At times, the electoral organisations arrange a more automated count of the
results as a whole using the computer, fax, telephone, etc., which permits them to have an official, albeit provisional scrutiny of all voting stations within a few hours after the close of voting.
It is, however, a very costly procedure, and its actual implementation depends on the communications capabilities of the country. It is an operation that can only be undertaken with certain possibilities of success in countries above a specific level of development. In other countries lacking the technological infrastructure, it can lead to spectacular failures, to the detriment of the credibility of the electoral organisations. A recent case of improper use of technology for the country was the 1995 election in Equatorial Guinea, in which they tried to use a sophisticated system of sending results by fax via satellite through Canada, at an exorbitant cost and with disastrous results.
On other occasions, the option to do 'quick counts' is chosen, that is to say, counts of a meaningful sample of the total of the voting stations and then extrapolate those results. This is not an exit poll (see Public Opinion Surveys and Exit Polls), because real votes are taken into account. But neither is it a scrutiny, because the count affects a relatively small number of voting stations or ballots from each voting station. Quick counts can be organised by the electoral organisations themselves (though it is uncommon for them to do so because of the risk of announcing erroneous or misleading forecasts), by the political parties, by the media or by national or international observation organisations.
Quick counts often serve a different purpose from obtaining reliable results fast, at times far ahead of the results being announced officially (Haiti in 1990, Bulgaria in 1991 and Albania in 1992). Quick counts, then, are more often used to verify the information offered by the electoral organisations independently. When its essential purpose is to act as a mechanism of control, above all in elections of political transition, they are referred to as parallel counts or parallel vote tabulation (PVT).
Regarding their intents and purposes, it has been indicated that quick counts have:
- prevented manipulation of the results (Philippines in 1989 and Panama in 1990)
- convinced the party in power of the necessity to accept unexpected results (Chile in 1988 and Nicaragua in 1990)
- convinced the opposition that the victory of the governing party was above board, in spite of the fact that there were suspicions of fraud (Paraguay in 1989 and Bulgaria in 1990)
- even substituted the count itself, since there was a long delay in the official count and around 25% of the votes were never counted, as occurred in 1990 in Haiti with the count carried out jointly by the United Nations and the OAS, in spite of the fact that it was a very limited sampling of 1% of the votes and had been designed only to find out whether or not a second round was necessary.
One of the primary decisions to be made when doing ordinary quick counts or counts aimed at being instruments of control, such as parallel vote tabulation (PVT), is what size the sample should be (i.e., what percentage of the total votes they should comprise), to avoid significant mistakes. It is very difficult to establish a universal model for a quick count sample size. It is, however, possible to pinpoint which factors will contribute to whether it has to be greater or smaller.
The more consolidated democracy is in a country and the longer the essential elements of the electoral system have been in place, the quick count sample size can be relatively smaller. The historical introduction of data processing techniques to tabulate electoral results permits the establishment of polling station types that, one by one or jointly, systematically obtain similar results as those throughout the national territory, so a quick count can be done exclusively on them. On the other hand, the less these conditions exist, the bigger the sample will have to be, which is of particular significance in transition elections, most in need of PVT mechanisms.
The sample size will also depend on the degree of political, social, cultural, ethnic and even religious homogeneity of the country and on the existence of politically decentralised units with their own subsystems of parties. Obviously, the more politically significant factors of this type there are that have to be taken into consideration, the larger and wider the sample will have to be.
The type of popular consensus on which the count has to be made also plays a part. No doubt the simplest count is when there are only two possible
choices (e.g., in a referendum or in the second round of a presidential
election), and it becomes progressively more complicated as the options, or voting choices, and the possibilities of personalisation of the vote increase.
Lastly, there is the degree of uncertainty about the subjects making up the quick count sample. When there is no stratification model of the population according to particular pre-set features (socio-economic, cultural, sexual, racial, identification with political models, etc.), one has to fall back on samples that do not contain any differentiation, which means that their size has to be increased considerably. This is what normally occurs in transition periods, especially in developing countries.
As a consequence of all these factors, the size of a quick count, in order to be reliable, remains universally variable. At the one extreme we would have the experience in the presidential elections in the United States, where there is a suitable combination of all the factors. This means that it will be possible to carry out a quick count of about 2,000 ballots for the whole country. At the opposite extreme is the experience in the Philippines in 1986 in which MANFREL did a quick count of 70% of the voting stations.
On the whole and probably in a rather random way, one might say that there is an inclination towards doing quick counts on 10% of the population in the case of transition elections (e.g., Chile in 1988, Panama in 1989 and Bulgaria in 1990).
It has already been pointed out that quick counts can be carried out by political parties, mass media or by national or international observation
organisations. However, in the interests of preventing electoral fraud, they will have more impact and greater importance if they are done by people who are independent of the electoral contenders. In this regard, the most independent quick counts have been those done by non-partisan national or international observation organisations, especially because of the international recognition that they have gained.
Precisely because the organisations that do quick counts are independent of the contenders and see to it that the process is above board, is not uncommon for them to clash with the electoral organisations legally commissioned to perform the same tasks and that indirectly become subjects of control. Thus, in Bulgaria in 1990 the Bulgarian Association for Free Elections had to cope with great official difficulties, including a prohibition by the Central Electoral Management Body on circulating their results the day before the elections, which was overruled only at the eleventh hour. In Nicaragua again, in 1990, the United Nations and the Organisation of American States reached a kind of compromise with the Supreme Electoral Council, duly fulfilled, consisting of not disseminating any results before having a formal consultation meeting with the said electoral organisation.
In spite of the overall favourable results of the quick counts, it cannot be forgotten, however, that these operations are closely linked to electoral observation and to transitional processes. As already indicated when we referred to observation in general, therefore, their very success in contributing to the strengthening of democracy will inevitably lead to their demise.
They are, therefore, manifestations of growth that will tend to be substituted by the administration's own quick count mechanisms to offer provisional results and for the control of the operations of these organisations by the political parties. All this, naturally, occurs within the context of a free press, without which all these mechanisms would be devoid of meaning.
