Count

Author: Zheng Congling
Publisher:
Publish Date: 2004-01-01
Features: Section 4: Summarization and Tabulation of Statistical Data
I. Organizational Forms and Techniques of Statistical Data Summarization
(1) Organizational Forms of Statistical Data Summarization
A large portion of statistical data processing involves summarizing the totals of various groups and the overall totals, known as statistical summarization. The organizational forms of statistical data processing are essentially the same as those of statistical summarization. Since statistical summarization is a very labor-intensive task, a scientific and comprehensive organizational form is necessary to ensure its smooth execution. The basic organizational forms of statistical data summarization include hierarchical summarization and centralized summarization. Combining these two forms results in comprehensive summarization. Hierarchical summarization is the most commonly used form.
1. Hierarchical summarization refers to the method of summarizing survey data in a top-down manner according to a certain statistical management system. China's statistical reporting system adopts this form. Its advantages include meeting the needs of various regions and departments for data, facilitating on-site verification and correction of raw data. However, its disadvantages are that it involves multiple levels of summarization, which is time-consuming and prone to errors.
2. Centralized summarization involves concentrating all raw data in the highest organizational agency or a designated institution for summarization. This form is often used for rapid censuses with high timeliness requirements and important surveys with stringent summarization demands. Its advantages include eliminating intermediate steps and utilizing modern summarization techniques to improve efficiency and quality. However, it cannot promptly meet the needs of local or grassroots leaders and makes it difficult to verify and correct data.
3. Comprehensive summarization refers to the method of summarizing basic data required by all levels hierarchically, while other survey data is summarized centrally. China's third national census adopted this form. Data such as the total number of households and population, as well as population data grouped by gender, ethnicity, and education level, were summarized hierarchically. Other census data were centrally summarized by provincial and central statistical agencies using computers. This organizational form meets the needs of all levels while saving time.
(2) Techniques of Statistical Data Summarization
To ensure accurate, fast, and efficient statistical summarization while saving human and material resources, it is essential to adopt appropriate summarization techniques. Common techniques include manual summarization and computer-assisted summarization.
1. Manual Summarization
The main methods of manual summarization are as follows:
(1) Tallying Method. This involves marking points or lines on a pre-designed summary table as indicators. It is suitable for summarizing the number of total units. When summarizing, determine which group a total unit belongs to and mark a point or line in the corresponding group on the summary table. Finally, count the number of points or lines in each group to obtain the unit count. Common tally symbols include "||," "≠," and "※." The tallying method is simple but can only summarize unit counts, not indicator values. Excessive marking may lead to errors, so this method is generally used only when the number of total units is small and there is no need to summarize indicator values.
(2) Recording Method. First, transfer the survey data to a pre-designed summary table, then calculate and sum up the totals for each group and the overall totals and indicator values, and finally fill them into the statistical table. The recording method can summarize both unit counts and indicator values and is convenient for verification and calculation. However, it is time-consuming and prone to errors if there are many recording items. Therefore, this method is more suitable when the number of total units is small and the grouping is simple.
(3) Folding Method. This involves folding the values of the same item on the survey forms along a line for summarization and directly entering the results into the statistical table. This method is suitable for summarizing indicator values and is simple and easy to implement without requiring a pre-designed summary table. It is widely used by statisticians. The disadvantage is that if an error is found during summarization, the entire process must be repeated, and it is difficult to identify the cause of the error.
(4) Card Method. This involves using special extraction cards as tools for grouping and counting. When survey data is extensive and grouping is detailed, the card method is more accurate than tallying and simpler than recording and folding methods, ensuring higher summarization quality and efficiency. The card method is generally used for processing large-scale specialized survey data. If the survey data is limited, this method may be uneconomical.
2. Computer-Assisted Summarization
Using modern computer technology for statistical summarization and calculation is a new development in summarization techniques and an important indicator of statistical modernization. The process of computer-assisted summarization can be divided into the following five steps:
(1) Programming. Based on the summarization plan, develop computer programs for statistical grouping, summarization, and table creation. Standardized summarization programs can be stored and reused multiple times. Such pre-prepared computer programs are generally referred to as software packages, and the computer operates according to the programmed instructions.
(2) Coding. This involves converting a symbol system that represents information into another system that is easier for computers or humans to recognize and process. Summarization information includes both numerical and textual data. Coding is the process of converting textual information into numerical information. For example, assign appropriate codes to the groups to be summarized and indicator names. The quality of coding not only affects the speed and accuracy of data entry but also impacts the results of data processing.
(3) Data Entry. This involves recording the coded data and actual numerical data into storage media (e.g., magnetic tapes or disks) by data entry personnel using input devices. The computer then converts these data into electromagnetic signals that it can recognize through its own devices.
(4) Data Editing. This involves the computer checking the input data according to a set of pre-defined editing rules. If a group of data exceeds the allowable error range, it is returned for re-examination and correction. Individual errors within the allowable range are corrected according to the editing program rules. The key to editing effectiveness lies in whether the established editing rules are reasonable.
(5) Table Creation and Printing. After all data has been edited, the computer creates statistical tables according to the pre-defined summary table format and summarization hierarchy and prints the results through an output device. Computer-assisted summarization not only significantly reduces manual labor but also brings major changes to the entire statistical process.