The author of our most famouse love song Laza Kosti? puts love and his song in the area „Between Wakefulness and the Dream“. May the poet and those who love him forgive us for the metaphor which comes to mind to a public opinion researcher when these verses are mentioned. The wakefulness is somehow being linked to the public opinion researcher and the dreams to the research client. The research should provide certain answers about the wakefulness itself (about reality and public) and the client expects the research to help him realize dreams which are connected to the part of reality he is dealing with. That smaller or larger gap between the wakefulness and the dream marks the difference between the science that the researcher is using in perceiving the wakefulness and all those wishes, hopes, dreams of politicians, entrepreneurs, public relations officials who are ordering surveys. The role of public opinion and similar surveys are to be found on the way from the rough wakefulness to the delicate dreams about that very wakefulness.
Surveys are helping us bring the picture of reality, that we want to know more about, closer, make it clearer, better known and better understandable. Some researchers are more, others less successful in doing that. We are trying to be successful in our work. We are not claiming, of course, to be able to have every new knowledge brought by the survey make things more visible. Sometimes the survey will only complicate things and show us how the things are not as simple as we thought. But such a knowledge is of strategic importance for making important decisions. Surveys can sometimes show that we are more or less different than the picture that we have about ourselves. If we learn something from those findings, the survey was suitable and useful.
Sometimes the client expects the research to confirm his opinion and to document the notion that he is having about the research subject. That happens most often in the political sphere. I am not claiming that there are no researchers who will fall in the trap of wanting to proove the prejudicial opinion of the client, but most of the researchers are trying to present the most objective image of reality possible. The mentioned (fake) surveys are just there to test the vanity of the client, they are of no use to anybody and they are harmful for the client and for the researcher.
Surveys are not value-free, if the researchers stick to the values which are coming out of scientific principles and ethical values of the profession. The fact that surveys are not value-free, because they are following ethics and science, does not lead to the conclusion that they are necessarily „flattering“ the client.
There are of course prejudicial surveys, those with „ordered results“. The centre „Demostat“ does not conduct such surveys, „Demostat“ is conducting objective surveys, whether the client likes that or not.
The survey design is the task of the researcher, but with serious consultations with the client. Determining the objectives of the survey is especially important and that needs to be the joint task of the researcher and the client. The researcher should not impose objectives to the client, especially when the client is not really able to state precisely what he would like to accomplish with the survey, but also the client should not impose objectives to the researcher, if those objectives are unrealizable or if they are in contrast to certain standards (scientific, ethical, legal standards) which the researcher needs to follow.
The researcher has to inform the client about all aspects of the research, which can influence the credibility of the findings and the survey price. Without a doubt, a smaller sample survey is less expensive, but also less reliable; an online survey is certainly cheaper than a face to face survey, but it is also less reliable.
It is a joint task of the researcher and the client to design the survey with the most favourable relation between the validity and reliability of survey findings and the price.
Many survey classifications, stricto sensu, make no sense, starting with the most widely spread classification into in-dept research and applied research or the classification into quantitative and qualitative research. In-dept research is senseless if it has nothing to do with a certain application, just like the applied research is senseless if it is not based on an in-dept research. But in the research work this rough classification is used in order to cover up the senseless „in-dept“ research and the „applied“ research, because they are of no use. The research concept, that we in „Demostat“ are insisting on, is based on a firm connection between the in-dept and applied research. Only with such a connection it is possible to generate good public opinion research with valid and reliable results. Our standard public opinion research is basicly neither a quantitative nor a qualitative research but it is both at the same time.
Because of the communication with potential clients we are forced to use the usual research classifications and that is why we will mention some of the usual types of research, for which „Demostat“ has the appropriate capacities.
Desktop research. For us in „Demostat“ this is the pre-research, which includes work which has to be done previously to every serious research project. It includes the collection of all selected relevant data, which have been produced within other research projects, documents from the public administration or records from some other ressources and their analyses. It is the analytical use of data produced by others. Within a good desktop research theoretical analyses of other authors are collected and analysed and then they are being linked with quantitative data, which allows the creation of the most optimal platform for the analysis. „Demostat“ is using the desktop research also as an explanation for the intended research projects.
Action research. It means the ful research participation in all three essential phases of every action: designing the change/action, realization of action/change and results of the action i.e. introduced change. A starting research is foreseen before the beginning of the action planing just like also the final research after the finalized action as well as a number of partial research efforts in the meantime. Basicly it is part of the social engineering which is followed by a number of ethical questions during realization, among other things, also because it is frequently disputing the scientific status of this research type. Nevertheless, action research is unavoidable in testing some of the social policy measures and the scope of application is not so narrow as it may seem. It can even be applied in public policies and in politics itself. Anyway, the Research and Publishing Centre „Demostat“ has personnel capacities to conduct this type of research and it is inclined towards participatory action research.
Panel research. Repeating the research in certain intervals in order to follow eventual changes in the researched population usually is called panel research. This type of research should not be confused with the online panel research, which is a type of modern imitation of the panel research. It includes the selection of a group of interviewees („web panel“), which is then being used for research on different topics. Interviewees are normally receiving a certain remuneration, they are contacted by phone or internet or the questionnaire is posted on a platform and the interviewees can reach it themselves. We should say that in the search for cheap „research“, online panels are becoming trendy in the area of social research. That makes sense in countries with a higher level of internet use, but in countries like ours it is only useful for some public or market research activities. The face-to-face interview is almost disappearing in the most developed countries, but in our country it is not possible to avoid this interview type. In the USA it is possible to collect good data for the election prognosis also with an online questionnaire, which is something that was proven at the last presidential election, but that would be a total disaster in our country. (The use of the online panel research in marketing research activities is not only less expensive and less time consuming, but it is also more realistic).
The client can of course also ask for a complete interpretation of research findings and a plan of concrete activities and finalized projects. That depends on the agreement. When it comes to politics and to the area of public policies, that is above all the task of think-tanks, and the centre „Demostat“ is also fulfilling that task with the help of an interdisciplinary team. We can gather thinking people, we have credible colleagues and we can hire experts, who are able to do this type of analysis, to create recommendations and projects. when it comes to the method of work, our work relies on the creation of a synergy between researchers, clients i.e. decision makers, users of the findings and other stakeholders and the research object. All of that takes place in all phases of the research, starting with the preparation activities. In this case we are insisting also on post-research activities, above all on the discussion between the stakeholders about the research results and also about the concrete plan of proposed changes, which are normally part of the finalization of such projects.
The research team of „Demostat“ prefers to use the representative sample (random sampling) and then also quota sampling. A certain level of stratification of the basic population is foreseen in both cases, which means that the basic population is divided into specific parts (strata) based on relevant criteria and then within these parts simple random or quota sampling takes place.
The representative sample includes a random selection of interviewees. In the so-called random sample each unit of the basic population has to have equal chances to be part of the sample.
The representative sample is based on the idea about the biggest possible match between the main and most relevant characteristics of the selected part of the population and the basic population. If the differences between the sample and the basic population are statistically significant, then it is a frequent mistake in the public opinion research – the unrepresentative sample.
The standard procedure for collecting answers of interviewees that we are in „Demostat“ instisting on is the direct personal (face-to-face) interview. „Demostat“ has a network of experienced and trained interviewers which can cover the complete territory of Serbia. The optimal dynamics for the realization of a half-hour interview in 75-100 cities in Serbia at a sample of 1,500 interviewees foresees five days. Another two days are necessary for the control of the interview process at a sample of five percent.
The centre „Demostat“ offers also the phone interview (CATI - Computer Assisted Personal Interviewing) and the online questionnaire (CAWI – Computer Assisted Web Interviewing).
For some research activities the in-dept interview and the focus group discussion (a group interview of focus group members) are especially important and „Demostat“ has both the people and experience necessary for the realization of these activities. Of course, we don’t see the technique of data collection called the in-dept interview or focus group as a quality research but only as additional procedures in gathering basic information for a standard research. (The use of focus groups in marketing research activities is completely different and it makes sense to use it in that area.)
We are not even close to talking about a quality research if that research only includes counting the answers gained from interviewees. The research does not mean that you only ask questions and get answers. The researcher should not transfer the problem, that he is supposed to solve, to the interviewee and then take the answer for granted. In the strategy of data collection the centre „Demostat“ offers special attention to the critical approach to answers gained from interviewees.
There are many problems when it comes to the validation of answers. We need to ask ourselves first if we are inquiring the public opinion or the secret, private opinion. If we are promissing anonimity to the interviewee, is he then giving us his public opinion or his private and secret opinion. So the big question remains if the interviewee would express in public the position that he has told us in confidence as his opinion in anonimity.
There are also questions to which the interviewee doesn’t know how to answer and he is embarrassed to say that he doesn’t know so he randomly selects one of the offered answers. Thus the researcher has to assess the level of information of the interviewee in advance in order to avoid asking questions most of interviewees are not able to answer.
Giving socially acceptable answers happens frequently in the communication between interviewers and interviewees. It is therefore recommended to avoid asking questions which provoke socially acceptable answers.
Declining to answer specific questions happens increasingly often. For instance, there are less and less interviewees ready to answer questions about their income and the income of the household. The centre „Demostat“ is collecting indicators of the income and the financial situation of individuals indirectly, based on peripheral indicators.
It happens also a lot that interviewees refuse to answer out of fear or they give false answers out of fear. For instance, during the pre-election survey for the presidential election in 2017 a huge number of interviewees avoided answering who they would give their vote and also many of them said that they were afraid to say who they would vote for. The centre „Demostat“ minimises this by using answer gathering procedures, in which the interviewees are encouraged to answer all questions – in a face-to-face interview. All in all, we are not treating the answers of interviewees as the final research product but as only one of a few sources, although as the most essential source for the good interpretation of research findings.
We are trying to go a step further than just to use bare statistical data – numbers, tables, graphs – and to point to their meaning.
It is easy to give the client numbers, tables, graphs and let him interpret them as he likes. That is however often the wrong way to do it. It is the job of the researcher to point to the main directions of interpretation when it comes to the facts determined within the research. Only then the client can start interpreting the gained information in the context of his experiences, his rules of behaviour and business principles. Only with such an approach the numbers become explored facts. Uninterpreted numbers remain hieroglyphs. It is not the objective of the researcher to send only numbers to the client but to point to the exits out of this labyrinth of numbers. Only correctly interpreted numbers can become the instruction for action.