International Mission for Iraqi Elections (IMIE)

Evaluation of the Out-of-Country Voting Program for the January 30, 2005, Iraqi Elections

Executive Summary
Part 1: General Assessment
Part 2: On-site Monitoring
1. Overview
2. General observations
3. Registration process (8 indicators)
4. Display of the voters lists (4 indicators)
5. Polling process (16 indicators)
6. Counting process (13 indicators)
7. Conclusion
Appendix 1 – List of Polling Centres by Country and City
Appendix 2 – Questionnaire
Appendix 3 – Instructions for Monitors

Executive Summary

The Out-of-Country Voting (OCV) Program for the January 30, 2005, Transitional National Assembly elections was implemented and carried out by the International Organization for Migration (IOM), as mandated by the Independent Electoral Commission of Iraq (IECI). The program was established in 36 cities in 14 countries. The International Mission for Iraqi Elections (IMIE) sought and received accreditation from the IECI to observe the OCV process. To this end, the IMIE was able to recruit monitors in 13 of the 14 countries (all except the Netherlands), who completed detailed monitoring reports on the four main steps of the process: registration, display of lists, polling and counting of the votes.

The first part of the evaluation draws from a variety of information sources: reports from monitors, relevant IECI regulations and procedures, relevant IOM OCV material, copies of OCV reports prepared by the government of Japan and the Inter-Parliamentary Union, as well as three IOM reports on registration, polling and counting.

The second part of the evaluation provides a more detailed analysis, featuring qualitative and quantitative data based on IMIE monitors' on-site reports. A questionnaire was developed to standardize the assessment of 41 indicators on a five-point scale.

Overall, the IMIE finds that the OCV program was carried out effectively at all stages and can be qualified as a successful experience. The IECI and the IOM are to be commended for this favourable outcome.

The IMIE monitor reports show that nearly all aspects evaluated received a high score and were well conducted. Low scores are the exception and point to specific and isolated cases.

Additional reports indicate that the IOM was very successful in carrying out its mandate, particularly in light of the security situation in Iraq and the tight implementation time frame. These reports further show that the IOM succeeded in reaching out to its intended audience, in training poll officials, registering electors, conducting the polling process and counting ballots.

Eligibility requirements for OCV promoted inclusiveness, but in certain cases also gave rise to problems in determining what types of documentation were acceptable to prove eligibility. Increased efforts to publicize eligibility requirements, including, where possible, advertisements in minority and host-country languages, may be considered for future electoral events.

The decision to have separate registration and polling processes meant that considerable time and effort were needed to participate. While postal registration and/or polling could have been considered, the potential for fraud may have been too high given the difficulties with the Iraqi postal system.

Approximately 25 percent of those initially expected to register for OCV actually registered. While a 25 percent participation rate is higher than usual for OCV programs, a more realistic initial estimate might have freed up resources. In total, some 265,000 Iraqis voted through the OCV Program in the January 30, 2005, elections (93 percent of those registered outside of Iraq).

Of all the aspects of the OCV process, it would appear from monitors' reports that the display of voters lists was the least satisfactory, likely because the IECI procedures did not permit accredited observers to inspect the voters lists and because they were displayed for only one day.

Preliminary results were communicated in time to Baghdad and no complaints were filed.

Because out-of-country votes cannot be assigned to a particular governorate, OCV is not contemplated for the forthcoming constitutional referendum. The IECI should ensure that this limitation is well understood by Iraqis living outside Iraq.

Areas for improvement at future electoral events include:

• gaining a better sense of the number of potential OCV registrants, to maximize resources;
• clarifying and publicizing the eligibility requirements more effectively;
• considering the addition of centres, particularly in large countries;
• increasing voter education;
• clarifying the role and functions of OCV officials to ensure stricter adherence to the rules; and
• considering more restrictive rules for the presence of media, along with stronger enforcement mechanisms.

Part 1: General Assessment

The Out-of-Country Voting (OCV) Program for the January 30, 2005, Transitional National Assembly elections was implemented and carried out by the International Organization for Migration (IOM), as mandated by the Independent Electoral Commission of Iraq (IECI). The program was implemented under IECI Regulation 10/2004, which provided the legislative framework for registration, display of the voters lists and polling, and IECI Regulation 16/2005, which established procedures for vote counting outside Iraq. The IOM fulfilled its mandate in close co-operation with the IECI, which had complete authority over policy decisions.

The IECI estimates that more than 265,000 Iraqis registered and voted through the OCV Program at the January 30, 2005, elections.^[1] The OCV Program was a commendable effort on the part of both organizations. The Iraq OCV Program was established in 36 cities in 14 countries (see Appendix 1 for details) within a very short period of time – a significant achievement, particularly since OCV is not a required practice, even in long-established democracies.

The International Mission for Iraqi Elections (IMIE) sought and received accreditation from the IECI to observe the OCV process. To this end, the IMIE was able to recruit monitors in 13 of the 14 countries (all except the Netherlands), who completed detailed monitoring reports on the four main steps of the process: registration, display of lists, polling and counting. The first part of this assessment draws from a range of information sources, such as: reports from monitors; relevant IECI regulations and procedures; relevant IOM OCV material; copies of the reports on OCV prepared by the government of Japan and the Inter-Parliamentary Union; and three IOM reports on registration, polling and counting. The second part of this evaluation provides a more detailed analysis based on the IMIE monitors' on-site reports.

The IOM reports on registration, polling and counting,^[2] which informed this assessment, give a convincing impression of a strong, dedicated and competent organization, and testify to the competence of those responsible for the IOM's success in fulfilling its mandate. Most problems encountered were minor and primarily due to factors outside of IOM control. These reports leave little doubt that the OCV operation was conducted satisfactorily and that it therefore contributed substantially to the overall legitimacy of the January 30 Iraqi elections.

The general regulatory framework, including all relevant IECI regulations and preparations as implemented through IOM procedures and described in IOM manuals on registration, polling and counting, provided an adequate basis for the OCV process.

The IOM did a remarkable job of reaching out to its intended audience. It is generally acknowledged that Iraqis abroad were well informed about their rights and about how, where and when to vote.

Considering how the security situation in Iraq could have affected the entire OCV operation, it is highly commendable that security was taken very seriously, and that it was fully understood by all staff and registrants/voters. The fact that no incidents spoiled the operation should be noted as a very positive achievement. Thanks to high levels of security, fear and intimidation do not seem to have been determining factors of non-registration, and registrants and staff apparently felt very little fear of violent attacks on the process.

Generally speaking, the OCV registration process went very well. The operation was conducted smoothly by well-trained officials in a very positive, not to say festive, atmosphere. The IOM officially reported only one case of fraudulent behaviour. The IECI's decision to extend the registration period by allowing additional opening hours and two extra days demonstrated a commitment to ensuring access to those who wished to register.

The list of potential documentation required for eligibility was not overly strict, which had the advantage of promoting inclusiveness but may have led to minor problems with the documentary evidence. The possibility of using only one piece of documentary proof (the Iraqi passport, if it contained all the necessary identification items) was introduced during the process, as was the policy of applying discretion when people had travelled long distances.

The IECI's decision to conduct separate registration and polling processes can be justified on the grounds that it allowed for the democratic option of scrutinizing the voters lists and may have prevented certain logistical problems. However, it also meant that considerable numbers of Iraqis had to spend both time and money on two trips. According to an IOM survey conducted among more than 12,000 Iraqis living in the 14 countries, non-registration was explained by a number of factors, including lack of information about candidates (13 percent), perceptions that the electoral process as a whole lacked legitimacy (13 percent), lack of interest (12 percent), and inconvenience of travel distance and the necessity of two visits (10 percent).^[3]

Of all the steps involved in the OCV process, it would appear from monitor reports that the display of the voters lists was the least satisfactory, possibly because IECI procedures did not permit accredited observers to inspect the lists.^[4] In addition, the two-day extension of the registration period actually left only one day for displaying the voters lists, with the unintended consequence that very few people seem to have used the opportunity to inspect the lists.

In general, the polling process ran smoothly, long queues were avoided, and well-trained officials helped to avoid serious procedural problems. The atmosphere was, once again, very positive, even festive in some of the centres (or outside them). Some 265,000 Iraqis voted during the three days of OCV (over 93 percent of those registered).

A number of minor incidents have been reported in the IOM End of Polling Report, including the quality of the marking ink, various attempts to demonstrate political and/or ethnic identification by staff, attempts to register during polling, proxy and family voting, and loss of registration slips. In several places, the media were reported to be not only present, but involved to an extent that bordered on intimidation, and were often alleged to be disrespectful of voters' secrecy, despite the guidelines established by IECI Regulation 11/2004 and described in the Manual for Registration and Polling for OCV personnel. IOM and IMIE observer reports also agree that campaigning took place too close to polling stations. Various problems with language were mentioned, both for voters and observers. The availability of voting and information material in some minority languages was often an issue.

The counting process was even more satisfactory than polling. Preliminary results were communicated in time to Baghdad, and no complaints were filed. The counting process seems to have been carried out according to the prescribed procedures, albeit in some places rather slowly and inefficiently, possibly due to overstaffing, lack of experience and insufficient guidance from the responsible counting officer. Observers were most critical about the failure, in some places, to display a copy of the results forms for the observers and party agents.

When planning for future events, a number of aspects should be given further attention. Only about 25 percent of the expected registrants turned out to register. This can be explained by a number of factors, including the expectation of many Iraqi expatriates not to return to Iraq and the resulting impression that there was no need to participate in Iraqi elections. While 25-percent participation is much more than what is observed elsewhere (in general, about 10 percent), a more realistic estimate might have freed up resources and avoided duplications. While it is difficult to gauge potential participation rates in OCV programs in post-conflict environments, additional time would have allowed for a better assessment of the demographics, for example, by conducting a survey among potential out-of-country voters.

In order to address the difficulties of having to travel twice – to register and to vote – postal registration and/or polling could be given consideration. However, in light of difficulties with the Iraqi postal system, the potential for fraud may be too great. Simultaneous registration and polling also brings with it the threat of identity and eligibility fraud.

Voter eligibility documentation should be reconsidered in view of the January 2005 experience. There is obviously a trade-off between, on the one hand, the inclusion of all individuals with even the slightest interest in participation and, on the other, the formality of documentation one can and should require. If the inclusion is intended to be very broad, the formal documentation could be less restrictive. In addition, increased efforts to publicize eligibility requirements and ensure their availability in multiple languages would be beneficial. This should include a careful review of all translated material to ensure quality control.

Communication systems between the Joint Operations Centre and the individual OCV centres and offices should be strengthened to avoid discrepancies and ensure consistency in the process across the countries and centres.

Time and financial resources permitting, the translation of voting materials into relevant minority and host country languages would benefit many voters, as well as election officials, party agents and observers.

Despite the overall success in recruiting, training and deploying some 7,000 election workers in only 67 days, some of the workers misunderstood their role and the importance of their functions. These could be further emphasized, so that stricter observation of the rules would prevail in all instances. This would help decrease the incidences of proxy and family voting, and bolster the seriousness of the electoral process. In addition, more restrictive rules for the presence of media, especially to prevent their interference with the OCV process, would be appropriate, along with stronger enforcement mechanisms.

Finally, the IMIE notes that the framework established for the forthcoming constitutional referendum is not compatible with an OCV program. According to Article 61(C) of the Transitional Administrative Law, "The general referendum will be successful and the draft constitution ratified if a majority of the voters in Iraq approve and if two-thirds of the voters in three or more governorates do not reject it."^[5] Because out-of-country votes cannot be assigned to a particular governorate, OCV is not contemplated for the referendum. The IECI should ensure that this limitation is well understood by Iraqis living outside Iraq.

Main recommendations to the IECI for the future:

• Gain a better sense of the number of potential OCV registrants to maximize resources.
• Clarify and publicize the documentary requirements more effectively.
• Consider adding centres, particularly in large countries where great distances must be travelled.
• Increase voter education.
• Clarify role and functions of OCV officials to ensure stricter adherence to the rules.
• Consider more restrictive rules for the presence of media at OCV centres, along with stronger enforcement mechanisms.

Part 2: On-site Monitoring

1. Overview

The IMIE was accredited by the IECI to monitor the implementation and proceedings of the OCV Program, including registration, display of the voters lists, polling and counting of votes.

Part 2 of this evaluation discusses the OCV Program as witnessed and assessed by IMIE on-site monitors. It features both quantitative and qualitative data. A questionnaire was developed (see Appendix 2 – Questionnaire) to standardize the evaluation of 41 indicators on a five-point scale. Monitors were also invited to provide additional comments supporting their observations, in the questionnaire or a separate document. In this part, the quantitative data are presented and documented in light of additional comments provided where relevant. This part of the report focuses on the items assessed by the questionnaire; it does not take into consideration contextual material and documentation other than that reported by the monitors. The first part of this report (General Assessment) provided an evaluation based on a broader range of information, including the three post-mortem reports published by the IOM (available at www.iraqocv.org).

1.1 Recruitment of monitors

The recruitment process began on December 28, 2004, when independent electoral management bodies (EMBs) in countries where OCV would take place were contacted to provide monitors. In countries where no EMB existed, foreign affairs ministries were contacted. Given the difficulty of recruiting monitors in sufficient numbers, regional or local election officials and citizen advocacy groups were contacted in some instances. The Inter-Parliamentary Union was solicited when other sources were not able to provide monitors.

Despite extensive efforts to recruit monitors, not all OCV centres were monitored during the process, and a number of centres had only one monitor to provide an assessment. The IMIE was able to recruit monitors for all of the countries where OCV was taking place, except the Netherlands, where no one responded to the recruitment initiatives.

While most monitors were recruited within the first two weeks, late recruitments were made up to and including January 29, 2005. As soon as the monitors were recruited, the IMIE provided them with a standardized monitoring form to fill out, along with a set of simple instructions (see Appendix 2 – Questionnaire, and Appendix 3 – Instructions for Monitors).

1.2 Response rates

Based on the number of monitors recruited, 120 reports were expected. Of these, 101 were submitted, for a response rate of 84 percent. Out of 74 polling centres, 53 were monitored, or 72 percent. Despite these limitations in terms of coverage, the IMIE considers that the monitoring process was sufficiently complete to draw general conclusions and lessons for future implementation of similar programs.

1.3 Data manipulations and considerations

The OCV Program was a four-step process consisting of registration, display of the voters lists, polling and counting of the votes. Taking place at different moments in time, each step of the OCV Program was evaluated separately.

Because only a small number of monitors followed the entire process, many monitors provided partial assessments. As a result, those recruited late in the process, sometimes well after the registration and display periods, could provide an evaluation of the voting and counting processes only. While in some cases monitors combined their efforts to cover the entire process for a particular polling centre, in other cases several monitors were present at the same time and in the same centre, each providing his or her own report. In sum, the monitoring process involved various arrangements in the field that were out of IMIE control, but that had to be taken into account for data compilation and analysis purposes. The three most important cases of data manipulation are described below.

Partial reports provided by different monitors

In the United Kingdom, three observers returned a separate report for the same polling centre, each of which contained data for one specific part of the OCV Program. Because there were no overlaps, these questionnaires could be merged together into a single report for the polling centre.

Reports covering more than one polling centre

In Jordan and the United States, some of the monitors produced one report covering more than one polling centre. In such cases, we replicated the scores for each one of the centres mentioned in the report.

Reports aggregating the assessments of more than one city

One of the reports submitted for Iran provided an aggregated evaluation covering more than one city (Tehran, Qom and Mashhad). This report was removed from the analysis because it was impossible to identify the strengths and weaknesses of program implementation at the polling centre level.

1.4 Other considerations

Countries with single assessments

While some countries had over 20 monitors to cover as many as eight or nine polling centres, two countries, France and the United Arab Emirates (UAE), had only one monitor each. In France, the single monitor covered the one polling centre that had been set up, while in the UAE, the monitor covered one of two centres. The results for these two countries at the country level are therefore more likely to be affected by personal judgments. The fact that they both received "excellent" scores (5/5) on all items also invites caution when interpreting the results.

Potential discrepancies in monitors' assessment schemes or expectations

Despite the IMIE's efforts to recruit skilled monitors, their background and experience in monitoring electoral events varied widely. In this context, and because the circumstances prevented more extensive training for the monitors, the results presented here must be interpreted with caution. It is likely that the variance in the assessments is due in part to variations in expectations or assessment schemes from one monitor to another, both within and between countries. This problem is inherent to large-scale multinational studies even when more extensive training is possible.

Given these difficulties, the evaluations provided here should be seen as general indications of the proceedings and should not be interpreted in a statistical manner. While scores were given numerically, most of them on a five-point scale, the IMIE's approach in analyzing and interpreting the data is qualitative rather than quantitative.

2. General observations

The IMIE organized the monitoring and evaluation process around four main steps of the OCV Program, which are presented in this report in sequential order: registration, display of the voters lists, polling and counting of the votes. Each step was measured by a series of indicators evaluated on a five-point scale (where 1 means "poor," 2 means "weak," 3 means "good," 4 means "very good" and 5 means "excellent"). This report is based on the average scores at the polling centre level, which is the primary aggregative unit used.

As shown in Table 1, not all of the four steps were documented in the 13 countries where monitors were present. In fact, the polling process is the only step that was evaluated in all 13 countries. No data were received for the registration process in Jordan, Syria and Turkey. Similarly, no data on the display of the voters lists were received from Syria and Turkey. In France, Germany, Iran, Jordan, Syria, Turkey and the UAE, no information was provided about the counting of the votes.

Table 1: Overview of the monitoring data received ( = received)

All four steps of the OCV Program that were evaluated were generally found to be highly satisfactory across the board, with only a few isolated exceptions. Of the four steps, polling and counting stood out with the highest ratings. As these were the two last steps of the OCV process, it is possible that a positive learning curve may have played a role in strengthening the program's implementation from one step to the next.

The display of the voters lists appears to have been less satisfactory. Evaluations for this step were the lowest of all four steps in six countries (Denmark, Australia, Canada, Iran, Jordan and the USA). These observations lead the IMIE to conclude that the display of the voters lists raised the most important concerns across the board, most particularly in Denmark. This could be related to the fact that IECI regulations and procedural manuals did not allow monitors to inspect the lists.

The registration process falls in the middle range, with "very good" ratings in all countries. Apart from France and the UAE, where "excellent" scores were given (by a single monitor in each case, it should be noted), the registration process was rated higher generally in Australia, Canada and Iran, but was reported less satisfactory in Sweden.

Aggregating the data at the country level naturally hides problems encountered locally. For this reason, an analysis at the polling centre level is required to provide a better sense of how the various steps of the OCV Program were carried out in the field. The following sections present in greater detail how these four steps were evaluated based on a series of specific indicators. Eight indicators were used for the registration process, four to assess the display of the voters lists, sixteen for the polling process, and thirteen for the counting of the votes.

3. Registration process (8 indicators)

The assessment of the registration process was based on the following eight indicators:

• Number of election officials to manage the registration process
• Quantity of registration material
• Fairness and impartiality of the registration process
• Provision of adequate security
• Physical set-up of the registration centre
• Clarity and completeness of instructions on the registration process
• Ease of use of application forms
• Voters' awareness of the voter eligibility criteria

While most registration centres generally received nearly "excellent" scores on all eight indicators, a few centres received a greater number of slightly lower scores (still rated "very good," or between 4 and 5 on the five-point scale). For example, Randwick Town Hall in Australia received three marks below 4 (rated "good"); Bromma in Sweden had four "good" marks; and Washington and Skokie in the USA received four and five marks rated "good," respectively. Aggregated at the country level, the overall score for the eight indicators was lower in Germany and Sweden. None of the eight indicators produced unsatisfactory marks (below the "good" mark).

It should be noted that no information was received for the registration process from Jordan, Syria or Turkey. Only one monitor submitted a report for France (covering the unique centre in that country), and one for the UAE (covering only one of the two centres).

The following section features the eight indicators used to assess the registration process, from the highest to the lowest average scores given across all registration centres.

3.1 Number of election officials to manage the registration process

The number of election officials to manage the registration process, which included IOM staff, was rated the highest of all registration process indicators, with near "excellent" scores in all centres. In fact, many monitors reported that their centre was more than adequately staffed. One possible explanation is the IECI's overestimation of registration attendance, on which the IOM based the number of personnel required. In some cases, some of the staff were redirected to other centres as required. The monitors' assessments showed that the abundance of registration staff was viewed positively, as it provided the flexibility needed to deal with the extended daily hours and two additional registration days without any problems. In contrast, it was reported that centres in Iran required tighter queue control, and that registration staff could have been more numerous in some centres. Monitors gave slightly lower scores on this aspect for Germany (Butzweiler Hof), Australia (Randwick Town Hall), the USA (Skokie) and the UK (London), without further explanation.

3.2 Quantity of registration material

According to the monitors' assessments, registration material was largely provided in sufficient quantities. Slightly lower ratings are found in a few areas, including in Australia (Randwick Town Hall), Germany (Butzweiler Hof) and the UK (particularly in Manchester). The absence of very low scores suggests that there were no shortages or serious problems to report for this indicator.

3.3 Fairness and impartiality of the registration process

The third highest scores were given to the fairness and impartiality with which the registration process was carried out by IOM staff, with all scores but one above 4 (rated "very good" or above). Fairness and impartiality were perceived to be lower in most centres in Iran, with only one "excellent" rating given for Qom. In contrast, all centres in the USA received "excellent" ratings except Skokie, which was rated "good" for this indicator. No explanation was provided by the monitor for this result.

3.4 Provision of adequate security

Most monitors gave a "very good" score to the security provided in the registration centres. In their additional comments, security was reported as being very efficient, and security officials as co-operative, neutral and particularly non-intrusive. The highest ratings for security were given in Canada, France, Germany and Iran. The lowest scores, below 4 on the five-point scale (but still rated "good"), were given to Sweden (Bromma), the UK (London) and the USA (Detroit and Skokie).

3.5 Physical set-up of the registration centre

On the whole, the physical set-up of registration centres was rated "very good." While Iran and the USA received the highest number of "excellent" scores, those below 4 (but still rated "good") were given to Skokie and Washington in the USA, Bromma and Gothenburg in Sweden, Manchester in the UK, and North York and Scarborough in Canada. Explanatory comments, when present, reported insufficient size or awkward layout of the facilities. It should be noted that locations for OCV centres were identified by the IECI jointly with, and upon approval of, host governments.

3.6 Clarity and completeness of instructions on the registration process

With an average score just above 4 ("very good"), clarity and completeness of instructions on the registration process was rated very satisfactory overall, although it was among the three lowest-rated steps of the registration process. Iran, the USA and Australia generally received higher marks than other countries. Lower scores (below 4, but still rated "good") were given to Randwick Town Hall in Australia, Copenhagen in Denmark, Bromma and Gothenburg in Sweden, Glasgow in the UK, and Skokie and Washington in the USA.

3.7 Ease of use of application forms

Also evaluated at just above 4 ("very good") was the ease of use of the application forms. Difficulties were reported for Randwick Town Hall in Australia, Butzweiler Hof in Germany, Glasgow in the UK, and Skokie and Washington in the USA, but no monitor provided any explanatory comments.

3.8 Voters' awareness of the voter eligibility criteria

Lastly, also rated "very good" was the awareness of voter eligibility criteria. While very few centres obtained an "excellent" mark on this indicator, the lowest marks were given to Randwick Town Hall in Australia, Bromma and Skärholmen in Sweden, Abu Dhabi in the UAE, and Washington in the USA. Some monitors provided comments explaining that, from the second day of the registration process, as a result of a late decision by the IECI, the Iraqi passport alone was deemed sufficient proof of identity. While the decision to allow the passport to serve as a single proof of identity was a positive development, the lateness of the decision meant that a number of eligible voters had potentially been disenfranchised the day before.

4. Display of the voters lists (4 indicators)

The second step evaluated by IMIE monitors was the display of the voters lists, as prescribed by Regulation 10/2004: "The voter register for a given country will be displayed at the voter registration centre(s) of that country during a two-day period immediately after the voter registration period." According to the IECI Manual for Registration and Polling, "The Provisional Voter List for each Station will not be publicly displayed, but will be made available for inspection under staff supervision upon request. [...] The only persons who may inspect the Provisional Voter List are: Out-of-country registered voters, on presentation of an OCV Registration Receipt and a photo ID document; accredited political entity agents, on presentation of an accreditation badge and a photo ID document."^[6] It should be noted that neither Regulation 10/2004 nor the IECI Manual for Registration and Polling allowed international observers to inspect the lists of electors.

An interpretation of the results must take into account the fact that the IECI's late decision to add an extra two days for registration had the effect of compressing the time left for the display period. Also, the IMIE was not able to obtain assessments on the display of the lists from Syria or Turkey. Monitoring for Jordan is almost non-existent, and only about half of the reports covering this step in Australia were submitted to the IMIE.

The display of the voters lists was evaluated on the basis of the four following indicators:

• Voters' access to the voters lists upon request
• Readiness of the voters lists to be displayed at the prescribed time
• Privacy of the voters lists at all times
• Availability of the voters lists for the inspection period

Among the four main steps of the OCV Program, scores for the display of the list are the most dichotomous (with both very high and very low scores across the board). It should be noted that the four indicators used to assess the display of the lists tend to vary together, pointing to a general consistency at the centre level.

Most centres in a majority of countries were rated above 4 ("very good") for the display of the list. Lower scores were given in particular to Shepparton in Australia, North York in Canada, and Copenhagen in Denmark.

Again, it should be noted that France and the UAE received "excellent" scores across the board, but had only one monitor each.

4.1 Voters' access to the voters lists upon request

Of the four indicators related to the display of the voters lists, public access obtained the highest rating (nearly "excellent"). While all three centres in Sweden received "excellent" scores for public access, the lowest ratings went to Shepparton in Australia, Mississauga in Canada, Copenhagen in Denmark, and Detroit in the USA, for which explanatory notes indicate that the monitors were unable to see the lists being displayed.

4.2 Readiness of the voters lists to be displayed at the prescribed time

The second highest rating (just above 4) went to the timing of the display of the voters lists. Problems were reported in the same centres where monitors had raised concerns about public access.

4.3 Privacy of the voters lists at all times

The privacy of the voters lists was judged less satisfactory overall. However, ratings falling below 4 (but still rated "good") were observed in only four centres: Copenhagen in Denmark, Butzweiler Hof in Germany, and Detroit and Washington in the USA.

4.4 Availability of the voters lists for the inspection period

The availability of the voters lists for the inspection period was ranked the least satisfactory indicator on average. While most registration centres were rated at least 4 (or "very good"), a few "excellent" marks are noteworthy: Noble Park and Preston in Australia, Mississauga, Ottawa and Scarborough in Canada, Paris in France, Kermanshah and Mashhad in Iran, Bromma and Gothenburg in Sweden, Glasgow in the UK, and Abu Dhabi in the UAE.

In contrast, many centres did not reach 3 (rated "weak"): Shepparton in Australia, North York in Canada, Copenhagen in Denmark, Skärholmen in Sweden, and Washington in the USA, where definite concerns with the amount of time provided to inspect the voters lists were noted. The lowest scores in this respect were given to three centres (North York in Canada, Copenhagen in Denmark, and Skärholmen in Sweden). Although it is not possible to conclude from the data that the lists may never have been displayed or available for inspection in certain locations, these "weak" evaluations certainly should raise concerns.

5. Polling process (16 indicators)

In the 13 countries where IMIE monitors were present, all 53 polling centres for which the IMIE received assessment reports covered the polling process. The assessment of this stage was based on the following 16 indicators:

• Correct use of seals on the ballot boxes prior to opening of polling
• Quantity of polling material
• Correct use of indelible ink once each elector had voted
• Verification that voters did not already have their fingers inked before voting
• All voters signing the voters lists
• Provision of adequate security
• Number of election officials to manage the polling process
• Assistance provided to all voters requesting it
• Political entity agents' permission to observe the proceedings
• Security of location for marked ballots
• Security of location for unmarked ballots
• Verification of registration receipt and a photo ID before handing ballot to voters
• Accounting for all paper ballots
• Clarity and completeness of instructions on the polling process
• Physical set-up of the polling station
• Privacy of the vote

Overall, the various indicators evaluating the polling process received very high marks, second to the indicators related to the counting of the votes. At the country level, France and the UAE received "excellent" scores for all 16 items (caution is necessary, as only one report from a single monitor was submitted for each country). Australia, Canada and Jordan received nearly "excellent" scores, with only a few indicators that may have presented room for improvement. In contrast, the UK and the USA both received lower marks overall. The following section presents the results for the 16 indicators, listed from highest to lowest average score.

5.1 Correct use of seals on the ballot boxes prior to opening of polling

The correct use of seals on the ballot boxes received the best evaluation of all 16 indicators, with a large number of "excellent" marks found in Canada, Denmark, France, Jordan, Sweden, Syria, Turkey, the UAE, and in all polling centres but one in Australia (Randwick Town Hall) and one in Iran (Tehran). Lower – but still "very good" – marks went to Tehran (Iran) and Skokie (USA), where a monitor suspected that one seal had been removed between the second and third days of the polls. In London (UK), it was reported that the staff had voted before the opening of the polling station at 7:00 a.m., outside of the monitoring window.

5.2 Quantity of polling material

The quantity of polling material was reported to be largely sufficient (rated "excellent") in the vast majority of polling centres. Similar to the abundance of staff reported at the registration stage, this may also have resulted from the IECI's overestimation of attendance at the OCV polls. No specific information explains the slightly lower scores (still rated "very good") given to Randwick Town Hall (Australia), Ahvaz, Kermanshah and Orumiyeh (Iran), Jaramana (Syria), Manchester (UK), and Rosemont (USA).

5.3 Correct use of indelible ink once each elector had voted

The use of indelible ink to mark each voter's finger to ensure that voters could not vote twice was rated very successful in most centres. Initially, the procedure required that voters have their fingers inked upon reception of their paper ballots. This resulted in stained ballots, and the procedure was changed to have the voter's finger marked after casting his or her vote. Many monitors reported this problem in side notes but still rated the procedure highly satisfactory. There was only one score below 4 (rated "good"), for Washington (USA), with no explanation provided.

5.4 Verification that voters did not already have their fingers inked before voting

Also among the top-ranked indicators, and strongly correlated with the previous one, is the verification by polling officials that voters did not already have their fingers marked with indelible ink before being handed a paper ballot. Only one lower evaluation (rated "good") was given in Jaramana (Syria), without further explanation.

5.5 All voters signing the voters lists

The regulations required that all voters sign the voter register in order to be handed a paper ballot. This procedure was reported to be followed correctly in the vast majority of polling centres, with only one lower score in Bromma (Sweden). It should be noted that no data for this particular aspect were obtained from France, Germany, Jordan, Syria, Turkey, the UK or all centres but one in Iran.

5.6 Provision of adequate security

Security was one of the aspects which received very explicit and positive appraisal and which was rated very efficient by most monitors. The entire OCV process was carried out without major incident or threat. However, some minor incidents were reported, with monitors indicating that security staff was highly responsive, co-operative and non-intrusive. Nowhere in the side notes did a monitor record a complaint regarding security; however, a number of positive comments were reported. The fact that election staff and voters felt secure was also reported to have contributed to the success of the OCV Program. Only one lower score was given (rated "good"), in Manchester (UK), but without further explanation.

5.7 Number of election officials to manage the polling process

As was the case for the assessment of the registration process, the number of officials to manage the polling process was considered more than sufficient. In fact, the number of personnel may have been found to be greater than what was required in most polling centres. Some monitors reported that the abundance of staff resulted in an overenthusiastic atmosphere that may have detracted from the professionalism of the event. Despite this, most centres received a nearly perfect mark on this aspect, except Butzweiler Hof (Germany) and Manchester (UK). No specific comments explained these two lower marks.

5.8 Assistance provided to all voters requesting it

The very high scores for this indicator demonstrate that assistance was adequately provided to all voters requesting it. Moreover, some monitors added that these services were generally available in Kurdish and Arabic to accommodate most voters. Despite the IOM's implementing the IECI's policy to provide services in Arabic and Kurdish, as was the case for in-country voting, a few monitors indicated that the IECI could have provided for services in more languages. That said, only one centre (Los Angeles, USA) received a score below 4 (but still rated "good").

5.9 Political entity agents' permission to observe the proceedings

The access granted to political entity agents to monitor the process received an "excellent" mark in most centres in which it was evaluated – a rather limited number of centres overall. No data were received from France, Germany, Iran, Jordan, Syria, Turkey, the UK, or Calgary in Canada. Monitors in Skokie and Washington in the USA rated this aspect below 4 ("good"), without explaining their assessment.

5.10 Security of location for marked ballots

The security of the location where marked ballots were placed was rated "good" in most polling centres. Noteworthy are the excellent scores given to centres in Australia (all except Randwick Town Hall), Canada (all except Mississauga), Denmark, France, Jordan, and the UAE. "Good" marks were given to Manchester (UK), Chicago and Skokie (USA). No data could be obtained for Munich (Germany).

5.11 Security of location for unmarked ballots

The security of the location where the unmarked ballots were stored was rated above "very good" on average. In a few polling centres, namely Scarborough (Canada), Copenhagen (Denmark), Bromma and Skärholmen (Sweden), London and Manchester (UK), and Detroit and Skokie (USA), the location for unmarked ballots was reported to be less secure than for the marked ballots. Lower scores given to Manchester (UK), and Detroit and Skokie (USA) constituted the most serious concerns in this area. In contrast, the location for unmarked ballots appeared more secure than that for marked ballots in Kermanshah, Mashhad, Qom and Tehran (Iran), Gothenburg (Sweden), Glasgow (UK), and Washington (USA). No monitor provided comments on the security of the location of the paper ballots, whether marked or unmarked.

5.12 Verification of registration receipt and a photo ID before handing ballot to voters

Verification of registration receipts and photo IDs before voters were handed a ballot was also rated in the upper mid-range of the evaluation spectrum ("very good"). The verification process received an "excellent" score in Australia (except Auburn and Randwick Town Hall), Canada (except North York), France, the UAE, Turkey (Ankara) and all centres in Jordan. Of greater concern were Detroit (USA), rated "good" and, more particularly, Glasgow (UK), rated "weak."

5.13 Accounting for all paper ballots

Most monitors indicated that all paper ballots were accounted for in their polling centres, particularly in Australia (except Shepparton), Canada (except Mississauga), Iran, Jordan and Sweden. Scores closer to 3 (above "good") were given to Shepparton (Australia), Glasgow and Manchester (UK). In Detroit (USA), the lowest score of all (rated "weak") could indicate that the monitor in that centre observed some irregularities, but no additional information was provided to document the case.

5.14 Clarity and completeness of instructions on the polling process

The clarity and completeness of instructions provided to voters on the polling process were rated "very good" in most centres, but slightly less satisfactory in Bromma and Skärholmen (Sweden), Skokie and Washington (USA), and lowest in Manchester (UK).

5.15 Physical set-up of the polling station

The physical set-up of polling centres was among the lower-rated indicators, with 9 of the 53 locations rated under 4 ("good"). According to the monitors' reports, polling locations were rated "good" in Fairfield (Australia), North York and Scarborough (Canada), Gothenburg and Skärholmen (Sweden), Detroit and Washington (USA), and even "weak" in Skokie (USA) and Manchester (UK). Some monitors reported that the location set-up in their centre did not facilitate efficient queue control, but this was not a widespread situation, since many others reported a very straightforward and expeditious process. The only conclusion to draw from these ratings is that probably not all buildings were equally suitable for operating a polling centre, and that efficiency of physical organization and queue management was uneven across polling centres. It should be noted that the IOM did not have a choice about polling locations, which were identified by the IECI in close collaboration with, and with the approval of, the host governments.

5.16 Privacy of the vote

Lastly, the privacy of the vote received the lowest rating of all indicators, with an average score of 4 (still rated "very good"). Twenty-three centres were rated "good" or less (two were rated "weak," Skärholmen in Sweden and Washington in the USA). Monitors made various comments respecting what they considered to be violations of the privacy/secrecy of the vote, including attempts to vote by proxy, directives to illiterate voters on how to cast their ballots (particularly women), cast ballots being shown publicly, and the presence of media largely described as invading voters' privacy at the time of casting their ballots. According to the monitors' side notes, this last point seems to have been widespread and particularly problematic.

6. Counting process (13 indicators)

Of the four steps monitored by the IMIE under the OCV Program, the counting process was rated highest. Unfortunately, this stage was also one for which the IMIE received the smallest number of reports, with no monitoring in France, Germany, Jordan, Syria, Turkey or the UAE, and only partial monitoring in Australia (Auburn, Fairfield and Preston) and Canada (North York, Ottawa and Scarborough). Nevertheless, where data could be obtained, it can be claimed that the counting process generally went very well and in accordance with the prescribed regulation.

The assessment of the counting process involved the 13 following indicators (listed from the highest to the lowest score):

• Ballot boxes correctly resealed after the count
• C04 Results Forms correctly signed by the table supervisor(s)
• Sufficient count officials to manage the counting process
• Intactness of the ballot box seals before the beginning of the count
• Good order during the counting process
• Match between the number of ballots counted and the number of votes cast
• Political entity agents' presence and ability to witness the counting process
• Preparation of the count location officials
• Provision of adequate security
• Compliance with the Count Location Procedures during the process
• Physical set-up of the counting location
• Understanding of the Rules for Determining Validity of Ballots
• Display of the C04 Results Forms for observers and political party agents

It should be noted that all centres in Australia received an "excellent" mark for all 13 indicators, as did Ottawa (Canada) and Rosemont (USA). In contrast, the lowest scores were given to all centres in the UK, Skokie and Washington in the USA, Skärholmen and Bromma in Sweden, Copenhagen in Denmark and North York in Canada.

6.1 Ballot boxes correctly resealed after the count

All centres that monitored this aspect gave it an "excellent" rating. We can therefore conclude that this instruction was carefully followed in all centres.

6.2 C04 Results Forms correctly signed by the table supervisor(s)

As indicated by the "excellent" scores given in all centres, the table supervisor(s) signed the C04 Results Forms as required. This instruction was carefully followed in all centres, as no marks below "excellent" were given.

6.3 Sufficient count officials to manage the counting process

Slightly below the "excellent" mark of 5 points, the number of count officials seems to have been sufficient in virtually all centres monitored. Only in Copenhagen (Denmark), London and Manchester (UK), and Washington (USA) did monitors provide slightly lower scores, the lowest mark being 4 points (still rated "very good") in these three centres.

6.4 Intactness of the ballot box seals before the beginning of the count

All centres except two (Washington and Skokie in the USA) received an "excellent" score, indicating that the seals were intact before the beginning of the counting process. No explanatory data were provided to explain the slightly lower mark given to Washington. Skokie obtained the lowest evaluation for this aspect ("poor"), the monitor for that centre reporting that the number printed on one of the seals was out of sequence with the others and mismatched the last information he had written down the day before.

6.5 Good order during the counting process

By and large, the count proceeded in an orderly fashion. All centres in Australia and Sweden received "excellent" scores for the order during the counting process, in addition to all but one in Canada (Scarborough) and one in the USA (Washington). Copenhagen (Denmark) and all centres in the UK (Glasgow, London and Manchester) received 4 points ("very good").

6.6 Match between the number of ballots counted and the number of votes cast

All centres in Australia, Canada and Sweden received an "excellent" score for this indicator. The other centres scored above 4 ("very good"), except London in the UK, which was rated "good." No comments were provided to explain this lower evaluation.

6.7 Political entity agents' presence and ability to witness the counting process

Likewise, all centres in Australia, Canada and Sweden received "excellent" scores with respect to the ability of political entity agents to be present and witness the counting process. Slightly weaker ratings were given to Skärholmen (Sweden) and Manchester (UK), and particularly lower scores to Washington and Skokie (USA). In that last centre, officials were reported to be less inclined to co-operate with observers, and it was difficult to obtain information and access to the polling material.

6.8 Preparation of the count location officials

Overall, the count location officials were reportedly well prepared, particularly in Australia and the USA, as witnessed by the "excellent" scores for most centres in these two countries. Only in Skärholmen (Sweden), Manchester (UK), and Detroit and Washington (USA) did the count location officials receive slightly lower marks with respect to preparation, and even in those instances it was "very good."

6.9 Provision of adequate security

Security at the counting stage received very satisfactory scores, as it did at other stages of the OCV Program. While security was given a nearly "excellent" rating in Australia, Canada and Sweden, two centres, namely in Manchester (UK) and Washington (USA), received a slightly lower mark. No comments were provided to explain these two cases.

6.10 Compliance with the Count Location Procedures during the process

The Count Location Procedures were generally properly followed during the counting process, particularly in Australia and all centres in the USA, except Skokie and Washington. London and Manchester in the UK also received lower marks for this aspect, but no further details were provided.

6.11 Physical set-up of the counting location

In general, monitors positively appraised the physical set-up of the counting locations. While all centres in Australia and the UK received "excellent" scores on this aspect, a few centres received slightly weaker marks: North York and Scarborough (Canada), Gothenburg (Sweden), and Detroit and Washington (USA). The lowest score – "good" – was given to Copenhagen (Denmark).

6.12 Understanding of the Rules for Determining Validity of Ballots

The understanding of the Rules for Determining Validity of Ballots received an "excellent" score in all Australian centres, in Ottawa and Scarborough (Canada), Copenhagen (Denmark), Gothenburg (Sweden), Glasgow and London (UK), and Los Angeles and Rosemont (USA). The only significantly weaker rating was given to Skokie (USA). None of the lower scores were accompanied by explanations.

6.13 Display of the C04 Results Forms for observers and political party agents

Lastly, among all indicators used to assess the counting process, the lowest average rating was given to the display of the C04 Result Forms for observers and political party agents. Particular problems were noted in the centres of North York (Canada), Skärholmen (Sweden), Detroit and Skokie (USA), with the latter two receiving a "weak" score. The only centre for which additional comments were provided is Skokie (USA), where monitors' access to polling material in general was reported to be particularly difficult.

7. Conclusion

Notwithstanding the methodological challenges outlined in the first part of this report, the IMIE received evaluations for 72 percent of the centres involved in the OCV Program, providing a relatively complete picture of most aspects of the program and its implementation. The overall conclusion stemming from this assessment is that virtually all aspects evaluated received very high scores and can be presumed to have functioned particularly well. Lower evaluations are the exception and generally point to very specific and isolated incidents.

The registration process was reportedly very well organized and conducted in a secure environment with sufficient staff who were well trained and who acted in a fair and impartial manner. There was sufficient registration material, the registration forms were generally found to be easy to use, and the instructions provided to registrants were clear and complete. Some monitors found that electors could have been better informed of the registration eligibility criteria. It should also be noted that many monitors indicated that having separate registration and voting periods, combined with the long distances to travel and the travelling costs involved, may have deterred many eligible Iraqi voters from exercising their right to vote. Likewise, the IECI's late decision to allow the Iraqi passport as a single piece of eligible ID may have resulted in the disenfranchisement of those Iraqis who attempted to register earlier.

Following the registration process, IECI Regulation 10/2004 provided a period for the display of the voters lists to allow for verifications and challenges to voter eligibility. The extension of the registration period over two additional days left only one day for exhibiting the lists. Most centres obtained very satisfactory scores for the four indicators used to assess this stage. However, some centres returned low scores on all four indicators, possibly highlighting problems related to the display of the voters lists or difficulties monitors encountered in accessing and scrutinizing them freely.

The polling process was assessed on the basis of 16 indicators, all of which were rated very satisfactory in all 53 polling centres for which the IMIE received assessment reports. According to these reports, the voting process took place in well-organized environments with adequate security and sufficient numbers of well-trained staff. There was enough polling material to accommodate all voters. Voting instructions were clear, and all voters requesting additional assistance received it.

Voter identification, requiring voters to show their registration receipt and photo ID, also went smoothly, and no problems were reported with respect to voters signing the voters lists before being handed a ballot. Polling centre workers correctly verified that electors did not already have their finger inked (indicating that the voter would have voted already). It was reported that, to avoid stained ballots, inking of the voters' fingers was best done after the ballot had been cast. In the vast majority of centres, the seals provided to secure the ballot boxes were used correctly (there was only one exception reported). Both marked and unmarked ballots were held in secure places, and all ballots were accounted for. Political entity agents were able to observe the process freely. As was the case for the registration stage, the festive ambiance surrounding the election was reported to be problematic at times.

The counting process, which was evaluated on the basis of 12 indicators, is the stage of the OCV Program that received the highest appraisal from monitors – but also had the smallest coverage. Where data could be obtained, however, most centres received nearly "excellent" scores on all aspects, indicating that the counting procedures were closely followed and that the entire process was highly satisfactory.

In sum, the IMIE concludes from this assessment that the OCV Program was carried out effectively at all stages and can be qualified as a successful experience overall. While the occasional lower marks provided by monitors indicate where potential improvements could be made, they by no means call into question the general success of the program. The Independent Electoral Commission of Iraq and the International Organization for Migration, which organized and implemented this program, should definitely be commended for this favourable outcome.

Appendix 1 – List of Polling Centres by Country and City

Appendix 2 – Questionnaire

Appendix 3 – Instructions for Monitors

The out-of-country registration and voting assessment is comprised of two parts. Part 1 consists of completing the Out-of-Country Assessment of Registration and Voting Form. Part 2 consists of completing a detailed assessment report summarizing the observations of the assessor.

Part 1

OUT-OF-COUNTRY ASSESSMENT OF REGISTRATION AND VOTING FORM

Step 1

Complete the information in the area at the top of page 1 including your country, the city where the registration/voting centre is located, your name and the address of the registration/voting site location. If you are not monitoring the activities of all of the polling stations set up in that location, please indicate the polling station numbers as well.

Use the following guidelines when evaluating each statement found in sections 1, 3 and 4.

Rating Scale

Excellent

Confirms that standards, policies and guidelines were followed thoroughly thereby ensuring an extremely effective performance of the function.

Very good

Confirms that standards, policies and guidelines were followed in a more than adequate fashion ensuring an effective performance of the function.

Good

Confirms that standards, policies and guidelines were followed in a satisfactory fashion thereby ensuring adequate performance of the function.

Weak

Confirms that standards, policies and guidelines were followed in an inadequate fashion resulting in an ineffective performance of the function.

Poor

Confirms that standards, policies and guidelines were followed in a disorderly fashion resulting in a very ineffective performance of the function.

Step 2

Complete the information found in sections 2 and 5.

Step 3

Tally the subtotals from sections 1, 3 and 4 and enter the total in the box found in Section 7.

Part 2

DETAILED ASSESSMENT REPORT

Step 1

Keep a detailed daily account of your observations with respect to the conduct of the registration and voting process.

Step 2

Produce a summary report detailing your observations of this initiative. Include in your report processes that were successful, areas for improvement and any suggestions for future electoral events.

[1] IOM Iraq Out-of-Country Voting Program – End of Polling Report (February 2005).

[2] Available at www.iraqocv.org.

[3] Information provided to the IMIE by the IOM, April 9, 2005.

[4] The IECI Manual for Registration and Polling states: "The only persons who may inspect the Provisional Voter List are: Out-of-country registered voters, on presentation of an OCV Registration Receipt and a photo ID document; accredited political entity agents, on presentation of an accreditation badge and a photo ID document." (p. 31)

[5] Law of Administration for the State of Iraq for the Transitional Period, Article 61(C).

[6] Manual for Registration and Polling (p. 31), provided by the IOM to the IMIE by e-mail on June 29, 2005.