stations:
	1. to the relevant regional office (RO)
	2. to the Help Desk
	The objective of these link-ups was not to transfer data from the SRO to the RO or to the Help Desk, but to enable the RO and the Help Desk to solve problems at the SRO and to conduct on-line training sessions without the need for their physical presence in the SRO. The RO would override the local computer network and use their communications link-up to solve the problem or conduct the training required.
	The transfer of data from SROs to ROs was done by pick-up truck.

	The applications needed for computerization of the field setting were installed in the main server and in the SRO PCs. These applications include:
1.	Manpower application: registration of new workers, details of bank accounts, job descriptions, worker output calculations, etc. handled at the SRO.
2.	Institution and kibbutz enumeration: the enumeration of institutions and kibbutzim was carried out by a special system which required its own Enumerator's Report Book (ERB), and the separate processing of kibbutz data and institutional data.
3.	Regular enumeration: this comprised the computerization of the regular ERB and of the ERB for inhabitants outside localities.
	This application was divided by enumeration stages -- drop-off, pick-up, clean- up and closing of the Enumeration Area (EA).
	The three applications were linked together so that it was possible to transfer data between all applications. For example, to calculate an enumerator's wages required data from his ERB.

	This paper discusses the process by which the regular enumeration was computerized.
	The work procedure of manpower administration and institutional enumeration shall not be discussed in this paper.


	Work Procedure:
	Each time an enumerator was due to report to his supervisor, he had first to key in the ERB data added since his previous report to the supervisor. The output of the newly keyed-in data included a cumulative follow-up report and an analysis of the quality of the enumerator's work. The enumerator would then take these reports into the meeting with his supervisor.
	The supervisor would inspect the computer reports together with the manually completed ERB and the questionnaires collected by the enumerator. The follow- up reports were filed with the supervisor immediately after the meeting. The other reports, for the purpose of evaluating quality of enumerator ERB recording, required the supervisor's written comments and reaction, which then had to be at returned to one of the keying stations in order to update the computerized ERB as per the supervisor's comments.

	Work Procedure At Each Enumeration Stage:
	As mentioned above, the ERB application was divided by enumeration stage, as follows:
4.	Drop-off
5.	Pick-up
6.	Clean-up
7.	Closing of the Enumeration Area.
	A brief description of the work procedure at each of the four stages follows (you may find it useful to refer to the specimen Visitation Record Page (VRP) provided)

	1. Work Procedure at Drop-off Stage:
	At the drop-off stage, the enumerator would manually record in the VRP in his Report Book the data on the buildings, dwellings and households he has visited, keeping to their order on his enumeration route.
	When he brings his ERB to be keyed in, the keying operator, who has a simulation of the VRP on his/her screen, would key in the data from columns (1) to (9).
	Once the data have been keyed in, output, including a "Incorrect Records in ERB" report (marked R2 ), would be generated. This report lists every line containing a logic error by the enumerator with a description of the type of error. In the enumerator's meeting with his supervisor, the latter would manually correct the mistakes and return the report to the keying operator for correction of the computerized ERB.

	On conclusion of the drop-off stage, or as soon as an enumerator has completed drop-off in his area, the keying operator would be asked to run a program comparing the addresses registered by the enumerator in his ERB with those in the computer Mapped Address file. The program outputs a report of all "Address Coverage Check" (marked R5 ) showing the discrepancies between the two address sources, which indicates the possibility that the enumerator has not covered his entire route. If this proves to be the case, the enumerator would be required to return and drop off questionnaires at the missing addresses.

	2. Work Procedure at Pick-up Stage:
	At the pick-up stage, the enumerator returns to the dwellings he visited at the drop-off stage and either fills out a questionnaire or classifies them as dwellings for which a questionnaire does not have to be filled out (empty dwelling, non-residential dwelling or building, dwelling whose residents are not subject to enumeration). At the visit, the enumerator filled in columns (11) to (21) of the VRP for every row filled in at the drop-off stage.
	Each time the enumerator is due to meet with his supervisor, the keying operator would key in the data added by the enumerator since his previous meeting with the supervisor at the pick-up stage (in cols. (11) to (21)). Four reports are obtained as a result of this keying:
	"Records for Supervisor" (marked R1)
	"Incorrect Records" (marked R2 )
	"Refusals Record" (marked R3 )
	"Total and Turnouts" (marked R4 )
	Report R1 lists those lines for which a supervisor decision on further treatment by the enumerator is required. The supervisor records his decision in col. (22) and adds the following data from the first page of the questionnaire:

Household classification in Col.(24)


	Total number of individuals in Col.(25) *
	Type of questionnaire in Col. (26) *
	Report R2 lists the rows containing a logic error entered by the enumerator in his ERB. These errors are corrected by the supervisor on the hard copy of the output.
	After the meeting, the supervisor sends Reports R1 and R2 to the keying operator for keying the new data into the computer. The reaction to these data would be expressed at the next meeting between the enumerator and the supervisor.
	Report R3 lists those households that, according to the enumerator's records refused to fill out a questionnaire: this listing is retained by the supervisor for follow-up.
	Report R4 includes summaries and completion of work regarding cases handled by the enumerator to date. Its purpose is to monitor the output of each enumerator, which helps to prevent delays during the course of the work and to detect them at an early enough stage so that they can be overcome.
	At the end of the pick-up stage the supervisor prepares the manual and computerized ERBs for clean-up stage enumeration. This preparatory work
	of the supervisor is necessary since at the clean-up stage the enumerator handles only specific cases, as described below.

3. Work Procedure at the Clean-up Stage


	The clean-up stage is intended to complete work left unfinished from the pick-up stage for one of the following reasons;
1.	The enumerator did not complete work on certain cases;
2.	The household refused, at the pick-up stage, to fill out a questionnaire;
3.	The household lived in a dwelling that was closed at the pick-up stage.

From the point of view of computerized work procedures, the clean-up stage is very
similar to the pick-up stage.


	Three reports are issued at this stage --R1, R2 , and R4. The structure of R1 and R2 is similar to that of the equivalent reports at the pick-up stage; Report R4, however, has a different structure although it fulfills the same purpose as Report R4 at the pick-up stage.


	___________
	*Note: Cols. (24) to (26) do not appear in the manual ERB but only in the computerized ERB. Their purpose is to check the match-up between the first page of the questionnaire and the ERB.

4. Conclusion of Work in Enumeration Area

The conclusion of work in the enumeration area (EA) is characterized by the
generation of three summary reports:


	"Incorrect Records" (marked R2 )
	"EA Production Turnout" (marked R7 )
	"ERB Table F" (marked R6 )

At the EA closure stage, the number of logic errors in Report R2 decreased to
only those critical ones which, if left uncorrected, will not allow the application to
permit closure of the EA.

Report R6 provides a summation of the ERB corresponding to the summation at the
bottom of each VRP, and summarized in Table 6 at the end of the manual ERB.

Once he has corrected all errors and approved the ERB summation, the supervisor
requests a keying operator to correct the computerized ERB and generate an
"Output" report (R7). This report constitutes the basis for calculating the
enumerator's wages.

The Advantages of Computerized Work
Procedures

There is no doubt that computerization of the fieldwork improved various aspects of
both implementation and data quality. Had we remained with the former manual method
of operation used in the previous census in Israel, we would never have reached
the levels of performance achieved.

The advantages and contributions of the methodology, first used in the 1995
census, are as follows:

Improved Coverage: This was felt during the procedure which was carried out at
the conclusion of the drop-off stage. A computer file had been prepared earlier of
all the addresses recorded on the Census map. During the drop-off stage, the
enumerator recorded every address he visited in his manual ERB; these addresses
were then keyed into his computerized ERB file. At the conclusion of the drop-off
stage, this list was compared with the list drawn from the Census map; this
comparison helped to reveal those addresses in the EA that the enumerator missed
and those addresses which he had visited but belong to other EAs.

The discrepancies between the two lists were examined and verified by the
supervisor. If a discrepancy occurred due to enumerator error, the enumerator
was requested by the supervisor to correct the error.

Improved Supervisor Control of Enumerators: The numerous computer-generated
reports used to monitor enumerator progress and discussed at the enumerator-
supervisor meetings replaced much of the work that supervisors formerly did
themselves manually. This freed the supervisor to devote more time to other
aspects of their duties and to execute them more efficiently, such tasks as more
thorough inspection of completed questionnaires, dealing with households who
refused to fill out a questionnaire, and the like.

If all the computer-generated checks and updating of data had had to be done
manually by supervisors, then many of them would simply not have been done. The
supervisor's work benefited from the greater precision of computer-generated work
and the indispensability of frequent data updates.

Uniformity of Enumeration: Computerization ensured uniformity of enumeration in
every census office in the country, something which should be a feature of every
census, but is virtually impossible to control in a non-computerized procedure. This
uniformity was achieved by using identical computerized applications in every sub-
regional office and by the Help Desk giving identical answers to the same problems,
wherever they occurred.

Questionnaire Quality: There is no doubt that computerization freed supervisor time
for more thorough checks of completed questionnaires, which in turn raised the
quality of the data eventually transferred to computer files.

Keying the data from the first page of the questionnaire into the computerized ERB
also contributed to improving the match-up between the data obtained from the ERB
and the data taken from the first page of the questionnaire. This was another
positive factor leading to better questionnaire quality.

Improved Oversight of Fieldwork by Senior Management: At the end of every
working day in every SRO, the data newly added that day were backed up onto
diskettes; the next day, they were transported by van to the Regional Office. The
data on the diskettes were transferred to the ROs computer, and from there
transferred via telephone line to CHQ at the end of every working day. Thus,
senior management could monitor fieldwork in almost real time (the maximum delay was
two days). This permitted the early detection of delays or other problems in the
SROs, and granted senior managers the rapid response capability needed to
adequately handle most of the problems and delays which were detected.

The Availability of a High-Quality ERB Data File: Immediately upon conclusion of the
census, a high-quality computerized ERB was available for use. This permitted the
rapid publication of ERB-based high-quality interim population and household data
reports.

The availability of a detailed computerized ERB file always permitted easy access
for the purpose of conducting all manner of checks and comparisons with the
questionnaire data.

The Disadvantages of Enumeration
Computerization

The following disadvantages were the result of computerized enumeration:

Difficulty in Changing/Correcting Procedures: Field application was carried out in
accordance with procedures defined by the census planning team. Problems arose
in certain EAs that could have been solved by making a modification or change in
work procedure applied in a particular EA. At times, changes were not possible
because computer-dependent solutions, such as correcting an applications problem or
changing a work procedure in a certain EA, could not be implemented in a uniform
fashion in every one of the census offices.

Over-dependence on the Computer: Computerization of enumeration led to over-
dependence on the part of workers in SROs. Although each computerized procedure
had a manual counterpart, the workers did not observe the alternative manual
procedures but, instead, operated according to the slogan, "If the computer doesn't
work, neither do we".

This state of affairs occasionally gave rise to work delays.

Conclusion

The disadvantages resulting from computerization of enumeration do not outweigh
even a small portion of the many advantages.

In the final analysis, computerization met all the census planners' main objectives and
targets which they set up for the system:


	- improved coverage of the census population;
	- high-quality completed questionnaires;
	- a detailed ERB data file of adequate quality.