Correlation and causality

We have two sentences:

Sentence 1: In 1979, a followup study by Combs and Slovic showed that the skewed probability judgements correlated strongly (.85 and .89) with skewed reporting frequencies in two newspapers.

Sentence 2: murders are more available to memory because they are more reported-on

It is claimed that Sentence 2 need not be true. So we made 3 hypotheses yesterday and are now attempting to disentangle the puzzle. The first two of the hypothesis seemed to be not probable. In the last hypothesis the idea that correlation, might not mean causation, was thought about.

A quick search reveals that indeed correlation need not imply causation. In this site there is something really cool informing the absurdity of a statement that “all correlations imply causation”. A link to Tyler Vigen’s site called spurious correlations, is provided here as well. This seems to be the cure for thoughts that correlation and causation are related. It takes absurdity to a whole new level to prove its point. Thank you Tyler.

A couple of examples where correlation does not need to mean causation, from the Tyler Vigens site are the follows: 1) The number of people who drowned by falling into a swimming pool correlates with the number of films Nicolas Cage appeared in. 2) US Spending on science, space and technology correlates with suicides by hanging, strangulation and suffocation. 3) Per capita cheese consumption correlates with number of people who died becoming entangled in their bedsheets.

I cannot imagine that Nicolos cage movies cause swimming pool related deaths or the other way around. But that doesn’t need to mean that it is false. Just that further investigation is needed to confirm if there is causation. Or can we just reject the hypothesis?Correlation is the first step after which controlled studies can reveal if there is causation or not. The only thing we derive from the above examples is that we need to investigate if correlation would mean caustion for a given case. I guess this is what Eliezer was talking about. :)

Selective reporting

Selective reporting can be defined as reporting that does not cover all activities equally. For example, if murders are more often covered as compared to suicides depite the higher frequency of suicides, we term this as selective reporting.

Ideally we would like to avoid selective reporting as it further deteriorates the quality of the information we possess, resulting in poor quality of future judgements that are based off of the ‘availability bias’.

In ancestral days there was little going on and very few people in a clan. The effect of selective reporting would be quite minute compared to our current world where so much happens and even lesser gets to our ears. In the ancestral days, information would filter down say about 6 times before it reaches you, as you were in a small clan and things got passed on by word of mouth. But now, there are many more people before information reaches you and hence a heavy/large selection is at place. For example, the information passed on by religion got selectively reported over time and amongst families and regions, that my mom says one thing should be done to satisfy the gods, where as my neighbour of similar religion says another trivial thing needs to be done. If this were in a small tribe, there is very little chance I guess that information would be as is now- severly selectively filtered.

As we saw earlier selective reporting affects availability bias. It is expected that with more severe selection, our availability bias is affected even more.

Open Issue

How is it concluded by Eliezer that there is an availability bias at work, while talking about the 1979 study?

Can we just show the middle finger to correlations that we think might not be causation related.

Discussion on bill gates, what does he want to say with

The objective frequency of Bill Gates is 0.00000000015, but you hear about him much more often. Conversely, 19% of the planet lives on less than $1/day, and I doubt that one fifth of the blog posts you read are written by them.