Marco's case solution.https://crowdflower.atlassian.net/browse/ENT-912

gistfile1.txt

I added an "aggregation=avg" to the cml:checkboxes element. I copied the job and ran it on the internal channel and submitted my judgments. The results came out in the hidden field "results_percentage", so at least that part works.
As for making test questions, I don't know how to do that, because as they said in their email, the choices are created dynamically in javascript. If he wants to completely re-do the way he's built the job, here's one idea I can think of.
For each unit in their csv, create new columns titled
"category_1_name"
"category_2_name"
"category_3_name"
"category_4_name"
Then split out the contents of the column "categories" into those columns. In the cml he would do
<cml:checkboxes label="" validates="required" instructions="" name="" class="" gold="true" aggregation="agg">
{% if category_1_name != '' and category_1_name != 'No data available' and category_1_name != empty %}
<cml:checkbox label="category_1_name" value="category_1_name"></cml:checkbox>
{% endif %}
{% if category_2_name != '' and category_2_name != 'No data available' and category_2_name != empty %}
<cml:checkbox label="category_2_name" value="category_2_name"></cml:checkbox> 
{% endif %} {% if category_3_name != '' and category_3_name != 'No data available' and category_3_name != empty %}
<cml:checkbox label="category_3_name" value="category_3_name"></cml:checkbox>
{% endif %}
{% if category_4_name != '' and category_4_name != 'No data available' and category_4_name != empty %}
<cml:checkbox label="category_4_name" value="category_4_name"></cml:checkbox> 
{% endif %}
<cml:checkbox label="None of the above" value="none"></cml:checkbox>
</cml:checkboxes>
So it would only show the choices if they are in the data. Then he can dig gold.