Tweet
Recent years have seen a number of fights over net neutrality. Recently kos reminded us ("What almost wasn’t") that
Almost four years ago, campaign finance "reformer" groups (dinosaurs, all of them), set out to destroy political blogging.
KOS is correct to see the historical importance of this battle, and in spades. The framers of our Constitution had no idea what blogs would become, but they knew something about the power of the free and unfettered press. In the 18th century, pamphleteers functioned like modern day bloggers. Perhaps the best known American pamphleteer was Thomas Paine.
A study five years ago showed "…that Internet connectivity is positively correlated with spread of democracy at high levels of significance. " (Co-Evolution of Information Revolution and Spread of Democracy, by Walter Frisch)
Obviously, this study needs to be updated, but it is clear that the Internet has become the "5th Estate," so if we lose Net Neutrality, we lose everything, including democracy.
Bloggers, do your stuff, and don’t let anyone stop you!
Bob in HI
16 Comments
High levels of significance?
Puh-leeze. The size of the effect needs to be quantified in order assess its practical importance.
Gratuitous example:
Make a few sufficiently precise (and accurate) measurements and you can show at miniscule levels of significance that 1 g is not 10 M/s^2. Well, that’s because it’s 9.8 M/s^2. But is the difference important for most back-of-the-envelope computations? No.
Did you actually read the study?
The complete study is accessible through the links provided.
BTW, the authors thank Al Gore, among others, for providing feedback.
In the summary, the authors state,
Bob in HI
Thanks, Bob. I especially love that with the Net, we don’t have to be independently wealthy to start “pamphleting”. That has to be a part of why some of “the powers that be” have their shorts in knot over it.
The powers that be are not happy. Most journalists jailed worldwide are bloggers. Telling the truth in times of deceit has become dangerous.
http://www.andhranews.net/Intl…..-77556.asp
Thanks Bob. I heard an interview on NPR this morning with Bill Keller from the NYT doing the usual whining about the state of Journalism and how bloggers will never be able to replace the NYT. He specifically said that bloggers would never be able to set up a Baghdad bureau. Evidently it never occurred to him that it might be possible for English speaking Iraqi reporters/bloggers to report to the American public directly via the internet. I don’t think Bill likes losing his seat of privilege as the arbiter of what is and isn’t news.
Yeah. Why don’t you point that one out to Spencer Ackerman?
Kellerman is terrified. We are coming for his job, and he knows it.
Just look at what Gannett did this past week to more of its outlets — downsizing thousands across the country, out of greed, in part due to the economy, but much of it due to the continuing trend of readers moving away from the static, push-only media to new media that actually listens and interacts with its readers and does its damned homework. They can become a preferred provider in information movement, or they can become a casualty at some point down the road.
Unless the powers that be can reduce the size of the pipe and own the throughput, they can’t be certain they will exist once the print-only newspapers are gone. What they haven’t reckoned with, though, is that information wants to move, it’s a force of nature, and it will find a way to do so even if it means moving back to tin cans and string.
That is a great quote Rayne — mind if I borrow it?
I think what Keller and the rest of his buddies fail to grasp is the very nature of their gate-keeping and privileged access has distorted their ability to do objective reporting. As a result people desperate for a better source of information have had to go elsewhere. And looky here, here we are. A lot of people pooling their resources and experiences can take different stories apart, look at them objectively from different perspectives and learn a lot in the process. Guess I’ve never been good at accepting the “official” story at face value ; )
Not my quote, per se, phred; here’s a source that started a paper on the subject, unfinished, which gives some of the background behind “info moves,” while discussing two other positions on information (copyright and transcopyright). Wish they’d have finished the paper…
If it’s a fundamental human right to be able to move unrestrained, all that is human including the information that humanity creates should move with it. That’s the critical distinction and relationship that Brand did not make when he first discussed information’s movement.
I cannot see a reason why users should not pay a fee — think of moving a car along a tollroad and the costs involved in creating that road, or the taxes required to create a highway. But there’s an enormous difference between charging a reasonable cost for enabling the movement of information, and becoming a virtual highwayman, holding information hostage until a ransom is paid. The tollroad provides value, the highwayman does not.
Which now brings us to the question: is NYT a tollroad or a highwayman?
Thanks for the link Rayne. I agree that the sense of “free” information is too often miscontrued as “gratis” rather than “unconstrained”. One of the hubby’s frequent mantras is that hardware should be free, i.e., one should be able to choose one’s hardware (phone, mobile device, etc.) independent of service provider (AT&T, Verizon, etc.) and software (MacOS, Windows, etc.). In both the cases the freedom sought is simply to not be constrained by the vested interests of big corporations desperately trying to stifle competition and innovation. Seems to me we are dealing by and large with highwaymen in our fixed-market economy.
Yes, actually I did read the report.
And as Gertrude Stein famously said about Oakland, there is no there there.
Simply stating that a result is statistically significant (or highly statistically significant or supremely statistically significant) is meaningless in terms of assessing the importance of the result. Statistical significance is computed relative to a null model, that is, it’s a point hypothesis. I don’t believe point hypotheses, so the question is not “is the parameter zero?”, it’s “how different is the thing from zero?”
Maybe there IS something there, but you can’t tell from the report, because the author doesn’t present any tables of his results. The data are ecological (concerning populations rather than individuals) in nature. Ecological correlations are typically much higher than correlations for individuals.
If you want to assess the importance of this work, you need to know the parameter estimates from his model (odds ratios would work fine, too), and confidence limits on the odds ratios.
Yup, the old “free as in free puppy” versus “free, as in free beer.”
I have no problem with taking the free puppy, but don’t tell me I have no choice, that I MUST take the f*cking Great Dane.
LOL! Honestly Rayne, never figured you for the lap dog sort… (ducking and running ; )
But the statistical significance of correlation coefficients IS measured in terms of “how different is the thing from zero?”
I will grant you, however, that statistical significance is not the same as significant in a meaningful sense. For example, the correlation between the number of storks in German villages and the birth rate might be statistically significant, but correlation is not the same as causation.
I am intrigued by the resulting confirmation of our founding father’s insight with respect to the importance of the “Fourth Estate.” To paraphrase the study’s conclusion, it is not that democracy supports a free press, but rather that it is a free press that supports democracy.
So, perhaps we can’t count on democracy to maintain Net Neutrality, but we need net neutrality to maintain and promote democracy.
Bob in HI
Not really. Statistical significance is not a measure of effect size. Consider the common, garden-variety T-statistic:
T = SqRoot(n)*(Y-bar – Mu0)/S.
If the population mean isn’t Mu0 (and since I don’t believe point nulls, it’s not), then T has a noncentral T distribution. The probability that |T| exceeds any fixed value approaches 1 as n grows without bound. Since the observed significance level (a/k/a p-value) is the probability of observing a central T exceeding the calculated value, the observed significance level approaches 0 as n grows without bound. This is (mathematically) true even if the difference between the true population mean and the hypothesized mean is 1:10,000.
The bottom line (and it’s not nearly as well understood as it ought to be) is that statistical significance is not strongly related to practical importance. It’s entirely possible that in a noise-laden process you can have evidence of a practically important result that isn’t statistically significant. It’s far too common that by taking inordinately large sample sizes that trivial effects are found to be statistically significant.
The way around this conundrum is to define parameters that measure the features that are interesting and important to the inferential problem. Once those are defined, use appropriate interval methods to estimate the parameters.
I haven’t even touched on the pseudo-problem of causal inference. (I call it a pseudo-problem because you can’t use statistics to address causal question. Statistics can suggest causal patterns, but you need science to figure them out.)
However, on the bottom line, I don’t need statistical evidence to believe that net neutrality is supporting mechanism for democracies.
OT – Bob, if you are checking this still, could you let me know the first name of Anita’s husband? (I’m formerly NZ Expat and went to grad school with Anita.)
Well, you can now make that former husband, Ben.
Bob in HI