I also wrote about last year’s population and housing census (with reference to 2021) – RPL 2021. Then I expressed my fear that the general distrust of state authorities, as evidenced by numerous sociological studies, will negatively affect the quality of the census results.

Unfortunately, the publication of the first preliminary results confirmed my fears. Between 2,484,926 and 2,656,477 people appear in the tables by ethnicity, mother tongue, and religion with the information not available. In other words, in the respective sections, nothing is known about approximately 14% of the population of Romania.

Practically, with such results, not only can we not know how many members an ethnic or religious group has, but we cannot say anything about their evolution. For example, published data tells us that in 2011 there were 1,227,623 Hungarians living in Romania, and only 1,002,151 were registered in 2021. Apparently, the number of Hungarians has decreased by 225,000. But can we know how many Hungarians are “hidden” among the 2,484,926 people with “unavailable information”? Maybe there are 300,000, but in fact the Hungarian population has increased.

Similarly, the number of Orthodox believers decreased in the ten years between censuses from 16,307,004 to 13,989,584 (as a share of the total population from 81% to 73%). But how do we know that among the 2.6 million with “unavailable information” 2.3 million Orthodox “disappeared” between the last two censuses have not been found?

Short, the huge amount of missing information (missing for one in seven people) makes data on ethnicity, mother tongue and religion from the last census unusable.

From the clarifications received from the National Institute of Statistics, we understand the following:

• 1.1 million self-check questionnaires were incomplete;
• 500,000 questionnaires completed by reviewers in person were also incomplete;
• for about 1 million individuals who were not verified by self-completion or peer reviewers, individual data completion was provided “by imputation using statistical methods, from statistical and administrative sources”; obviously, these sources lack information about ethnicity, native language, and religion.

(It would be interesting if the INS released a county-by-county breakdown of the reasons for missing this information. It would also be nice to know how much of the lost information is due to refusal to respond, and how much is due to carelessness or failures of the registration platform data).

Undoubtedly, the lack of certain information in statistics, in particular the results of the census, is inevitable, especially in the case of indicators with an important subjective component. But it is desirable that the improvement of recording devices led to a decrease in the amplitude of approximations. Unfortunately, in the case we are analyzing, the trend is the opposite: in 2001, the most missing information was registered regarding religion – 11,734, i.e. 0.05% of the population; in 2011 there were 1,259,739 (6.3%), and now there are 2,656,477 people with unavailable information, which is 13.9% of the population being tested!

What are the practical consequences of the inaccuracy of such data? For example, according to the law, in settlements where the share of a national minority is more than 20% of the population, citizens can contact local state authorities (and must receive a response) in the language of that minority. Also, the current legislation establishes the financing of cults “relative to the number of believers, according to the last population census.” The Law on Elections of the Verkhovna Rada stipulates that organizations of national minorities can submit candidacies if their number of members is at least 15% of the number of citizens who, according to the last population census, recognized themselves as belonging to this minority. (Examples continue.) Well, not all of these laws will be exactly enforceable, or their enforcement will be contested, since the census gives such approximate results.

I know that RPL 2021 was held under special circumstances. It had to be postponed due to the COVID19 pandemic; data collection was carried out at a distance of more than six months from the reference date (December 1, 2021). The direct inspection took place during the summer vacation, which made it difficult to contact the respondents.

The RPL 2021 methodology was a hybrid, innovative (for Romania) combination of online self-verification (CAWI) and personal review by reviewers. The eye part also had an element of novelty, because data registration was no longer done on old paper forms, but on an IT platform (CAPI). There are many versions that lead us to believe that the computer tools contained errors that led to the loss of information. It seems obvious to me that the “test census” did not sufficiently test the reliability of the computer hardware.

Another factor that made the census difficult was the general distrust of the population in the authorities. It was known from all public opinion polls of the last decades, as well as from the most recent reactions of Romanian society to official inquiries. We remember how some of our fellow citizens reacted during the recent pandemic when governors tried to impose masks and then recommended vaccinations. Most likely, those who wore a mask under their chin on public transport (although they could be fined) and then refused to be vaccinated were about the same number as those who did not cooperate with the reviewers.

Finally, it should be noted that the intensive migration of Romanians, especially of a circular nature, creates additional problems for the population survey. For each of the millions of Romanians who have gone to work abroad, it is difficult to determine whether they live in fashion normally there or here

I no longer saw the overt persuasive intervention of politicians, as was the case in 2011, but I also did not see them participating in explaining the need for such an approach. Local authorities were also inactive in organizing the census, the most obvious result of which was a decrease in the number of recruiters (although the remuneration was attractive).

Several commercials and posters promoting the census did not withstand the necessary professional information and persuasion campaign. Even some well-wishers ran into difficulties and probably gave wrong information by answering an online questionnaire with many terms whose meaning was not obvious. Read the whole article and comment on Contributors.ro