Finances of the G1000

G1000 principles of fundraising

The G1000 was an initiative by citizens for citizens. As an independent organization, it relied on crowd funding. The entire project was funded by donations.


Moreover, for the first two phases, we did not ask for public money. This choice for limited private donations had important consequences for the project. It required a non-relenting flow of energy towards finding money, which also conditioned some methodological choices as we explain below.


The methodology | The recruitment | Random selection | Targeted recruitment for difficult-to-reach groups | Dropout


The methodology

From the beginning, we were convinced that the G1000 should respect three important principles: diversity, inclusion and independence. These three principles are reflected in each of the methodological choices we had to make notably concerning the recruitment of the participants and the design of the script (as well as in the agenda setting process of the first phase). Nonetheless, beside these three overarching principles, methodological choices are also unavoidably driven by practical considerations, which need to be stated openly.


The recruitment

The key question when gathering citizens to discuss political issues at a deliberative event concerns the selection of the participants. There are many techniques available for recruiting participants (Caluwaerts & Ugarriza, 2012; Reuchamps, 2011). Many deliberative events, especially when they are initiated by government organizations, rely on self-reporting (Ryfe, 2005). This means that there is an opt-in possibility by which volunteers can answer a broad call for participation. Often this technique also relies on snowball sampling in which organizers rely on the participants to bring along others they know. Another technique often used is drawing samples from existing panels (Caluwaerts, Forthchoming 2012; Price & Capella, 2002). Especially when the groups are small, and guaranteeing representativeness is not the main aim, members of panels are often called upon.


Random selection

However, normatively and methodologically, the most appealing technique for recruiting participants of deliberative events is random selection (Bohman, 2007, pp. 351-352; Fishkin & Farrar, 2005). The reason why randomization is so normatively appealing is because it gives every citizen an equal chance of being selected to participate. Moreover, randomization ensures that the multitude of public opinions is present in a group and it thus “produces discussion among people who think and vote differently and would not normally be exposed to one another” (Fishkin, Luskin, & Jowell, 2000, p. 660) .


This is also the reason why the G1000 opted for random selection. Besides methodological soundness, the recruitment procedure aimed at maximizing the diversity of opinions among the participants, in order to avoid “informational inbreeding among likeminded citizens” (Huckfeldt 2001, p. 426). Citizens can, after all, only find themselves in a situation of genuine deliberation when they are faced with competing claims and opinions (Caluwaerts & Ugarriza, 2012) . When everyone at the table shares the same opinion, there is very little contestation within the group, and under such circumstances, deliberation does not lead to well-considered opinions and well-argued positions.


Nonetheless, we did not seek for representativeness. In fact, we did not put forward any claims to representativeness because when 1000 people are invited, if one does not show up, there is no statistical representativeness anymore. Rather, diversity – and not representativeness – was the central principle governing the G1000, and the random selection of participants from the population was generally considered to be the most promising technique for ensuring this diversity.


Because it proved very difficult, too lengthy and much too expensive for our crowd funded budget to draw a sample from official census lists, we asked an independent recruitment agency (GFK Significant) to contact participants through Random Digit Dialing. This technique generates random phone numbers for fixed and mobile lines and in Belgium has a penetration rate of 99%. Every inhabitant – who has a fixed or a mobile telephone – thus had an equal chance of being selected for participation in the G1000. However, for such invitations, the “yes” response rate is always very low: around 1%, so 100 phone calls for 1 yes. This figure may be surprisingly very low, but it should be qualified. Indeed, response rate for a telephone political survey ranges from 10% to 50% (on the Internet, it’s a bit higher because the respondents are slightly different from the overall population) and for such survey, no commitment is asked from the respondents. In the case of an invitation to participate to a deliberative experience, the commitment is much higher: ordinary citizens are asked to spend one (or sometimes more) free day to discuss topics for which they often have no clue and possibly no interest. So for the recruitment of the G1000 participants, we expected a normal response rate of 1%. In fact, it went up to 3% because the experience was quite well known. In addition to the phone calls by the independent recruitment agency, we also organize a follow-up call or visit (it was up to the participant) by one of our many ambassadors, who were other citizens interested in the G1000 and willing to spend some of their free time in its organization. The task of the ambassadors was to answer the questions of the participants and above all to reintroduce a human face to the event.


In order to guard over the quality of the participant sample, the random selection was checked for certain predefined population quotas. More specifically, our selection guaranteed that the sample resembled the population with regard to gender, age and province. This last quorum was considered crucial in order to guarantee a proportional representation of both linguistic groups.


In the end, these quota seem to be well respected in the group of final participants. 52% of the participants was female, 48% was male, which is a perfect reflection of the gender composition of the population, and which was rather unexpected since women are found to be more likely to drop out of such deliberative events (Ryfe, 2005). Moreover, 61% of the participants were Dutch-speaking versus 39% of the French speakers, which is also an accurate reflection of the population. And there was a large diversity in age groups, with the youngest participant being 18, and the oldest one being 85.


Targeted recruitment for difficult-to-reach groups

Despite the careful process of random selection, however, we knew that there was possibly a stronger dropout among the groups who traditionally feel less at ease with politics, or who are less interested in the subject. Moreover, some people are simply much harder to reach which further contributes to self-selection effects.


This consideration urged us to slightly expand our recruitment strategy. Because we valued the diversity at the table so much, and because we wanted to optimize the possibilities for social learning and creative thinking, we reserved 10% of the places to persons who were least susceptible to answering positively to our invitation. In order to reach these groups, we contacted numerous grass-root organizations dealing with socially vulnerable people such as homeless people or people from a foreign origin. The other 90% of the participants were selected by random recruitment.


This strategy of relaying our invitation through intermediary social associations is often suggested because of the bond of trust these organizations have with the underprivileged groups (Ryfe, 2005). Moreover, our strategy of recruiting specific target groups seems to have worked since the diversity at the tables was one of the main points of praise the G1000 project received from the international observers.



Despite all recruitment efforts and despite all the hard work of the volunteers to keep the participants motivated, we knew that it was very likely the G1000 would not reach its symbolic target of 1000 participants. As is common in deliberative practice, we experienced a dropout rate of about 30% among the people who had confirmed their participation shortly before the event, with the final number of participants amounting to 704. This has to be put into perspective, however. Unlike many other events, the participants for the G1000 did not receive any financial compensation for their participation. For instance, in Deliberative Polls, which are comparable events, the participants receive a flat fee of up to 300 euro simply for attending the event. In the case of the G1000, we could only compensate the transportation of the participants by train. Moreover, the 11 th of November 2011 was a very sunny holiday and there was a train strike, which was announced to last until 10 o’clock in the morning. This puts the dropout rate of 30% into perspective.