Survey research – from the first question to the final data
It is not possible to imagine the production of statistical data without reporting units - individuals, households or business entities. Just as their role is important, so the result of the survey, that is, statistical data, does not fulfil its purpose if it is not available to users. It is a kind of a circle in which all elements are interdependent and the quality of each link contributes to the quality of the entire process.
Although most of our infographics and texts are focused on statistical data and putting them into context, this text has another purpose: to briefly explain the process of conducting a research by collecting data from individuals and households. We are talking about survey research, which is used to obtain various types of data, from those on how frequently, when and why we travel, what our spending habits are, what our living conditions are, to those on how active we are and what our work habits are, but also how much and why we use the internet.
But let’s start from the beginning. What precedes the process of data collection? How is the need for certain statistical indicators assessed? Based on monitoring the needs of users, but also the national and European legislation, a decision is made on which research needs to be conducted. Prior to conducting a research, statistical methodologies should be prepared and drafted. A clearly defined survey methodology is the basis of the process of collecting the necessary data, in which, in the case of a survey research, the reporting units and interviewers of the Croatian Bureau of Statistics participate.
Quality preparation results in quality data
Have you ever participated in a survey conducted by the Croatian Bureau of Statistics or maybe wondered what those surveys are for? The obtained indicators are used for the development of various policies, they are part of scientific analyses and contribute to informing the public. They can also be part of the calculation of macroeconomic indicators.
In order for that to be possible, a survey must be well prepared, which includes two important steps - preparation of the survey questionnaire and selection of the sample. Surveys conducted by the Croatian Bureau of Statistics in accordance with the legal framework of the European Union are conducted simultaneously in EU countries using the same methodology. It is thus possible to monitor and compare their results. Therefore, the questionnaires, compiled by methodologists, are harmonised with the defined survey methodology, but at the same time, each country reserves the right to adapt them to their national needs.
Surveys are conducted on a sample, which is also defined by methodologists. The framework for defining the sample depends on the survey - those in which the Croatian Bureau of Statistics collects data on households and persons in households use the data of the last population census as the framework for the sample, while, for example, for those that aim to obtain data on certain habits of individuals, the directory of phone numbers may be used. In addition to the framework, the method according to which the sample is defined is also important. Most of the surveys carried out by the Croatian Bureau of Statistics use a stratified random sample, in one or several stages, in order to increase the accuracy of assessments. Stratification is a procedure that enables units from each individual group to be adequately represented in the final sample, which makes the sample more efficient and more representative. It all contributes to the quality of the survey implementation, but most importantly, to the quality of the data obtained. Based on the data collected on the sample, conclusions will be made about the entire population, more precisely, the parameters for the entire target population will be estimated.
Each answer recorded contributes to the survey
As for the survey process, it is carried out in the Croatian Bureau of Statistics using three methods: field collection using a laptop computer (CAPI - Computer-Assisted Personal Interviewing), filling in electronic forms (CAWI - Computer-Assisted Web Interviewing) and telephone interviewing via call centre (CATI - Computer-Assisted Telephone Interviewing). Computer-assisted personal interviewing and telephone interviewing is carried out by specially trained interviewers - employees who, in addition to general training, also attend methodological training for each survey they work on. A large number of surveys are conducted using the CATI method and interviewers in the call centre, contacting respondents of all socio-economic, educational and other characteristics, make up to several hundred calls in one day, depending on the survey they are conducting. What is their experience?
"In 99% of cases, these are pleasant conversations and the communication ends very friendly. I try to have a personal approach. At the beginning of the conversation, I explain why it is important to give accurate and honest answers and what the collected data will be used for. There are questions that people are sensitive to, then I explain to them that all information is confidential. When respondents understand the need for research and that their personal data are protected, they are ready to talk and answer all the questions," says the interviewer who has been conducting the Labour Force Survey (LFS) for the past two years. "In addition to recording answers, we, the interviewers, are also here to educate respondents. I often direct them to where they can find data processed and published in aggregate form. They are happy to see that their participation contributed to the implementation of the survey," she concludes.
Individual data are protected
After thousands of collected data, their processing follows. Depending on the survey, the processing can last from several months to even a year and a half. Three important procedures are applied during processing: detailed data editing, coding and removing direct identifiers. During the processing, in the coding procedure, some features of each respondent, such as age and sex, are marked with special codes (e.g. 1 for male, 2 for female). In the further process, data that can identify the respondent, such as the name, surname and address, are removed in order to protect the identity of the reporting unit. Every statistical research, including the survey research, is conducted on the basis of the Official Statistics Act (NN, Nos 25/20 and 155/23), which guarantees the confidentiality and secrecy of the collected data. Individual data are used exclusively for statistical purposes and are expressed in a collective (aggregate) form. Using the example of the Household Budget Survey, data for 2019 show that 50.4% of households had a dishwasher, but it is not possible to identify which households are in that share. This data will serve for some future analysis or assessment, all thanks to the reporting units’ response, their time and their willingness to participate in the survey.
A long time ago, the great Spanish writer Miguel de Cervantes said that we may judge of the whole piece by a small sample. He probably did not even think that, centuries later, this very thought would be the basis of many modern surveys. At the same time, it is a reminder that every respondent is very important and that by providing accurate and complete information, they make their contribution, becoming a piece in a large statistical mosaic.
Although most of our infographics and texts are focused on statistical data and putting them into context, this text has another purpose: to briefly explain the process of conducting a research by collecting data from individuals and households. We are talking about survey research, which is used to obtain various types of data, from those on how frequently, when and why we travel, what our spending habits are, what our living conditions are, to those on how active we are and what our work habits are, but also how much and why we use the internet.
But let’s start from the beginning. What precedes the process of data collection? How is the need for certain statistical indicators assessed? Based on monitoring the needs of users, but also the national and European legislation, a decision is made on which research needs to be conducted. Prior to conducting a research, statistical methodologies should be prepared and drafted. A clearly defined survey methodology is the basis of the process of collecting the necessary data, in which, in the case of a survey research, the reporting units and interviewers of the Croatian Bureau of Statistics participate.
Quality preparation results in quality data
Have you ever participated in a survey conducted by the Croatian Bureau of Statistics or maybe wondered what those surveys are for? The obtained indicators are used for the development of various policies, they are part of scientific analyses and contribute to informing the public. They can also be part of the calculation of macroeconomic indicators.
In order for that to be possible, a survey must be well prepared, which includes two important steps - preparation of the survey questionnaire and selection of the sample. Surveys conducted by the Croatian Bureau of Statistics in accordance with the legal framework of the European Union are conducted simultaneously in EU countries using the same methodology. It is thus possible to monitor and compare their results. Therefore, the questionnaires, compiled by methodologists, are harmonised with the defined survey methodology, but at the same time, each country reserves the right to adapt them to their national needs.
Surveys are conducted on a sample, which is also defined by methodologists. The framework for defining the sample depends on the survey - those in which the Croatian Bureau of Statistics collects data on households and persons in households use the data of the last population census as the framework for the sample, while, for example, for those that aim to obtain data on certain habits of individuals, the directory of phone numbers may be used. In addition to the framework, the method according to which the sample is defined is also important. Most of the surveys carried out by the Croatian Bureau of Statistics use a stratified random sample, in one or several stages, in order to increase the accuracy of assessments. Stratification is a procedure that enables units from each individual group to be adequately represented in the final sample, which makes the sample more efficient and more representative. It all contributes to the quality of the survey implementation, but most importantly, to the quality of the data obtained. Based on the data collected on the sample, conclusions will be made about the entire population, more precisely, the parameters for the entire target population will be estimated.
Each answer recorded contributes to the survey
As for the survey process, it is carried out in the Croatian Bureau of Statistics using three methods: field collection using a laptop computer (CAPI - Computer-Assisted Personal Interviewing), filling in electronic forms (CAWI - Computer-Assisted Web Interviewing) and telephone interviewing via call centre (CATI - Computer-Assisted Telephone Interviewing). Computer-assisted personal interviewing and telephone interviewing is carried out by specially trained interviewers - employees who, in addition to general training, also attend methodological training for each survey they work on. A large number of surveys are conducted using the CATI method and interviewers in the call centre, contacting respondents of all socio-economic, educational and other characteristics, make up to several hundred calls in one day, depending on the survey they are conducting. What is their experience?
"In 99% of cases, these are pleasant conversations and the communication ends very friendly. I try to have a personal approach. At the beginning of the conversation, I explain why it is important to give accurate and honest answers and what the collected data will be used for. There are questions that people are sensitive to, then I explain to them that all information is confidential. When respondents understand the need for research and that their personal data are protected, they are ready to talk and answer all the questions," says the interviewer who has been conducting the Labour Force Survey (LFS) for the past two years. "In addition to recording answers, we, the interviewers, are also here to educate respondents. I often direct them to where they can find data processed and published in aggregate form. They are happy to see that their participation contributed to the implementation of the survey," she concludes.
Individual data are protected
After thousands of collected data, their processing follows. Depending on the survey, the processing can last from several months to even a year and a half. Three important procedures are applied during processing: detailed data editing, coding and removing direct identifiers. During the processing, in the coding procedure, some features of each respondent, such as age and sex, are marked with special codes (e.g. 1 for male, 2 for female). In the further process, data that can identify the respondent, such as the name, surname and address, are removed in order to protect the identity of the reporting unit. Every statistical research, including the survey research, is conducted on the basis of the Official Statistics Act (NN, Nos 25/20 and 155/23), which guarantees the confidentiality and secrecy of the collected data. Individual data are used exclusively for statistical purposes and are expressed in a collective (aggregate) form. Using the example of the Household Budget Survey, data for 2019 show that 50.4% of households had a dishwasher, but it is not possible to identify which households are in that share. This data will serve for some future analysis or assessment, all thanks to the reporting units’ response, their time and their willingness to participate in the survey.
A long time ago, the great Spanish writer Miguel de Cervantes said that we may judge of the whole piece by a small sample. He probably did not even think that, centuries later, this very thought would be the basis of many modern surveys. At the same time, it is a reminder that every respondent is very important and that by providing accurate and complete information, they make their contribution, becoming a piece in a large statistical mosaic.