The Question
This question was posted on behalf of Alexandre Michaud, Senior Analyst, Elections Canada.

It has been reported to us that in Senegal and in the South Sudan referendum, mobile phones were used by polling station officials to report the preliminary results.  My questions are as follows:
1. Umbrella question: In terms of organization and implementing such system, what does it involve (e.g. time necessary to set up such system (planning, procurement, rules, executing,etc)
2. Anti-fraud mechanisms: how to check the authenticity of the election officials who called duplication, etc.?
3. Is it appropriate to publish preliminary results based of phone-based incoming results in a country where significant fraud may be expected (and low trust in the EMB)?  After verification of results, important changes may occur that may trigger social unrest?
4. Anything else that is noteworthy on this subject?

Summary of responses
It is pointed out that in the South Sudan referendum, the results relayed via mobile phone were not used as official preliminary results but rather as a means to regularly update the referendum commission and other stakeholders.

Organisation and implementation
In the case of South Sudan a communication plan was developed, detailing different types of communication and establishing a clear structure regarding what to report and when. This was crucial to ensure that information was not repeated or lost in the communication chain.

A detailed logistics plan will also need to be put in place, including an assessment of the mobile phone network, procurement of phones, software development, negotiations with mobile service providers to ensure that traffic will either be prioritised or handled separately, and training for officials.

It is advised to allow six months planning time and at least three months for an open tender process.

Checking authenticity
It is recommended that the central election administrator collect all the telephone numbers of polling centre chairpersons in advance so that calls can be authenticated on election day if required. A robust communications structure should prevent duplication. In the Kenyan referendum of 2010, in addition to sending results via SMS, each official was also required to submit them manually to the returning officers.

Appropriateness
Two experts do not believe that it is appropriate to use mobile phones to publish preliminary results, but that they can be a useful means to keep the election commission updated. While conceding that any change in results could trigger social unrest, another opinion is that this can be avoided with a transparent and robust system and with endorsement from political parties. 

Other issues
It should be borne in mind that mobile phone networks can be easily overstretched on election day if special arrangements have not been made with the provider. Security measures should also be taken to prevent data from being hacked in to and stolen. Furthermore, polling station officials sending SMSs should be clearly told whether they can expect to receive a reply or not. This will avoid officials re-sending results and potentially creating problems or confusion.

Examples of related ACE Articles and Resources
Encyclopaedia:
• Telecommunications Technology 
• Publishing Interim Election Results
• Transmitting Voting Station Results 

External Resources
• National Democratic Institute: SMS as a Tool in Election Observation, Ian Schuler, Innovation, issue 2, spring 2008. 

Names of contributors
1. Idriss Kamara
2. Vladimir Pran
3. Immaculate Njenge
4. Staffan Darnolf
5. Horacio Boneo
6. Arthur Minsat
7. Ingo Boltz
8. Khalid Waheed
9. Atem Oben Henry Ekpeni

It is true that during the South Sudan Referendum process mobile phones were used by Referendum Center Staff and State Referendum Committees to report on several issues, including voter turnout and preliminary results. But this was only meant to update the South Sudan Referendum Commission (SSRC) and other stakeholders on daily progress of voting process. The turnout figures reported were not considered official preliminary results or published until paper copies were submitted to the SSRC.

1. In terms of organization and implementation of such a system, a communication plan was developed which facilitates the communications required between the Referendum Centers, the Referendum sub-committees, the SSRC for the referendum pre-polling, polling and counting periods. The communication plan distinguishes between regular, scheduled communications and unscheduled communications that have to take place when an issue needs to be addressed urgently. Regular/scheduled communications between the Referendum Centers, the State Referendum Committees and the SSRC during the pre-polling, polling and counting periods followed an established structure of communications (including time of report, what to report, etc) in order to avoid repetition of communications or loss of information along the chain. E.g. during pre-polling and polling, the State Referendum Committee Chairperson or designated official will call the Chairperson of the Referendum Center according to the scheduled time and will collect the information outlined in the reporting template, consolidate for his/her state and send it to SSRC operations according to the reporting template. Some State Referendum Committees also submit completed copies of reporting template by fax, if any.

2. To check the authenticity of election official or Referendum Center staff who called, all State Referendum Committees were asked to collect mobile phone numbers of all Chairpersons of Referendum Centers, plus an alternative staff number in case the first number is unreachable. Since the State Referendum Committees recruited the polling staff and know them personally, especially the Chairpersons, it was unlikely that someone would attempt to impersonate them without the State Referendum Committees noticing. The same is true for the State Referendum Committee reporting to the SSRC.

Regarding duplication of reporting, the established structure of communications between the Referendum Center staff, the State Referendum Committees and the SSRC prevents repetition of communications or loss of information along the chain.

3. True, it is not appropriate to publish preliminary results based on mobile phone reporting. As indicated earlier, in the case of South Sudan Referendum, preliminary results were not published until paper copies of results received via mobile phones were sent and verified by the SSRC.

4. What is noteworthy on this subject, in a poor, conflict prone country like Sudan with limited communication means, mobile phone reporting can help the Election Commission or Referendum Commission to keep track of events, such as security threat, need for additional materials, or any other emergency that could jeopardize the smooth progress of polling.

This paper by my colleague describes some of the technical solutions that we've been dealing with when setting up SMS reporting systems. The article talks about our work in election monitoring (PVTs specifically), but  the principle is almost the same. Gateway services were provided by South African Clickatell  http://mobileactive.org/files/smselectionsinnovationmag.pdf

I wish to note that Kenya used mobile phones to transmit results during the constitutional Referendum held in August 2010. This was an exciting and historic moment for us at the IIEC. The system had been tested in 2 previous by-elections held in June and July. Indeed, after the referendum we can confirm that we have successfully transmitted results (100%) from all polling stations in the four by-elections held in Sep and October 2010.

Organization and implementing such system,

In terms of organization, it will depend from country to another but it is advisable to have a lead time of 6 months when you are initiating the programme. In our case it was a local solution that had been developed for the previous Elections Commission but was not used. We only needed to re-modify a few aspects. In our case in terms of open tendering it will require a minimum of 90 days to have an open tender.

You will need to work out a detailed work plan that include the, mapping of GSM network across the country to identify the areas with mobile phone network, purchase of phone for each polling station, development of the software, pretesting (which has to be field based). There are other items are critical that will need to be negotiated with the service provider in our case Safaricom, Airtel, Orange Kenya and Yu. The other aspect that takes a bit of time is the configuration of the mobile phones that will be utilized during the transmission. You will also need to develop guideline and training kit. The officials will require training on the system. We had our own closed network-not linked to the public. The network is only for data and not other usage to reduce on traffic due to several data entering.

Authenticity

In our case, we had set targets. The time the returns were expected at the head quarters was at least 36 hours after the close of poll. We had mapped out how long it would take the officials to get back to the tallying center and submit the results. Each official was also expected to physically carry the manually filled form and submit to the Returning Officers. This was one way of authenticating the results.

Appropriate to publish preliminary results based of phone-based incoming results

Indeed i want to note that any change in results could trigger social unrest. That is why the public must be informed at every point about the status of the results. The text messages in our view give an audit trail of the results. It is also important to keep track of the officials at each point and have a call center that monitors who has submitted results and what is the status of all polling stations. You must also involve political parties and encourage them to test the system.

 Others

The phones were pre-coded and any person who is not an official could not access the database.

Uganda is about to use a SMS-based solution for the Feb. 18 elections, so get in touch with EC for more details.

You can get information from Honduras through Ana Maria Tello from UNDP who was involved in the setting up of such a system (IFES was also part of the scheme).

Posted on behalf of Arthur Minsat

Hello,

The use of mobile phones to convey elections result is common practice including in crisis countries in Africa. Members of polling stations are allocated a phone number at the very last stage of the electoral process. This is underpinned by confirmations by electoral observers with their own organisations, which are normally conveyed with mobile phones. Representatives of political parties do normally transmit local elections results with mobile phones.

A security issue with mobile phone networks is that they can be easily overstreched on the elections day (as we experienced in Cote d'Ivoire). Moreover, it is rather easy to have the confidential information scamed by third parties who could use it to their own end. Therefore, transmitting elections results via mobile phone is rather delicate at the regional level after results are relatively well tallied.

Hope this helps,

Arthur Minsat.

Albania's CEC used SMS to collect participation numbers on election day. Polling station officials were instructed to send the updated number of voters in regular intervals throughout the day.

Technically, the system was basic but effective. We used a desktop GSM modem with a standard SIM card and off the shelf SMS processing software. 

The number of that SIM was included in the instruction packages for the polling station officials. Likewise, a white-list of cellphones in use by polling station officials was created before election day.

We did not build extensive security barriers (RAS servers with whitelists etc) but rather simply filtered the CSV message log we obtained from our software to exclude messages sent by un-authorized numbers. We did not take security precautions against number spoofing because the participation % numbers were not important enough to justify the expense.

Two possible caveats / lessons learned:

1. If you can, get a Priority SMS queue from the mobile phone operator. National SMS volume on election day is often high, and if your SMSs are processed as standard messages they may suffer long delays in arriving at the CEC; also, if message volume exceeds certain limits, some operators simply purge SMS queues so you may lose some messages. A Priority SMS queue solves that problem as it is reserved for the CEC, with guaranteed delivery time and no purging.

2. Either invest in SMS auto-response (sending a confirmation sms to polling station officials tha their SMS was received) or at the least make very clear that the CEC will receive their message, even if they don't get confirmation. In our case, poll workers wondering whether their participation numbers had been received repeatedly SMS-ed us with requests to confirm, re-sending their numbers, or even try to call our GSM modem to confirm.

Best,

Ingo

Election is a mega activity in which huge amount of funds and human resource is involved,besides , time. Every nation passing through this activity is  directly or indirectly busy all the times. The climax of elections is the result which speaks the fairness and credibility of elections or electoral process. If results are not credible ,whole exercise goes futile and not acceptable by local and international community. So the poling results play an important role in elections. Collection & compilation of result should always be transparent,free of doubt and open. Collection of result on mobile may create doubts and can be manipulated ,resulting to deteriorate the entire process. However, some mechanism can be evolved to eradicate or minimize the frauds which may occur during use of this technology.

It may save time but some rules have to be made to make it fool proof.

Mechanism to ascertain the authenticity of results will also be required to adopt.Some codification system has to be formed for transmitting results to the control room.

Results taken on phone are preliminary and are always subject to collection from polling station officers personally. 

Beside using mobile phones for reporting security, logistical and other statistical issues as explained above, it very dangerous to proclaim election results based on figures obtained through phone calls. But I must say there are situations that may require special considerations such as the distance and road factor, security and even natural disaster immediately after the polls preventing documents from getting to the centers on time. In such a case mobile phones can be used to give an idea of the trend pending the arrival of documents to substantiate the figures.

In cases   where distance, road and security may be an issue, the legal framework should take in to account such delays before proclamation. Using phones for reporting in elections entail excellent preparation, planning and implementation as in Kenya and Albania. 

I just want to thank everybody for their very useful contributions.  I found the information and analyses provided quite pertinent to the questions we had been reflecting upon in the context of our assistance projects.

Best regards,

Alexandre Michaud