Asking ‘The Community’ for their opinion about a proposed community decision is at the core of Civix Semi-Direct Democracy. In short We Listen Loudly.
Civix, We Listen Loudly
We frequently hear WestVan staff or councillors or others use the phrase: ‘The Community’ favours or opposes project rezone. We make a big deal out of public hearings where 20 people may speak. Why are these statements likely meaningless? Because The Community is never defined and more importantly the question is not asked in a structured manner (for example during public consultations) that enables statistical analysis.
The very first step is to define ‘The Community’ (of WestVan). what people, exactly, makes up ‘The Community’? Taxpayers? Voters? Residents? It varies?
44,000 Residents (everyone, older than 16, full time only, etc)
30.761 people who were registered to vote for mayor or councillors in 2018 election
16,660 representatives of housing units in WestVan (single & strata)
11,969 single family residential units
3,614 strata multi family residential units
504 business & commercial units
11,818 people voted for the mayor or councillors in the 2018 municipal election
10,000 ? people who might vote in a WestVan legal referendum
For illustration, we will define the 31,761 people who were registered to vote in the 2018 WestVan municipal election as The Community (also known as the statistical Population).
Problem: How can we measure what The Community’s opinion is on a specific question like are you in favour of WestVan building a $38 million dollar Arts Centre? The simple answer is we could ask each and every one of the 31,761 people that we define as The Community and then somehow force all 100% of them to respond to our question. Or we could initiate a legal referendum which means people may be more motivated to participate in the vote. Or we could use science and utilize statistical sampling of The Community.
If we ask people who are not members of The Community (as defined above) or if some people refuse to answer our question then intuitively we know our tally of results will be incorrect. And what happens if there are 31,761 WestVan voters and if only 75 people answer our question in the mall? Again intuitively we know that whatever the 75 people think will not be the same as what 31,761 people think. .
It would be foolish to state that: “we asked 75 people shopping in Park Royal mall and 71% said yes we want a new Arts Centre and therefore we conclude that ‘the WestVan community’ wants a new Arts Centre.”
What if only 15 of the 75 people asked were WestVan residents? Again intuitively we know that 15 peoples’ opinions will not reflect the opinion of The Community of 31,761 registered WestVan voters.
There is a scientific process that can be used to ask a few people from a specific group (The Community) a question and then be able to conclude something about what The Community’s response would be. The general term to describe this science is statistical sampling. Check out a couple of videos below.
First we need to very carefully defined the group of people whose opinion we want (example WestVan municipal voters, Ambleside retail store owners, community centre weight room users) and that specific group is called the “population”. Then we have to know how many people are in the population. "For example, there are 31,761 registered voters in WestVan who could have legally voted for Mayor and Councillors in the 2018 municipal election.
Second, we start at the end statement we scientifically want to be able to make and work backwards. The end statement could look something like: We have surveyed all 31,761 registered WestVan voters [The Community or the population] and 100% responded to the: “specific consistent clear question” like are you in favour of the WestVan spending $38 million dollars to construct a new Arts Centre and pay for on going maintenance costs?
If 100% of The Community (31,761 people) responded, and the responses was that 75% of The Community oppose spending $70 million on building and operating a new Arts Centre, then we can safely conclude that 75% of ‘The Community’ opposes the construction of a new Arts Centre.
However, in the real world we would be unable to ask 31,761 people in The Community (population) and get responses from 100% of them.
We could use a random sample of 300 WestVan registered voters and ask them for their opinion about the Arts Centre. Assume the sample survey result was 67% (example only) of the people who took the survey, are opposed to the New Arts Centre.
We then extrapolate the sample results to give an estimate of what The Community wants; and because we are estimating from the sample result, we have to qualify what we say based on the sample size and the population size. For a sample size of 300 (and 31,761 population), we can say: “we are 95% confident that if The (whole) Community had voted, then the actual vote of The (whole) Community would fall in the range of (67% - 6%*67) and (67% + 6%*67) which equals a range of between 63% to 71%. The 6% is derived from the sample size and population and is called the margin of error.
Conclusion: Based on our statistically random sample of 300 WestVan voters, we conclude with 95% confidence that between 63% and 71% of all WestVan voters oppose the new Arts Centre. Explained as: if every resident was surveyed and responded then we would be 95% confident that the true actual opinion of all WestVan residents would be between 63% and 71% opposed; or for simplicity we could say that we are 95% confident that at least 63% of all WestVan voters oppose the new Arts Centre. [all numbers are examples only].
Being able to say that: We are 95% confident that 63% of all WestVan voters oppose the new Arts Centre is powerful, conclusive and superb for civix Direct Democracy governance model and our decision making process.
And that is the power of statistical sampling. Relative to the population size, we can ask a random sample of a very small number of people about their opinion and then generalize that to the whole population (The Community) based on a scientific method.
In essence, Civix’s core platform of Semi-direct democracy governance model relies on statistical sampling and then (using statistical mathematics) work backwards to make a statement like: Civix is 95% confident that at least 74% of all WestVan voters are opposed to the proposed Arts Centre; councillors would then vote accordingly.
Statistical sampling is the only way to be able to draw a conclusion about The (whole) Community based on the opinion of a small sample of people and applying statistical methods.
However, sometimes the results of the sample may be inconclusive or too close to decide and if so, then a public referendum is held; effectively the referendum increases the ‘sample size’ from 300 to possibly 9,000 and that should yield a more accurate measure of what The Community wants.
A random sample is critical and difficult to get in real life.
One challenge in applying the above theory is that obtaining a truly random sample of the population (residents or taxpayers or voters) is difficult and must be done with care. For one thing, chances are we do not know the key characteristics of 31,761 WestVan voters. But there are some facts we know. For example, about 11% of voters rent their housing unit. We know the percentage of people who were registered voters and who voted (38%) We also know how many people live in a particular postal code. And so we can ask questions about renting, voting and postal code where you live to help ensure statistical randomness.
We also know that it is difficult to get a true random response. For example, even though we theoretically select a random group of people to take our survey, we know that some people refuse / do not like to do surveys. Right there we have introduced an error that we can not control. Or if we use the internet to survey then we would miss the people who do not have internet access. Chances are we can not eliminate this bias and hence if our sample results are close to, for example, 50% we likely should not use the results to make any decisions. Then the fallback would be a legal referendum on the subject.