Twilight of Statistics: Record Number of ‘Unknowns’ in the Most Unreliable Post-War Population Census

The media has been extensively discussing demographics, as well as the national and religious structure of the population of Croatia based on the published final results of part of the population census. The State Bureau of Statistics simply published the results of the population structure by age and gender, as well as by nationality, religion, citizenship, and mother tongue.

Politicians and the media are particularly focused on the share of Serbs in Vukovar, which has fallen below 30 percent, thus eliminating the constitutional need for bilingualism, or the use of Cyrillic. This was even the main appeal on the front page of one daily newspaper, and political repercussions are still unfolding and will continue to do so.

The data on religion is contradictory, with the share of Catholics dropping by as much as seven percent (from 86 percent to 79 percent). Consequently, the share of ‘other Christians’ has dramatically increased. This time, there were no proposed categories, so respondents identified themselves as both Catholics and Christians. However, even when the categories of Catholics and ‘other Christians’ (excluding Orthodox and Protestants) are summed, their share has decreased by almost three percent.

In addition, the number of residents has decreased by 16,696 compared to the first published results in January, which means that Croatia has not lost nine percent of its population over ten years (more precisely 9.25 percent), but ten percent (9.64 percent).

Data Without Coherent Explanation

All of this has been discussed for almost a week now. Statisticians have simply released this data to the public without a press conference and without any coherent explanation, accompanied by a terse statement that is not even completely accurate. Namely, the category that has undergone the most dramatic changes in the corpus of newly published data on nationality, religion, and mother tongue has been omitted. This is the category of ‘unknown’.

The DZS was unable to determine the mother tongue for as many as 20,840 residents, which is more than double compared to the previous census. With a share of 0.54 percent, ‘unknown language’ has reached third place, behind Croatian (95.25 percent) and Serbian (1.16 percent), and ahead of Bosnian, Romani, Albanian, Italian, and other languages.

The increase in unknowns by nationality is even greater – in ten years it has more than tripled, so now Croatia has 26,862 residents whose nationality is unknown (along with 22,388 who do not declare their nationality). Thus, unknowns have become, as with mother tongue, the third category in terms of representation among nationalities (0.69 percent), immediately behind Croats (91.63 percent) and Serbs (3.3 percent). There are now more unknowns than undecided individuals, as well as more than Bosniaks, Roma, Albanians, Italians, and members of other nationalities.

Although unknowns rank fifth in religious identification, their number has increased more than six and a half times – from 12,460 in 2011 to as many as 83,045 in 2021, with a share of 2.14 percent. Ahead of them are Catholics (78.97 percent) and other Christians (4.83 percent, thanks to imprecision in that category), non-believers and atheists (4.71 percent), and Orthodox (3.32 percent). There are more unknowns than those who do not declare their religion, more than agnostics and skeptics, and more than Muslims and other religions.

The record holder for the share of unknowns is certainly the Istrian municipality of Kaštelir-Labinci, which has 20.16 percent of residents of unknown faith, 19.02 percent of unknown nationality, and 18.89 percent of unknown mother tongue. Ten years ago, out of 1,463 recorded there, only two had an unknown nationality (now 284), three had an unknown faith (now 301), and for two, the mother tongue was unknown (now 282). This shows the quality of the census in that municipality. However, unknowns are not an Istrian specialty. Three other municipalities in various parts of Croatia also have double-digit shares of unknowns in all categories – Slivno (Dubrovnik-Neretva County), Runovići (Split-Dalmatia), and Čaglin (Požega-Slavonia).

No New Data Until at Least June

There have been no more unknowns in Croatia since the chaotic (pre)war year of 1991. When asked why this is the case, statisticians should have had prepared answers at the time of the announcement of these results. If there are any coherent explanations for this data at all, it is simply this: the population census was poorly prepared and even more poorly conducted. Let us remember the failure of the computer system, which extended the deadline for self-enumeration, which was conducted for the first time; let us remember the enumerators, for whom it was only discovered at the last minute that they needed vaccination certificates, so there were not enough of them in the end, and the deadline for household enumeration was extended. All of these were indications of what we are now witnessing – unreliable results. If they were in school, state statisticians would have to take a retake. There is no retake here, and one can only hope that this unprofessional group will not organize the next census in 2031.

This has been the second consecutive census with issues. After the spring of 2011, we waited for years for the final results. For this census, it will only be seen when we will receive data on the classification of the population by marital status, educational attainment, computer literacy, household structure… These are things that interest business more than this first set that interests politicians. However, when all this will be published – is unknown. The DZS has only casually announced that other data will be published successively, and that ‘their publication will be timely announced in the Calendar of Statistical Data Publication on the DZS website.’ When asked when this will be, we did not receive an answer. We only know that we will not receive new, most likely equally questionable data before June 2. Until then, there are indeed announcements on the DZS website that do not mention new census results.