SURVEY RESULTS Mark Kimura did a survey between October 9-12, 2014. There were 468 valid responses from the online version of the survey and 116 valid responses from the paper version (at Pahoa Community Meeting on 10/9). There were some respondents who do not live in lower Puna but have businesses, properties or job in lower Puna–since Kimura specifically asked those who currently live in lower Puna and those who recently left due to the lava flow, he had to exclude those responses. He may have to consider including them in a future survey. Also, Kimura will not include confidence intervals for every result– “If you’ve taken a stats course, you can estimate it by the sample size, population size and the percentage,” he writes about the survey. “I don’t want to make the picture too messy by adding every technical detail.”
This chart has some notes from Kimura:
Note: I really should have included “Household work and retired as separate options. For this time around, I included them in “I do not work” just for the sake of estimate the number of residents who need to commute regularly. Household production is unambiguously defined as work by most economists as far as I know (although it’s difficult to quantify).
Note: These numbers don’t add up to 100% because I consolidated Upper Puna, Keaau, and Volcanoes even though they are separate answers in the survey. It’s an error I made in the survey design and the result you see here includes the compromise I had to make as a consequence.
Note: The percentage for “Hilo” is 12-21% with 95% confidence level. 17% is the mean value from this particular sample.
Back-of-the-envelope calculation: One of my older infographics shows that there are 2,254 workers in lower Puna. And according to this survey result, 43% of the “workers” in lower Puna go somewhere outside of lower Puna to work, which is approximately 1,000 people.
If you have never taken statistics and are curious what I mean by “confidence level” or “confidence interval,” read on. Let’s think about this case: You want to find out the ratio between men and women in the world. You already know the answer–it’s approximately 1:1 (slightly more men in reality). Of course, you can’t count all of the world’s population, so you need to take a sample–or just pick part of the population. That sounds good, right? But if you just pick 3 people for your sample, and two of them happen to be male and only one is female. Would you conclude that 2/3 of the world’s population are male? Of course not. You know that would be a “sampling error.” This error becomes much smaller as you increase your sample size, but you are never free from it. Statisticians came up with a way to deal with it; you can never accurately tell the real percentage, but you can say, like, “the real percentage lies somewhere between A% and B% and we are confident about it by X%.” As you increase your sample size, the range between A and B gets narrower, or you will be more “confident” about the mean value you got from your sample. The convention for X is usually 95% or 99%.
Notes on above graphic:
A lot of respondents who answered, “I’m going to stay no matter what,” for the previous question also chose to answer this question, and Kimura decided to include their answers–perhaps their previous answer was more about their level of determination rather than real options, or perhaps they answered this question as a hypothetical situation even though they are truly determined to stay.
Either way, this question is more about what they COULD do IF they lost their houses or if lower Puna became unlivable for any reason.
Regarding “Other” –several of them were “camping,” which Kimura says he should have mentioned in the image. Most of the other “other” answers were hard to classify, too specific or personal to mention, too generic, too much information to summarize or not answering this specific question. Examples: “Island Chain, Kingdom of Hawaii is my hohm” (unclassifiable), “Renting a place if necessary” (too generic – where?), “I’m gong with the flow” (maybe an answer in a sense, but not an answer to what we are trying to figure out here), “I have options, but undecided yet” (reasonable, but didn’t check other answers).
(The following graph is not part of the survey, it’s a distance chart that Kimura puts out periodically.)
Reference: USGS Maps
Back to the survey:
Note on above graphic:
Some people chose “Other” and explained their decision making process. “I read each of them and sometimes I had to add their answers to ‘I haven’t decided…'” Kimura explains. For example, if somebody says they are leaving if the lava hits their houses or if the air quality becomes unbearable, in which case the answers should fall into “I haven’t decided. I will decide depending on the progress of the situation.” The important message here is, a large number of people haven’t made up their minds and a lot of people are determined to stay at the same time.