On Research Validity

I seldom read from "Economist" because I don't like the overall research quality there. Recently a graph like this (you don't have to subscribe to see) caught my attention from a discussion (which I didn't participate in) in one of the famous social forums:


As everyone can see the height with which both sets of the dots spread, it is not VERY fair to use linear regression to conclude -- years ago a research says boys are better at science and maths than girls with similar data distribution, where variation is bigger than the difference between the groups. Although the validity of the claim was criticized, damage was done.

It would have been much more honest if the two sets of data were plotted with red and green to show the widespread overlapping, instead of pale blue and pale violet(?).

When there are more resources (aka high GDP per person), a country can act with more options when given against a challenge, and democracy is often needed to sort out the differences in people's opinion. However, it is a matter of choice what system to use. The graph clearly shows the huge difference in outcomes for rich and poor countries under the threat of different epidemics.

However, it should be noted that since the TYPES of diseases that break out in different living conditions will be different, comparison is not really fair too.

Also, 95% confidence is considered weak in the hard sciences... my guess is that with a 99.5% confidence, there will be no significant difference but since epidemics are not THAT often, insufficient data points justify the lower confidence level used,

The last thing I would like to comment on is the theory of "poor country does not see benefits by adopting democratic system in face of epidemics). To my knowledge the article doesn't address this issue (though I don't have a subscription so I cannot be sure, I inferred this from the discussion).

According to this graph, scientifically significant differences in the death rate start at around $3K-4K GDP per person at purchasing-power parity at 2020, and disappear at around $30K-40K... I assume all these data be in USD. If epidemics represent the sort of typical case where a government and its people need to make quick decisions and powerful actions to overcome the challenge, this research actually hints at "a narrow window" during which people can clearly decide that democracy is better,

If the system of a country can only be changed drastically without conflict when 99% of the people give consent to (in HK, 1% of dissent already means nearly 80K people in a single city, enough to start a large scale "riot"), Hong Kong has definitely "passed" the point where consensus stemmed from objectivity can be attained -- our PPP in 2018 is well over 60K.

Well, strange that my conclusion and the Economist's conclusion(pro-democratic) is different.

If you ask me to resolve the political dissent in Hong Kong, I would suggest people to acknowledge the fact that supporters of democracy would generally have an income, when changed to PPP, falling between 3K to 30K USD, while the average GDP PPP in HK be something around (at least) double of that.

Yes, the chasm that has widened in Hong Kong over the years smells like it is more than about "political" dissent or delusions that can be "brainwashed" away. My suggestion would be dividing the people into two systems -- the poorer ones might have the democracy they deserve, while the richest should just submit themselves to something else -- yes it is still one country two systems, just that the rich ones should bow down to another power, which is generally already so in the first place.

Scientific evidence sugggests the poorer people in Hong Kong would benefit from a truly democratic government. So what would the rich and powerful say about this? Support the poor (PPP <35K) to leave China since CCP will not allow this choice? Like how Denmark supported the immigrants who did not successfully become a productive member of the Danish society to get back to their origin country by giving them a bunch of cash?

p.s. Since China still has more than 600 million still earning 3K PPP or less per year, this research actually supports that China is not ready for democracy yet.

p.s.s. I have always enjoyed seeing different things from the same research than the paper's author claims.

題目 : 論文。報告。心得 - 部落格分类 : 學術文藝

tag : 心理歷史學