Surveyors can use the list of boosters and barriers to establish three different prior modes. To do this it is essential that they finalise and, if necessary adjust, the list of barriers and boosters following the completion of Stage 2 (see Stage 2: Analysis of findings).

Simple BBQ prior mode:

The simplest approach to deciding the mode of the prior from the list of boosters and barriers is to score all findings equally, add the booster scores to 0, subtract the barrier scores from 100 and take the average of the two findings. To do this assign each booster and barrier a score of five, and total each. Subtract the total of the barriers from 100 and add the total of the boosters to 0. Then take the average of the two resulting numbers:  [math] Prior mode = left (Sum of Boosters + 0 right )+left (100 - Sum of Barriers right ) div 2 [/math] For example, Table 1 shows a list of boosters and barriers collected during the implementation a BBQ tool during a SQUEAC in Afghanistan: For the Simple BBQ prior mode, each of the Boosters and Barriers were assigned a score of 5% and totalled: Boosters had a total score of 40%; and Barriers a total score of 80%. The prior mode was worked out with the following calculation: [math] left (40 + 0 right )+left (100-80 right ) div 2 = 30% [/math]

Weighted BBQ prior mode:

Another approach for estimating the prior mode is to use scores or weights that reflect the relative importance or likely effect on coverage of each finding. To weigh each booster and barrier with the survey team, the survey leader should follow the following steps:

• Ensure that all of the available information collected from Stages 1 and 2 (including the results of Stage 2) are incorporated into the BBQ tool - displayed either on the original flipcharts or typed in a table and projected (like in Table 1).
• Each Booster and Barrier should have the triangulation information displayed next to it (including the sources who mentionned each factor, the method by which the information was collected, and if relevant, the demographic information of the individuals or groups who cited the factor).
• Then, with the survey team, he or she should go through the list and ask the survey team to attribute a score or weight to each barrier and booster based on its relative importance or likely effect on coverage.
• To avoid the sum of the boosters or barriers exceeding 100, the maximum score should be scaled. To do this, the survey leader should take the count of the boosters or barriers (choosing whichever has the most factors) and divide this number by 100. For example in Table 1, there were 16 barriers identified in total. 100 divided by 16 = 6.25. Therefore the maximum score or weight attributed to each barrier and booster was 6.
• To ensure that each of the team attributes weights which are not influenced by the other team members, the team leader should ask all the survey members to close their eyes and then raise their hand when the weight which they believe is most accurate is called out. The survey leader then counts up the total number of votes for each weight and assigns the most popular weight as the final weight.

With the final weights for each booster and barrier agreed and totalled, the prior mode can worked out using the same formula as the Simple BBQ. Using the information in Table 1 as an example: [math] left (22.5 + 0 right )+left (100-42 right ) div 2 = 40.25% [/math] This exercise can either be done with the entire survey team, or the team can be broken up into groups who can do the exercise separately, averaging their prior modes once they have completed the exercises. Boosters and barriers can also be weighed with community members in a village or community within the survey area. An account of the process involved of community weighting of boosters and barriers in Kirotshe health district in East DRC is available here. The prior mode from the community weighting should be worked out using the same formula as above. NB: All of the booster and barrier totals from the above exercises can be added to this template in order to define the average prior mode.