Census: How to statistically lie to ourselves

Ilo Trajkovski

What do lying and statistics have in common? Statistically, we ate sarma for New Year’s Eve. Namely, some ate meat, and others fasting cabbage. On average, everyone ate sarma! I remembered this anecdote recently, thanks to the perseverance of the Director of the State Statistical Office (SSO). His arguments against the inclusion of questions of ethnic and religious affiliation in the next population census reminded me of the title of the book How to Lie with Statistics”.

This small book was written in 1956 by journalist Darrell Huff. During the 1960s and 1970s, it became standard literature for studying statistics in Western countries, and later a bestseller with a circulation of one million and half copies, and only English. The main point of the book is the indication that statistics are not down to bare numbers, but that different “statistics” are indicators. They show something with numbers, they show states and processes in some mass units – natural, social, humanistic. But what they show is not visible from the statistical figures themselves. What is shown should be sought in the meanings of the statistical indicators, and they are contained in the definitions of the terms represented by them.

Thus, according to the director of the State Statistical Office, this is not pointed out in person, but as an indicator or representative of the authority of statistical science and the statistical authorities in the state – ethnic and religious statistics are not simple, but the real reasons for the failure of the census from 2011. Because of this political significance of ethnic and religious statistics, the director in our country proposes, and the Ministry of Justice raises his proposal in a draft law on population census. According to information available to the public, the next census will not include questions about the ethnic and religious affiliation of the persons to be included in the census.

Such a proposal is politically and scientifically interesting and calls for serious public debate. But it still does not exist, although, according to available information, the draft-law is in government procedure. Instead of organized scientific and expert debate, the public is only familiar with the reactions of another, if not, to political parties and to two or three journalists provoked by academics. What encourages this point of mine is that all reactions to the proposal for an ethnic and religious ignorant census is equally strongly criticized by all main and important political parties. We wonder what is the reason that makes them blow into the same horn, the constantly angry opponents – SDSM and VMRO-DPMNE, DUI, BESA and AA, but also Turkish, Roma and Serbian political leaders? And if there is such agreement among all relevant political parties, from where does the courage and power of the SSO arise, presented in the image of its director for such an increasingly independent proposal and performance? Could the SSO perhaps be the first example of so desirable institutional autonomy and professional integrity? Are we witnessing an open rivalry between statistics and politics?! But the Director, of course, is appointed by the Government of the Republic of Macedonia, and the Government is a party and, as such, in the selection of staff for filling positions in public institutions, they consistently adhere to the criterion of party affiliation! Viewed from that perspective, the case is more likely to be something else. We look for the answer in the title of the book “How to Lie with Statistics”.

How else can you understand this radical ending with the tradition of running ethnic and religious statistics? That seventy-year tradition in our country started with the census of 1948 and it was repeated in all subsequent censuses. Every one of them contains questions about ethnic and religious affiliation. True, terminology is different and definitions are different. But, in all eight census records so far, there was an indicator of ethnic and religious affiliation. Ethnicity, for example, was determined by different terms.

In the censuses of 1948, 1953 and 1964, the term “nationality” was used, and the term “ethnicity”, as identical to the term “nationality”, was introduced with the census of 1971. In the following census of 1981, then in the failed census of 1991 and the 1994 census, the term “nationality”, ie “national affiliation” was introduced with the same significance. This replacement of terms without precisely defining their meanings and specifics is best seen in the provisions of the Census of Population Act 2001, which culminated in the 2002 census. In it, in Article 9, paragraph 6, the ethnic is characterized by three characteristics of the persons covered by the census: their nationality, their mother tongue and their religion. In the same way, ethnicity was also defined in the 2011 Census of Population Law.

It is precisely the failure of this 2011 census by the director of the SSO that was used as the main argument for interrupting ethnic and religious statistics. Better and much easier moving towards one ethnic and religious “blind” statistics! Such ethnically neutral statistics appear to be in function of the danger that Darrell Huff has long warned us about. In order not to have such a manipulation of statistics and not be taken over the water thirsty, it is necessary to open a public, expert debate on this issue.

A good example of this may be the discussions that, in the course of the attempts to introduce ethnic statistics, have been conducted in France in the last ten years. The experience of the National Institute for Demographic Research (which does not exist in our country) in one of its studies, and the incentive of the then government to introduce indicators of ethnic and racial affiliation, has caused great reactions. The highlight of this was a scientific conference with the participation of all the main stakeholders (political parties, state institutions, associations for the fight against racism and discrimination, print and electronic media and 300 researchers). The conference did not result in a single conclusion about the need, assumptions, risks and consequences of introducing ethnic statistics. The two opposing sides remained in their positions. But after careful examination of the counter-arguments both sides are more informed about their strengths and weaknesses. In the end, the decision is again political, but based on the acceptance of scientific arguments and counter-arguments, and on the basis of manipulation and intimidation with statistical operations.

Ethnic and religious affiliation and statistics are not conflicting at all. What’s conflicting is the process of ethnicization and religionization of statistics and other public goods. It does not necessarily take place only with ethnocentricism or ethnic overstatement. The current proposal for some administrative-technical and technocratic justification of ignoring the ethnic aspects of the social life in our country can lead to this. Instead of statistics that shy away from ethnic and religious challenges, the country needs a move towards a science-based post-ethnic and post-religious statistics. This can be achieved with honoring, not by ignoring the phenomenon of ethnicity and diversity as an inexorable social fact, whose scientific and political importance no longer depends on the numerical proportions.