On a fine Sunday morning I received a mail with content I didn’t agree with, on a subject that had the potential to cause mass hysteria and it seemed to be about creating (read manipulating) the data on the subject and get the public to back it up. If you are wondering how is that even possible, then re-think butterfly effect. Exaggeration of the situation here? maybe yes, but did it have the statistical probability in my head, I would say yes!
The movement/petition was organised on Change.org and for those who don’t know what is change.org, come out of what I can only imagine a rather comfortable cave and leverage one of the most trusted product of modern times, Google search engine! As baffled as I was with the content of the said petition and felt like expressing my disagreement, I started looking out for methods to do that on the change.org platform. To my surprise, I didn’t find anything on there which I could use to express my disagreement. I then took a step back to think about the implication of this one option not being available. Lets visit the case study here - Change.org works towards getting petitions created & signed. It attempts to identify respective power stakeholders who they think they can move by showing the numbers behind the petition. However, no where on their platform did I find the following details available -
a) Number of people that were reached b) Their nationality c) Age group d) Income bracket e) Educational qualification This would give a very good perspective on what nationality people, have had their say in the matter and is that even justified for other nationalities to have a say in a matter that doesn’t concern them or even if they did, should that be taken into account? Which age group, educational qualification, income bracket etc gives a very good idea of what sort of distribution we are looking at and can potentially help in identifying patterns or biases for a certain kind of opinion that they may hold. Generalising is almost never correct so they need to capture this data to even understand what is it that the whole thing is about and HOW the output needs to be interpreted. Once they have the data captured and the landing page to sign the petition is presented, they also need to have an option to say, don’t know, don’t agree and not signing etc. The rational is rather simple - reach a 1,000 people and with a 5% hit rate you have 50 signs reach 1,000,000 people and with 5% hit rate you have 50,000 signs reach 1,000,000,000 people and with 5% hit rate you have 5,000,000 signs So, even though the numbers are increasing and 5M is not a small number but it is still only 5% of the total dataset. The 95% opinions are not even recorded. Further breakdown of the 5% & 95% data sets will reveal patterns which will give insight but to blatantly ignore the 95% makes no sense. Then to not even know about the directly, indirectly affected or bias based data only raises questions and doubts about the data on the platform seemingly being grossly inaccurate. Couple that with the attempt to create a bias, I wonder if that data should not be considered for any data analysis? In my opinion, as a product, Change.org, absolutely needs to keep track of the reaches, engagements, more data segmentation and impact in addition to recording the nay sayers and NOT just the yes’s. Needless to say what I have come to conclude on change.org, I leave it to your brain and its individual/mass capability to decide how you want to see change.org and would you put your trust in the data they present? Comments are closed.
|
Archives
February 2023
Categories |