The 1997 Viemam Demographic and Health Survey (VNDHS-II) is a nationally representative survey of 5,664 ever-married women age 15-49 selected from 205 sampling clusters throughout Vietnam. The VNDHS-II was designed to provide information on levels of fertility, family planning knowledge and use, infant and child mortality, and indicators of maternal and child health. The survey included a Community/Health Facility Questionnaire that was implemented in each of the sample clusters included in the women's survey. Fieldwork for the survey took place from July to October 1997. All provinces were separated into "project" and "non-project" groups to permit separate estimates for about one-third of provinces where the health infrastructure is being upgraded.
The primary objectives of the second Vietnam National Demographic and Health Survey (VNDHS-II) in 1997 were to provide up-to-date information on fertility levels, fertility preferences, awareness and use of family planning methods, breastfeeding practices, early childhood mortality, child health and knowledge of AIDS.
VNDHS-II data confirm the patterns of declining fertility and increasing use of contraception that were observed between the 1988 VNDHS-I and the 1994 lntercensal Demographic Survey (ICDS-94).
Kind of data
Sample survey data
The 1997 Viemam Demographic and Health Survey (VNDHS-II) is a nationally representative survey. Itwas designed to provide separate estimates for the whole country, for urban and rural areas, for 18 project provinces, and for the remaining non-project provinces as well. Project provinces refer to 18 focus provinces targeted for the strengthening of their primary health care systems by the Government's Population and Family Health Project to be implemented over a period of seven years, from 1996 to 2002 (At the outset of this project there were 15 focus provinces, which became 18 by the creation of 3 new provinces from the initial set of 15). These provinces were selected according to criteria based on relatively low health and family planning status, no substantial family planning donor presence, and regional spread. These criteria resulted in the selection of the country's poorer provinces. Nine of these provinces have significant proportions of ethnic minorities among their population.
Unit of analysis
- Women age 15-49
The population covered by the 1997 VNDHS is defined as the universe of all women age 15-49 in Vietnam.
Producers and sponsors
National Committee for Population and Family Planning
Macro International Inc.
General Statistical Office
Futures Group International Inc.
The Second Vietnam Demographic and Health Survey (VNDHS-1I) covers the population residing in private households in the country. The design for the VNDHS-II calls for a representative probability sample of approximately 5,500 completed individual interviews of ever-married women between the ages of 15 and 49. It was designed principally to produce reliable estimates of demographic rates (particularly fertility and childhood mortality rates), of maternal and child health indicators, and of contraceptive knowledge and use, for the country as a whole, for urban and the rural areas separately, and for the group of 18 project provinces. These 18 provinces are in the following geographic regions:
Six of the 18 project provinces are new provinces that will, in the near future, be formed out of three old provinces: Bac Can and Thai, Nguyen from Bac Thai; Hai Duong and Hung Yen from Hal Hung; Nam Dinh and Ha Nam from Nam Ha.
Northern Uplands: Tuyen Quang, Lai Chau, Lao Cai, Bac Can and Thai Nguyen.
Red River Delta: Hai Phong, Hai Duong, Hung Yen, Nam Dinh and Ha Nam.
North Central: Thanh Hoa and Thua Thien-Hue.
Central Highlands: Dac Lac and Lam Dong.
Mekong River Delta: Dong Thap, Vinh Long, Tra Vinh and Kien Giang.
Since the formation of the new provinces has not been formalized and no population data exist for them, this report will only refer to 15 project provinces out of 53 provinces in Vietnam (instead of 18 project provinces out of 61 provinces).
The sampling frame for the VNDHS-II was the sample of the 1996 Vietnam Multi-Round Survey (VNMRS), conducted bi-annually by the General Statistical Office (GSO). A thorough evaluation of this sample was necessary to ensure that the sample was representative of the country, before it was used for the VNDHS-II.
The sample design for the VNMRS was developed by GSO with technical assistance provided by Mr. Anthony Turney, sampling specialist of the United Nations Statistics Division. The VNMRS sample was stratified and selected in two stages. Within each province, stratification was geographic by urban- rural residence. Sample selection was done independently for each province.
In the first stage, primary sampling units (PSUs) corresponding to communes (rural areas) and blocks (urban areas) were selected using equal probability systematic random selection (EPSEM), since no recent population data on communes and blocks existed that could be used for selection with probability proportional to size. The assumption underlying the decision to use EPSEM was that, within each province, the majority of communes and blocks vary little in population size, with the exception of a few communes; i.e., within each province, most communes and blocks have a population size that is close to the average for the province. In each province, the number of selected communes/blocks was proportional to the urban-rural population in the province. The total number of communes/blocks selected for the VNMRS was 1,662 with tbe number of communes/blocks in each province varying from 26 to 43 according to the size of the province. After the communes/blocks were selected, a field operation was mounted by GSO to create enumeration areas (EAs) in each selected commune/block. The number of EAs that was created in each commune/block was based on the number of households in the commune/block divided by the standard EA size which was set at 150 households. The list of EAs for the whole province was then ordered geographically by commune/block and used for the second stage selection. Thirty EAs were selected in each province with equal probability from a random start, for a total of 1,590 EAs. Because of this method of systematic random selection, communes/blocks that were large in size had one or rnore EAs selected into the sample while communes/blocks that were very small in size were excluded from the sample. In each selected EA, all households were interviewed for the VNMRS.
To evaluate the representativity of the VNMRS, EA weights were calculated based on the selection probability at tile various sampling stages of the VNMRS: also, the percent distribution of households in the VNMRS across urban/rural strata and provinces was estimated and compared with the percent distribution of the 1996 population across the same strata. The distribution obtaiued from the VNMRS agrees closely with that of the 1996 population
CHARACTERISTICS OF THE VNDHS-II SAMPLE
The sample for the VNDHS-II was stratified and selected in two stages. There were two principal sampling domains: the group of 15 project provinces and the group of other provinces. In the group of project provinces, all 15 provinces were included in the salnple. At the first stage. 70 PSUs corresponding to the EAs as defined in the VNMRS were selected from the VNMRS with equal probability, the size of the EA in the VNMRS being very uniform. and hence sampling with probability proportional to size (PPS) was not necessary. The list of households interviewed for the VNMRS (updated when necessary) were used as the frame for the second-stage sampling, in which households were selected for interview during the main survey fieldwork. Ever-married women between the ages of 15 and 49 were identified in these households and interviewed.
In the group of other provinces, an additional stage was added in order to reduce field costs although this might increase sampling errors. In the first stage, 20 provinces, serving as PSUs. were selected with PPS. the size being the population of the provinces estimated in 1997. In the second stage, 135 secondary sampling units corresponding to the EAs were selected in the same manner as for the project provinces.
The results of the housebold and individual interviews shows high response rates. From a total 7,150 households selected in the sample, 7,031 households were occupied at the time of the interview, and 7,001 were successftdly interviewed, giving a household response rate of 99.6 percent. The household response rate was high in both urban (99.2 percent) and rural (99.7 percent) areas.
A total of 5,704 eligible women were identified in the interviewed households~ and 5,664 (99.3 percent) were successfully interviewed. Non-response was mainly due to the fact that respondents were not at home at the time of interview and during callbacks. Only one woman refused to be interviewed. The individual response rate was high in both urban (99.5 percent) and rural (99.2 percent) areas. The overall response rate for the VNDHS-II was 98.9 percent.
Dates of collection
Mode of data collection
Data collection supervision
Field supervision was conducted by tbe senior staff of GSO, members of the Technical Working Group, and staff from Macro International Inc.
Three types of questionnaires were used in the VNDHS-II: the Household Questionnaire, the Individual Questionnaire, and the Community/Health Facility Questionnaire. A draft of the first two questionnaires was prepared using the DHS Model A Questionnaire. A user workshop was organized to discuss the contents of the questionnaires. Additions and modifications to the draft of the questionnaires were made after the user workshop and in consultation with staff from Macro International Inc., and with members of the Technical Working Group, who were convened for the purpose of providing technical assistance to the GSO in planning and conducting the survey. The questionnaires were developed in English and translated into and printed in Vietnamese. The draft questionnaires were pretested in two clusters in Hanoi City (one urban and one rural cluster).
a) The Household Questionnaire was used to enumerate all usual members and visitors in selected households and to collect information on age, sex, education, marital status, and relationship to the head of household. The main purpose of the Household Questionnaire was to identify women eligible for the individual interview (ever-married women age 15-49). In addition, the Household Questionnaire collected information on characteristics of the household such as the source of water, type of toilet facilities, material used for the floor and roof, and ownership of various durable goods.
b) The Individual Questionnaire was used to collect information on ever-married women age 15-49 in the surveyed households. These women were asked questions on the following topics:
- Respondent's background characteristics (age, education, residential history, etc.);
- Reproductive history;
- Contraceptive knowledge and use;
- Antenatal and delivery care: Infant feeding practices;
- Child immunization and health;
- Fertility preferences and attitudes about family planning;
- Husband's background characteristics;
- Women's work information;
- Knowledge of AIDS.
c) The Community/Health Facility Questionnaire was used to collect information on all communes in which the interviewed women lived and on services offered at the nearest health facilities. The questionnaire consisted of four sections. The first two sections collected information from community informants on characteristics such as the major economic activity of residents, distance to civic services, and the location of the nearest sources of health care. The last two sections collected information from the nearest commune health center and the nearest "other" health facility (bealth center, intercommune health center or hospital), if such facilities were located within 30 kilometers of a sample cluster. For each facility visited information was collected on factors such as the type of services offered and the number of days services were offered per week, the type and.number of staff assigned and their training, and the equipment and medicines available at the time of the facility visit.
General Statistical Office
The first stage of data editing was done by the field editors, who checked the questionnaires for completeness and consistency. Supervisors also reviewed the questionnaires in the field. The completed questionnaires were then sent to the GSO by post for data processing. The office editing staff first checked that questionnaires for all households and eligible respondents had been received from the field. The data were then entered and edited using microcomputers and a software program developed for DHS surveys, the Integrated System for Survey Analysis (ISSA). Data entry was 100 percent verified. During the first three weeks of fieldwork, office editors and data processing staff were trained and supervised by a data processing specialist from Macro International Inc. Office editing and data processing activities were initiated immediately after the beginning of the fieldwork and were completed in mid-October t997.
Sampling errors, on the other hand, can be evaluated statistically. The sample of respondents selected in the VNDHS-II is only one of many samples that could have been selected from the same population, using the same design and expected size. Each of these samples would yield results that differ somewhat from the results of the actual sample selected. Sampling errors are a measure of the variability between all possible samples. Although the degree of variability is not known exactly, it can be estimated from the survey results.
A sampling error is usually measured in terms of the standard error for a particular statistic (mean, percentage, etc.), which is the square root of the variance. The standard error can be used to calculate confidence intervals within which the true value for the population can reasonably be assumed to fall. For example, for any given statistic calculated from a sample survey, the value of that statistic will fall within a range of plus or minus two times the standard error of that statistic in 95 percent of all possible samples of identical size and design.
If the sample of respondents had been selected as a simple random sample, it would have been possible to use straightforward formulas for calculating sampling errors. However, the VNDHS-II sample is the result of a multi-stage stratified design, and, consequently, it was necessary to use more complex formulae. The computer software used to calculate sampling errors for the VNDHS-II is the ISSA Sampling Error Module. This module used the Taylor linearization method of variance estimation for survey estimates that are means or proportions. The Jackknife repeated replication method is used for variance estimation of more complex statistics such as fertility and mortality rates.
The Jackknife repeated replication method derives estimates of complex rates from each of several replications of the parent sample, and calculates standard errors for these estimates using simple formulae. Each replication considers all but one clusters in the calculation of the estimates. Pseudo-independent replications are thus created. In the VNDHS-II, there were 205 non-empty clusters. Hence, 205 replications were created.
In addition to the standard error, ISSA computes the design effect (DEFT) for each estimate, which is defined as the ratio between the standard error using the given san pie design and the standard error that would result ifa simple random sample had been used. A DEFT value of 1.0 indicates that the sample design is as efficient as a simple random sample, while a value greater than 1.0 indicates the increase in the sampling error due to the use of a more complex and less statistically efficient design. ISSA also computes the relative error and confidence limits for the estimates.
Sampling errors for the VNDHS-II are calculated for selected variables considered to be of primary interest. The results are presented in an appendix to the Final Report for the country as a whole, for urban and rural areas, for seven survey regions, and for two groups of provinces. For each variable, the type of statistic (mean, proportion, or rate) and the base population are given in Table B.1 of the Final Report. Tables B.2 to B.13 present the value of the statistic (R), its standard error (SE), the number of unweighted (N) and weighted (WN) cases, the design effect (DEFT), the relative standard error (SE/R), and the 95 percent confidence limits (R±2SE), for each variable. The DEFT is considered undefined when the standard error considering simple random sample is zero (when the estimate is close to 0 or 1). Estimates and sampling errors of childhood mortality rates are presented only for the country as a whole, the urban and rural areas, and for the groups of project and non-project provinces. In the case of the total fertility rate, the number of unweighted cases is not relevant, as there is no known unweighted value for woman-years of exposure to child-bearing.
The confidence interval (e.g., as calculated for children ever born to women age 15-49) can be interpreted as follows: the overall average from the national sample is 1.86 and its standard error is 0.08. Therefore, to obtain the 95 percent confidence limits, one adds and subtracts twice the standard error to the sample estimate, i.e., 1.86+2x0.08. There is a high probability (95 percent) that the true average number of children ever born to all women aged 15 to 49 is between 1.7 and 2.02. Sampling errors are analyzed for the national samplo and for two separate groups of estimates: (1) means and proportions, and (2) complex demographic rates. The relative standard errors (SE/R) for the means and proportions range between 0.2 percent and 13.8 percent, with an average of 4.4 percent; the highest relative standard errors are for estimates of very low values (e.g., no education). If estimates of very low values (less than 10 percent) are removed, then the average drops to 3.5 percent. So, in general, the relative standard errors for most estimates for the country as a whole are small, except for estimates of very small proportions. The relative standard error for the total fertility rate is small, 2.9 percent. However, for childhood mortality rates, the average relative standard error is much higher, 16.2 percent.
There are differentials in the relative standard e~or for the estimates of sub-populations. For example, for the variable knowing any contraceptive method, the relative standard errors as a percentage of the estimated mean for the whole country, for urban areas, and for the Central Highlands region are 0.2 percent, 0.4 percent, and 4.0 percent, respectively.
For the total sample, the value of the design effect (DEFT), averaged over all variables, is 1.5 which means that, due to multi-stage clustering of the sample, the average standard error is increased by a factor of 1.5 over that of an equivalent simple random sample.
Other forms of data appraisal
Nonsampling errors are the results of mistakes made in implementing data collection and data processing, such as failure to locate and interview the correct household, misunderstanding of the questions on the part of either the interviewer or the respondent, and data entry errors. Although numerous efforts were made during the implementation of the VNDHS-II to minimize this type of error, nonsampling errors are impossible to avoid and difficult to evaluate statistically.
Use of the dataset must be acknowledged using a citation which would include:
- the Identification of the Primary Investigator
- the title of the survey (including country, acronym and year of implementation)
- the survey reference number
- the source and date of download
Disclaimer and copyrights
The user of the data acknowledges that the original collector of the data, the authorized distributor of the data, and the relevant funding agency bear no responsibility for use of the data or for interpretations or inferences based upon such uses.