Comment by mlthoughts2018 | Hacker News Reader

Comment by mlthoughts2018 | Hacker News Reader

What is the color and size scale here? It could be deeply misleading, for example if it’s a linear scale counting number of anomalies in a bucket (like number of +2.0 anomaly measurements in a given country).

From the upper right, it seems like it’s I guess the average per year per country, placed in bins for -2, -1, 0, 1, 2, ... except how were the bins constructed? In reference to global temperature in some specific year? Country-average in some specific year? Very hard to read or know the significance of it.

I hate visualizations like this for exactly this reason. The video hits me in the gut and makes me want to believe the final two decades of data are alarming, but it’s not clear at all what statistical claim is being measured, and how unlikely that specific claim would be under some counterfactual distribution.

I’m not denying (or confirming!) the implication of the data. Just pointing out that this is sort of like “ooh shiny” shallow attention grabbing, which IMO does more harm than good.

Statistical visualizations should not feel like they have an agenda in the idiosyncratic choice of axes, scale, colorbar, etc. It’s just not clear to me here that it’s not got one.

The intent of this visualization is to hit you in the gut. If you want something more concrete, there’s plenty of it out there. We can have both kinds of things.

I disagree in the strongest possible terms. Manipulating presentation of data to create an emotional reaction so that policy decisions are made emotionally is not ok.

If a presentation of the data that conforms to standard techniques for an unbiased representation would not result in “hitting you in the gut” then that has to be accepted, and worked with, not sacrificing intellectual integrity to elide the data and change things for an agenda.

This is a fine sentiment (double entendre intended), but some people, enough to sway public policy, cannot be reached by a simple presentation of facts. If these people control policy, and one party to the debate has no qualms about controlling them through appeals to their gut, what should one do? If in 100 years humanity is reduced to an impoverished remnant living in the ruins of its greatness, will they be grateful that we kept our rhetoric pure? (Note, I'm not talking about lying, only about presenting information in a way designed to appeal to the audience's emotions.)

I'm trying to get you to invest in my fund. I give you a set of cherry picked graphs that indicate I'm earning an unbelievable amount of money with minimal to no variance. Then somebody else shows you a complete graph showing my longterm profits (which are rather different than what I lead you to believe) and other data indicating that I was being somewhat less than completely honest with you.

What's going to be your opinion of the value of my fund? Of me? Did I do anything wrong, or can I just claim I was only trying to "appeal to the audience's emotion"? Myopic thinking and Machiavellian logic are a great combination to undermine the very ends you aim to achieve.

In your example you are attempting to deceive. That is the immorality. Not the appeal to emotions.

What if he really believes in the potential of his fund, and simply choses the graph that better conveys his belief- even if someone else might object that it's not the complete or most neutral representation?

If he isn't deliberately omitting information that would lead to a different conclusion -- deceit -- or using emotional manipulation to imply facts not in evidence -- also deceit -- then he's simply wrong.

In the case we're talking about, the emotional impact intended is simply that the information is alarming. You can drily say "the house is on fire", or with flushed face and wide eyes exclaim "the house is on fire!" The latter involves emotional information in addition to the relevant facts. If the house is not on fire, both statements have the same problem: they are false. If the house is on fire, the second is more likely to prompt action. The absence of emotional cues in the first statement will itself be taken as an emotional cue by an ordinary audience. "The house can't be on fire in the way I would normally interpret 'the house is on fire' because that would cause panic, and the speaker isn't panicking. Ho, hum. Are there any sports on?" In fact, the dryness of the presentation is frequently presented as evidence that climate change is a fraud. Denialists say "if you really believed this is true, you'd be doing X, and you aren't, so ha!" The beauty of this gambit is that the goal posts are so mobile. If you do X, they pull secret expected behavior Y out of their pocket and repeat the argument. These measures of sincere affect are like sharks teeth. There is always another if you pull out the one you see. You're damned if you do and you're damned if you don't. So, setting aside these particular denialists as insincere, one's best bet is to try to indicate your affect in your presentation, because this is a heuristic ordinary people use to gauge the honesty of their interlocutors: does the implied urgency of their argument match their tone.

Take the presentation data and contrast it against the longterm patterns seen in temperature [1]. This presentation absolutely omits information and relies on emotional manipulation to imply facts that are not in evidence -- in particular that the change in temperature is unprecedented and practically linear. This is literally selective sampling to present a misleading picture -- the identical behavior in our little metaphor.

If you genuinely think you're right, you don't suddenly turn to something very much bordering on falsehoods to make a point. In any discussion the strength of an argument is generally going to be measured based on the strength of its weakest point. And arguments, such as this article presents, are incredibly weak as they appear to be deeply misleading when one is presented with big picture data. And suddenly everything you've ever said becomes tainted, even if the other 99% of it was completely and wholly accurate.

[1] - http://upload.wikimedia.org/wikipedia/commons/7/7e/Global_te...

Ironic to complain about misleading visualizations and then link to a graph covering the last 800,000 years to imply that the recent change is just business as usual.

Do you not see the extremely recurrent pattern? One that has been occurring since before humanity even existed? This is really what makes things so incredibly challenging, and also why articles like this are presenting data in a less than honest fashion. We are undoubtedly going through a cyclical upswing as part of a regular cycle of climatic activity, but at the same time we're also adding a significant amount of additional CO2 to the cycle. The point is that it would be negligent for people to not consider human activity as a contributing factor, but it is no less negligent for people to not understand that we're also going through an ongoing warming phase that would have happened even if we were a 0 carbon output species. What would change is not the warming, but the magnitude.

Instead, like on so many topics in the social media era, people 'pick sides' and radicalize on these issues. And media like this is published to help further the radicalization. All the while in this process people on both 'sides' ends up voluntarily retarding their own logic and reasoning by working backwards from a predetermined conclusion instead of working from the breadth of data towards whichever conclusion (or lack thereof) that they find the most logical position.

But this is just different sentiment. Personally, I think that people are very skilled at identifying when they are being spoken to unhonestly or dismissively.

> But this is just different sentiment.

My point precisely.

> Personally, I think that people are very skilled at identifying when they are being spoken to unhonestly or dismissively.

How? By consulting their gut? Why do certain people lie so frequently if lying doesn't work?

Besides, we aren't talking about lying here, only presenting true data such that it has a particular emotional appeal. The goal is not to produce any false belief or any action contrary to the audience's interest. It is only to get the audience to register the information with the organ they use to process such information effectively: their gut.

Why do people lie? Maybe people are prone to exaggeration when its an idea they like or it's somehow attached to a position or stake they have taken (of)themselves.

But really the point I was trying to make by saying it's just another sentiment: perhaps the number of people put off by incomplete statistics is more than the people swayed by the dramatic results

I wouldn't say they are skilled, given the enormous false positive rate.

Manipulating data to "hit you in the gut" amounts to propaganda, of which people can be (rightly) wary, regardless of the veracity of the ultimate message.

A lot of things are propaganda. Showing a crying immigrant child is propaganda. Showing a picture of mourning fireman on the day of 9/11/2001 is propaganda. In each case we may want to debate what the picture is trying to say and whether it is appealing more to the emotions than rational debate. Or not. It depends.

On the other hand, when someone sees the firefighter and says "What a propaganda. These pictures never bother to present a concrete evidence that two planes hit WTC," then we kind of know rational discourse is over, which is what's happening with climate denialism.

> “On the other hand, when someone sees the firefighter and says "What a propaganda. These pictures never bother to present a concrete evidence that two planes hit WTC," then we kind of know rational discourse is over, which is what's happening with climate denialism.”

Let me state first that I do believe climate change is real and urgently needs to be given increased attention to find a solution.

But, secondly, your comment falls completely flat, and in my personal view, you are espousing something that is truly far, far more dangerous than even climate denial.

You’re attempting to say that a rational analysis of entire patterns of propaganda is equivalent to no longer participating in the rational debate. You’re trying to say that anyone who does not agree there is intrinsic credibility to the data-eliding emotional appeals must themselves no longer have credibility in the discussion.

This essentially lobotomizes our best and only chance to solve problems like climate change long-term, which is to use science as a constraint on political gamesmanship.

When you say, “... then we know rational discourse is over” you’re just falling right into politicians’ hands, who want to continue politicizing the issue while not actually doing anything about it that doesn’t happen to also serve their short-term profit interests, and figuring that if younger generations have to inherit a wasteland and figure out how to live in it, that’s not their problem, and they’ll happily consume now, while the getting is good.

Your type of meta-comment is the most frustrating to me in this whole debate, because you seem self-aware enough to know better.

Who says it’s manipulated? The complaint was that there wasn’t sufficient detail. Are we just assuming that any chart that doesn’t live up to the standards of a scientific paper must be manipulated?

Omitting sufficient detail to understand the scale and significance of a result is one of the most common types of manipulation in statistical data presentation. Not saying that this post was specifically trying to be manipulative, but it is very reasonable to presume the goal is manipulation and work back from there, since it is so, so common in published academic literature, newspaper infographics, etc.

So, you’re not saying it’s manipulative, but it’s reasonable to assume it is? That seems contradictory.

Why is it contradictory? I don’t have special knowledge to confirm if this was manipulative. But since manipulating the presentation of data is so common in academic papers, newspaper infographics, etc., it’s a reasonable prior belief. Generally being skeptical of data presentations is reasonable.

The scale is visible in the right top corner (literally in every frame of the video). Did you just happen to overlook it or are you asking something else?

I meant that the scale in the upper right is uninformative.

If the circles’ sizes and colors were based on country-specific average anomaly divided by country-specific standard deviation of anomaly, based on a post-industrial reference year, say in the 50s-80s, then the colorization and size changes would possibly be useful.

Basically bucketing Wald statistics instead of temperature anomaly, and ensuring it’s all always country-specific.

Reply on news.ycombinator.com