Semakan pada 05:03, 20 Mei 2013

220

Proses untuk memperolehi data statistik daripada dokumen yang tidak dibangunkan untuk tujuan statistik dipanggil pengekstrakan¹. Umumnya, apapun sumbernya, maklumat statistik adalah tertakluk kepada pemprosesan² sama ada secara manual³, mekanikal⁴, elektronik⁵ atau kombinasi kaedah-kaedah ini. Pemprosesan manual tidak melibatkan peralatan yang lebih kompleks daripada kalkulator⁶. Pemprosesan mekanikal menggunakan mesin-mesin penjadualan (224-2) atau mesin-mesin punch card (224-3); pemprosesan elektronik menggunakan komputer (132-2*). Tanpa mengambilkira kaedah pemprosesan, certain jenis-jenis operasi⁷ tertentu perlu dilakukan termasuk menyunting⁸ data, penjadualan (130-6*) dan pengiraan (132-2) dan penyediaan jadual⁹. Kerumitan operasi-operasi ini adalah bergantung kepada kaedah pemprosesan yang digunakan.

221

Penyuntingan data kebiasaannya memerlukan proses pengekodan¹ bagi beberapa entri pada dokumen asas². Skim pengekodan³ menguhubungkan setiap entri dengan keterangannya melalui kod nombor atau huruf tertentu. Buku kod mengumpulkan dan menerangkan kesemua skim pengekodan. Skim pengekodan dibangunkan untuk memudahkan pembahagian data mengikut kumpulan-kumpulan tertentu. Sebaliknya, klasifikasi⁴ adalah semata-mata kod individu di mana setiap

establishes a correspondence between an entry and its translation into numeric or alphabetic codes. The code book collects and describes the coding schemes used with a particular set of basic documents. A coding scheme is usually designed to facilitate later groupings of the data. In contrast, a classification⁴ is a mere list of individual codes where each heading⁵ is given one or several numbers. After the data have been coded, they constitute a file (213-3*) which can be converted to machine readable form. The second stage in the editing consists in the cleaning⁶ of the file, through elimination of errors by validity checks⁷ and consistency checks⁷; these can be internal checks within each statistical unit (cf. 110-1) or may result from the comparison of different units. After errors have been identified, they may be corrected in the original document or the file by some automatic procedure.

222

The edited data are rarely used directly; they are subjected to grouping (130-7) and tabulation (130-6*), and this normally leads to a presentation in the form of statistical tables (131-4). These may be the outcome of sorting¹, either manual or mechanical, resulting in the reorganization of the elements in a set according to predetermined rules, or more simply of a systematic count of the elements presenting a selected characteristic. The choice of elements or of characteristics may be based on the values of one or several quantitative attributes, or on the modalities² of one or several qualitative attributes. Few studies can do without computation, simple or complex, isolated or repetitive, and the computer (225-2) now allows calculation that would have been too lengthy by hand. These capabilities have led to the development of techniques of data analysis³. Deterministic and stochastic models (cf. 730) often require considerable computations, and so do simulations (730-6).

223

The stage of table preparation (220-9) aims at making the results of processing conveniently available in the form of listings¹, numerical tables (131-4) or charts (155-2), all of which are commonly used in descriptive statistics². The use of computer graphing³ and computer cartography³ permits the mass production of graphical presentation at a preliminary stage.

224

Purely mechanical processing (220-4) did not involve the use of electronic equipment¹ which has come to replace the earlier tabulating machines² or unit record machines² and is much more versatile. In most instances the information is coded (221-1*) first, and then transcribed onto punch cards³ by using a keypunch⁴. A card verifier⁵ is a device used to check the accuracy of the punching. These two types of unit record machines remain in common use since the punch cards are still a frequent way of entering data into the computer. The use of other types of unit record equipment such as the card sorter⁶ and the tabulator⁷ has declined. Increasingly, the data are entered directly on magnetic tapes (cf. 226-4) or disks (cf. 226-5) without resorting to punch cards.

225

Demographic research depends heavily on electronic data processing¹ using the computer². The term hardware³ refers to the physical component, whereas software⁴ supplies the user⁵ with ways to have access to the computer. Computer specialists⁶ include programmers⁷ who write programs⁸ conceived by system analysts⁹.

226

The hardware (225-3) components of a computer (225-2) include one or several central processing units¹, a central memory², one or more mass storage devices³ which use magnetic tapes⁴ or disks⁵ and a set of input - output devices⁶. The software (225-4) components include the operating system⁷, which has the task of efficiently managing the available facilities⁸ for the users (225-5) running the users’ programs⁹ and the processing programs¹⁰ which are preestablished programs (225-8) designed for the solution of standard problems.

227

A user (225-5) can process his problem by writing a program (225-8) in a general programming language¹ such as Fortran, Cobol, Basic or Algol, or a specific language, designed to use the processing programs (226-9) stored in the central memory (226-2) of the computer such as a data base management system² used to create and maintain a data bank², a survey processing program³ or a statistical package⁴. The devices which are used to enter and receive information from the computer can differ according to the mode of processing. In batch processing⁷, the normal input and output units are the card reader⁵ and the line printer⁶. A console⁸ is the normal input and output unit for processing in a timesharing mode⁹. In either instance the entry units may be spatially separated from the computer and processing under these conditions is accomplished by remote terminal¹⁰.

228

Any information processed in a computer (225-2) undergoes three main phases. First, data entry¹ or input¹ which may be done by using punched cards (224-3*) or by using an on line² device such as a keyboard console (227-8). Data which is already stored in the computer may be accessed from either central memory (226-2) or from one of the mass storage devices (226-3) and used as input data. The second phase, processing (220-2), is divided into two main types: numerical processing⁴ and non-numerical processing⁵. Statistical or arithmetic computations are normally the operations contained in the former while data manipulation operations are the focus in the latter. In a third phase, occasionally referred to as output phase, the processed results⁶ or output⁶ may be printed out on the line printer (227-6) or saved as a file on a mass storage device (226-3) for further processing. Results may also be diverted to a plotter⁷ to obtain processed results in the form of a graph or a figure.

* * *