Sources: Aviation Stack Exchange

If a citation is going to point to any of the Stack Exchange Q&A pages, it is extremely important to specifically cite the exact post or answer, since – not dissimilar to Wikipedia – the quality, consistency, and biases of Stack Exchange answers is paramount for evaluating the information presented, especially factual data to be fed into an infographic.

I personally am intrigued at these $800 economy, ten-hour flights, as well as a total omission of freight cargo in the underbelly. As presented, this flight has 180 passengers and runs for ten hours. This would suggest it’s not a common narrow-body, either the Boeing 737 or an Airbus A320, as even their largest available configurations can’t fit 180 people in a 2 class setup, let alone a 3 class setup. It could possibly be the Airbus A321, though.

My point is that if it’s a widebody aircraft or the A321, not hauling cargo would be some staggering malfeasance for a commercial revenue airliner. But I can’t follow-up on any of these queries, since the sources aren’t properly cited.

In agreement with the other comments, this is indeed a very dense diagram, specifically the right-side. Focusing on that some more, my chief concern is that this novel triangle representation is very easy to misread.

Let’s take the dot in the middle which has the arrow with “10M”. What would you say the car percentage for that dot is? The axis along the bottom of the triangle is labeled 0 to 100%, and the dot is just to the right of the 50% demarcation. So maybe 52% or 55% seems reasonable, yeah?

But the axis is deceiving: notice how the demarcation are all slanted at the bottom. The dot is actually representing about 42%, since although the axis is marked horizontally, the line which is 50% slopes north-east rather than straight up. You can see the 50% number itself is actually rotated 60 degrees counter-clockwise.

The public transit axis on the left of the triangle has its demarcations tilted clockwise by 60 degrees as well. Only the active transport axis matches the conventional Y axis.

For that UI/UX reason alone, I wouldn’t endorse this as a “great” depiction of statistical data. If a diagram can – intentionally or not – be used to mislead a casual reader, it’s not one we should put up on a pedestal.

I also had a gripe about the successive colors not being consistent for each mode of transport, but that’s minor and easily corrected. The tilted axes may require some reworking though.

Obligatory link to Statistics Done Wrong: The Woefully Complete Guide, a book on how statistics can and has been abused in subtle and insidious ways, sometimes recklessly. Specifically, the chapters on the consequences of underpowered statistics and comparing statistical significance between studies.

I’m no expert on statistics, but I know enough that repeated experiments should not yield wildly different results unless: 1) the phenomenon under observation is extremely subtle so results are getting lost in noise, 2) the experiments were performed incorrectly, or 3) the results aren’t wildly divergent after all.

I’m reluctant to upvote this, since it’s leaving out a lot of rather important caveats about the dataset. This depiction is presented as “the number of aviation incidents between the two giants since 2014 in the U.S. and international waters”. Here, “international waters” means the regions of the North Pacific Ocean, north Atlantic Ocean, and Gulf of Mexico, whose airspace services are delegated by ICAO to the United States, administered by the FAA. It’s not US airspace, but it’s administered as if it was, meaning accident reports get filed with FAA and NTSB, the source of this data.

The other caveat is that the total size of the Boeing fleet flying through FAA-administered airspace versus the total Airbus fleet is closer to 2-to-1, with nearly twice as many Boeing aircraft as Airbus aircraft, using 2018 estimates. This is including all the aircraft which US airliners currently operate, not just the newest ones they’ve bought in recent years.

Finally, in the reporting parlance, an aircraft “incident” means a non-serious injury event that happened. If major injuries or death occurred, that would be an “aircraft accident”. So an incident could include anything like:

• Returning to the airport because of an unruly passenger
• Another aircraft getting too close but not requiring evasive manoeuvres (aka minimum separation violation)
• Overspeeding of the aircraft, such as exceeding 250 knots while still below 10,000 ft
• An engine failure
• A door plug falling off, causing minor injuries to three people but no deaths
• A passenger getting their arm stuck in the toilet while reaching for their dropped phone

What reasons could Boeing aircraft have more incidents? Sure, they might be shoddily assembled. But it could also be a matter of fleet distribution: if Boeing makes more wide-body aircraft than Airbus, and thus carry more passengers, then passenger-related incidents would be higher represented for Boeing aircraft. Suffice it to say, this single graphic isn’t giving enough depth to a complicated situation.

but there’s no way for a reader to know if you’re an inept experimenter, got a bad batch of reagents or specimens, had a fundamentally flawed hypothesis, inadequate statistical design, or neglected to control for some secondary phenomenon.

I agree, to the extent that single, poor dataset can’t draw useful conclusions. But after (painstakingly) controlling for issues with this dataset and from lots of other similar datasets, there can still be some value extracted from a meta-analysis.

The prospect that someone might one day later incorporate your data into a meta-analysis and at least justify a follow-up, more controlled study, should be sufficient to tip the scale toward publishing more studies and their datasets. I’m not saying hot garbage should be sent to journals, but whatever can be prepared for publishing ought to be.

maybe experimenters are opting to do further research themselves rather than publish ambiguous results

While this might seem reasonable at first, I feel it is at odds with the current state of modern science, where results are no longer the product of individuals like Newton or the Curie’s, but rather whole teams and even organizations, working across universities and across/out of this world. The thought of hoarding a topic to oneself until it’s ripe seems more akin to commercial or military pursuits rather than of academia.

But that gut feeling aside, withholding data does have a cost, be it more pedestrians being hit by cars or bunk science taking longer to disprove. At some point, a prolonged delay or shelving data outright becomes unethical.

fine tuning your experimental design to get a conclusive result is a more attractive option than publishing a null result that could be significant

I’m not sure I agree. Science is constrained by human realities, like funding, timelines, and lifespans. If researchers are collecting grants for research, I think it’s fair for the benefactors to expect the fruits of the investment – in the form of published data – even if it’s not perfect and conclusory or even if the lead author dies before the follow-up research is approved. Allowing someone else to later pick up the baton is not weakness but humility.

In some ways, I feel that “publish or perish” could actually be a workable framework if it had the right incentives. No, we don’t want researchers torturing the data until there’s a sexy conclusion. But we do want researchers to work together, either in parallel for a shared conclusion, or by building on existing work. Yes, we want repeat experiments to double check conclusions, because people make mistakes. No, we don’t want ten research groups fighting against each other to be first to print, wasting nine redundant efforts.

or a significant result that you fear might need retracting later.

I’m not aware of papers getting retracted because their conclusion was later disproved, but rather because their procedure was unsound. Science is a process, honing towards the truth – whatever it may be – and accepting its results, or sometimes its lack of results.

My understanding – again, just from that book; I’ve never worked in academia – is that some journals now have a procedure for “registering” a study before it happens. That way, the study’s procedure will have been pre-vetted and the journal commits to – and the researchers promise to – publish the data irrespective of any conclusive results. Not perfect, but could certainly help.

The linked tweet in turn links to Adrian Barnett’s blog post: https://medianwatch.netlify.app/post/z_values/

The two large spikes in Z-values are just below and above the statistically significant threshold of ± 1.96, corresponding to a p-value of less than 0.05. The plot looks like a Normal distribution that’s caved in.

From my limited statistical understanding, the Z value measures standard deviations compared to the null hypothesis. This page on statistical analysis says:

Often, you will run one of the pattern analysis tools, hoping that the z-score and p-value will indicate that you can reject the null hypothesis

To reject the null hypothesis, you must make a subjective judgment regarding the degree of risk you are willing to accept for being wrong (for falsely rejecting the null hypothesis). Consequently, before you run the spatial statistic, you select a confidence level.

So from that, I understand the takeaway from the Z value graph is that if researchers are truly willing to publish studies which don’t reach a definitive conclusion, then the huge gap in the middle should be filled in. But it’s not.

And the danger is that valuable data from studies straddling the arbitrary p=0.05 line is simply being discarded by researchers, before ever reaching the journal. Such data – while not conclusive on its own – could have been aggregated in a metastudy to prove or disprove the effectiveness of medicines and procedures that have non-obvious or long-term impacts. That is a loss to all of humanity.

(Image credit: https://pro.arcgis.com/en/pro-app/3.1/tool-reference/spatial-statistics/what-is-a-z-score-what-is-a-p-value.htm)

A while ago, I read a book about how researchers inadvertently misuse statistical tests, along with how to understand what statistics can and cannot do, from the perspective of scientists who will have to work with datasets. It’s not terribly long, and is accessible with no prerequisite of any statistical experience. https://nostarch.com/statsdonewrong

EDIT: the author of that book has published its entire text as a website: https://www.statisticsdonewrong.com

I understand where you’re coming from, and fully agree that anytime someone goes to prison for something they didn’t do, society is doubly worse off: once because the wrong person has been jailed, and once more because the real culprit has evaded justice.

That said, what you’re describing is an issue with the practice of plea bargaining, not necessarily with giving less time for defendants pleading guilty. There are very compelling arguments that we should ban plea bargaining, as it’s extremely one-sided, among other things. But while plea bargaining is partly enabled because the sentencing guidelines allow leniency for pleading guilty, I would argue we should keep the latter.

As a society, we should incentivize people to voluntarily come forward and atone for their crimes. If a murderer pleads guilty and divulges the location of the buried body, the victim’s family can have a proper funeral service. But if that murderer instead flees, there’s a chance that officers can make an arrest, but there’s also a chance of successfully evading the law. Even if taken into custody, there’s no requirement that a hardened murderer needs to reveal the burial location, and our laws prohibit beating that answer out of anyone.

A principle in law is that different criminal behavior should be punished proportionally. Ruthless killing versus accidental death. An accident versus indifference to human life. A clouded conscience versus a maligned intention toward the victim’s family during a prolonged trial. This is what the sentencing guidelines seek to implement, moral hazards be darned.

I’m not sure I’d characterize any of the figures as “a super high chance of getting out of it”, unless you mean leaving in handcuffs. Bear in mind that defendants that plead not-guilty but are then found guilty at trial get a worse penalty than if they had pleaded guilty in the first place. The federal sentencing guidelines intentionally recognize that people who plead guilty are taking some responsibility for their crime, and so it shaves a few months off.

Defense attorneys are supposed to help a defendant weigh the bird in hand (a plea deal with the prosecutors) against the two birds in the bush (prospect of acquittal at trial). And that’s only if the prosecutor wants to even do a deal: they don’t have to, since sometimes justice cannot be served by anything less than a jury verdict. Other times, a lack of a plea deal is part of looking “tough on crime” or to set an example.

I phrased it that way because I’m also unsure as to how “ex-convict” should be used and how most people use it. I’ve heard other people say it to mean anyone who has been released from prison, although that doesn’t make much sense for someone who just serves their time.

As a result, so far as I’m aware, it’s colloquially ambiguous, and lawyers and jurists may have a more stringent definition they might use.

In the Pew Research article? I arrived at a trial acquittal rate of about 17%.

In fiscal year 2022, only 290 of 71,954 defendants in federal criminal cases – about 0.4% – went to trial and were acquitted, according to a Pew Research Center analysis of the latest available statistics from the federal judiciary. Another 1,379 went to trial and were found guilty (1.9%).

While that’s still about 1 chance in 5, that’s still some really bad odds when it comes to the matter of possibly being imprisoned. I imagine most Americans think they’d have better odds than that, but the data shows otherwise, to a scary degree.

Formally speaking, a conviction will attach once a defendant is found guilty by a trial court. Even while one or more appeals may be ongoing, it is accurate to describe the defendant as convicted. The status of a federal conviction sticks until such time the conviction is judicially overturned by a successful appeal, or when pardoned by the executive. But not clemency, which is a reduction in the penalty by the executive, but retains the conviction.

A person who has their conviction overturned or pardoned can no longer be accurately described as convicted. Although colloquially, it’s unclear if “ex-convict” is an acceptable description or not.

The latter group of defendants – the ones convicted by a jury – also receive heavier sentences, since the federal sentencing guidelines recommend that defendants pleading guilty before trial get a reduced severity score, potentially shaving months off the sentence, or omitting the custodial sentence entirely, replaced by probation.

Compared to the total number of federal defendants (using 2022 data), there appears to have been a slightly higher rate here of going to trial than defendants overall. Both sets demonstrate that when federal prosecutors bring cases, they don’t tend to miss. Also demonstrated is how federal trials rarely result in an acquittal.

Does this mean judges and juries are biased against federal defendants? Likely not, since again: federal prosecutors tend to only pursue a case they know they can win. Knowing this, it must be a tough job for federal public defence lawyers but someone has to do it.