Wednesday, 1 October 2014

Basis Of Singaporean Spirit - People Power: Logical Analysis vs Statistical Analysis

[This is part of the book: Basis Of Singaporean Spirit - People Power, click here to go to the main page]

The purpose of this chapter is to highlight that broadly, there are 2 ways to validate a claim – using logical analysis or statistical analysis.

Both methods are valid sciences and claims made using either method are all valid (assuming you know all the assumptions made).

There are differences in the methods and there may be situation where one is preferred over the other, or a combination of both is preferred.

Since this is a simple book for commoners, I am not going into the details of the definition; instead, I will go through a simple example to illustrate the point.

Imagine yourself in a shopping mall.
When you see many people coming into the shopping mall with wet umbrellas, it may strike you straightaway that it is currently raining outside. You could arrive at this conclusion using either logical analysis or statistical analysis.

For logical analysis, you based your conclusion on a number of premises (or assumptions). When these premises are true, the statement (or claim) will be true. For example:
Premise 1: People will use umbrellas to go to shopping mall when it is raining.
Premise 2: When an umbrella is used in the rain, it will be wet.
Claim: When we see people bringing wet umbrellas into shopping mall, it is raining outside.

For statistical analysis, you may have interviewed many people who brought wet umbrellas into shopping malls and obtained the following results:

Why people bring wet umbrellas into shopping malls?
99% - It is raining
0.8% - Umbrella was dirty and it was washed
0.2% - Others

Claim: When people bring wet umbrellas into shopping malls, there is a 99% chance that it is raining outside.

One of the problems with logical analysis is that the premises could be ill-formed. In this case, the umbrella could be wet for other reasons, e.g. after being washed, which is not depicted in the logical analysis.
One of the problems with statistical analysis is in the representativeness of the sample. The people that were pooled may not be representative and the statistics may not be appropriate to be extrapolated for other situations.
No matter which analysis is used to make a claim, one has to understand the potential limitations.


You may also compare logical analysis with statistical analysis by imagining a domino as shown below.

#########################################################################
[This section is on Technical explanation which is removed from this online version]
#########################################################################

Many of us know the basic of statistical analysis, such as average, median, etc. However, not many of us know about the logical analysis above.

Logical analysis is an important tool for analysing a claim because it served as an alternative as well as a complementary method to validate a claim. One of the aspects of logical analysis is what I called the "reflective qualification".

For example, if leaders said a particular thing is good, how do you know if it is really good? Well, by reflecting that statement back onto themselves.

Imagine this: A president of a country showed you some statistics on a survey of people, stating that more than 90% are proud to volunteer for combat vocation[i] and urge you to do the same.

Statistically, it sounded convincing and you may even have the urge to volunteer for that.

Applying logical analysis with reflective qualification, you will define the premise as "If combat vocation is really that good, the president's son should also be in combat vocation".

However, if you look at the president's son, who was medically fit, but was not in a combat vocation, you would be able to see a different perspective of the situation and question the validity and motivation of that statement.

Another application of logical analysis is using what I called "logical contrasting". For example, if ministers provided statistics stating that $5 a day is enough to live, but on the other hand, they had demanded for over $3000 per day of salary, more than enough to support the livelihood of 600 people (applying their statement), yet still claimed that they are underpaid. With this contrast, what do you think is wrong?

And of course, let's apply logical analysis together with statistical analysis on the statement that "you do not need degree to be successful in life". Well, without a single doubt, that statement is definitely true.

Anyway, let's apply the reflective qualification to the ministers first. Assuming they have 4 children and statistics showed that 25% of Singaporeans have a degree. If there are no deliberate actions, statistically, we should see that the ministers will have 1 child who is a degree holder, while the rest of them are not (25% quota). Is this the current situation? Or are all of them degree holders?

Now, let's apply statistical analysis for the situation that all of those 4 children are all degree holders. That would mean that there must be 16 people who are "not supposed" to have a degree (to maintain the 25% quota).

So where do you think these 16 people should come from?[ii] What do you think is the motivation for them making such a statement? I don't know, ask them.

Logical analysis is a very powerful tool to enable you to see and question people's motivation of making the claims and assess the validity of their statements.
Statistics is useful, but it has limitations, worst, it is often misused by people for ulterior motives.

For the purpose of this book, logical analysis is widely employed over statistical analysis mainly because statistical analysis, although very useful, is subjected to a very high probability of manipulation, either intentionally or unintentionally. [See "Please Spare Us the Statistics" for more details]

In summary, you should always question the basis of the assumptions of the statistics as well as employ logical analysis in order to have a complete assessment of the situation.

It is only when every one of the People understands and able to apply logical analysis, would their leaders start to think twice before they say or do anything, because they now know that the People will reflect their statements back onto them.

[This is part of the book: Basis Of Singaporean Spirit - People Power, click here to go to the main page]


[i] In Singapore, all males must serve at least 2 years of National Service. "Combat vocations" are roles that need to go to the battle field in times of war while "Service vocations" are roles that typically stay in the office and do not need to go to the battle field.
[ii] Examples are meant for illustration of concepts, not meant for asserting whether having a degree is better or not.

No comments:

Post a comment