Follow Slashdot blog updates by subscribing to our blog RSS feed


Forgot your password?
Take advantage of Black Friday with 15% off sitewide with coupon code "BLACKFRIDAY" on Slashdot Deals (some exclusions apply)". ×

Comment Published articles 1/1000 wrong -- PLOS article (Score 1) 208

Classical statistics mentions the significance level, alpha=0.05. It mentions beta -- (1-beta) is the power of the test to conclude the null hypothesis. Classical statistics never mentions R, the background ratio of true to false relationships in a field. While R lies in the interval [0,infinity], you could think instead about the background probability of true relationships. PLOS had an article several years ago that showed the probability a published article falsely touts a relationship as true, a probability they called the Positive Predictive Probability,
      PPV = 1 / [1 + alpha / ((1 - beta) * R))]
The person designing an experiment seeks a large power, 1 - beta, so is bounded away from 0 and at most 1, so this factor becomes irrelevant (remember, the article gets published). When R is much less than alpha; eg, R=0.001 is less than 0.05, then PPV is about
        R / alpha
or often
      R / 0.05
The background proportion of true relationships R dominates over alpha and over beta in the probability the relationship is true PPV.

You do a statistical test in a "field" of relationships where most of the relationships are wrong, otherwise any relationship stated has a good chance to be correct and the "field" is easy if not boring. Consider the search for some 30 genes that might cause a genetic disease out of 30,000 genes in a genome. Then R is 1 / 1000 and (about)
      PPV =. 1/(1 + 0.05/(1/1000)) = 1/51 =. 0.02
That is, such published genetics articles tout relationships that are very unlikely (0.02) to be correct.

The German pharmaceutical Bayer called a large sample of published article authors, duplicated their procedures, yet found 70 percent of the publications' touted results could not be confirmed (probably wrong). Many statistical tools will give fame -- hypothesis tests or even more so data mining tools -- these are often charlatan's tools.

Comment Warning: statistical ECOLOGICAL FALLACY likely (Score 1) 175

The ecological fallacy concerns making conclusions about individuals from aggregates (states).
From Wikipedia,
"An ecological fallacy is a logical fallacy in the interpretation of statistical data where inferences about the nature of individuals are deduced from inference for the group to which those individuals belong. ... The four common statistical ecological fallacies are: confusion between ecological correlations and individual correlations, confusion between group average and total average, Simpson's paradox, and confusion between higher average and higher likelihood."

An example is Red State/ Blue State and income.
Using states, we could conclude that poor states (southern or red states) vote republican.
Yet when run on individuals, rich people vote 60% republican.

The same thing happens in the pharmaceutical industry.
Rather than states, some researchers merely use other researchers published (aggregate) results,
collecting results from many academic pharmaceutical articles (each acting like a state).
Bayer found that they could not reproduce 75% of pharmaceutical academic articles.
When you aggregate (meta-statistics), the knowledgeable complain that you should use individual data, not aggregate data (from articles) to make proper conclusions.

Further examples of the "ecological fallacy" are numerous.
State confounds with another variable, and its extrication can take decades of research, though Bayesian hierarchical models with separate errors at each level can probably extricate the problem these days.

Using aggregate data to make conclusions about individuals has been rejected since the 1950 seminal article by the researcher William S. Robinson. That's 60 years ago! We should become more statistically literate!

Comment Vernier has done this for 30 years (Score 4, Interesting) 6

Particularly for schools, Vernier Software and Technology, has continually added sensors, now about 60 sensors, over the last 33 years.
Some sensors include carbon dioxide, water flow, radiation, respiration, soil moisture, spectrometer, UV.
I have about 20 of them for my child at home, and hope to get the blood pressure sensor for myself.
Being for schools, they are the least expensive sensors I've seen.
All these sensors plug into their Labquest 2 interface (or one of their older interfaces) which looks more like a smartphone with touch screen, WiFi, Bluetooth, and several device ports. Top 10 ranking high schools like Thomas Jefferson High School in Fairfax county use these, though that high school has also had genetic sequencers for the last 15 years.

I'm glad to see competition, but one should never overlook what has dominated this arena for decades.

Comment Fund Open Office rather than fund Microsoft market (Score 5, Interesting) 273

If you want a feature in Open Office, fund it. Better yet, considering the cost of Microsoft Office, put the funding of Open Office in the annual budget. Rather than giving $100 million a year to Microsoft, give $10 million a year to Open Office. With a programming / total-expenditures ratio of 1, open source funding is efficient.

Comment Nebraska AG lambasted economist for honest results (Score 1) 640

A few years ago, Kansas sued Nebraska, whose 100,000 plus wells left some Kansas towns with zero water, especially near the Republican River (named after a Pawnee subtribe of Indians known as the Republicans). A Nebraska agricultural economist investigated, publishing a research paper that revealed over several years that Nebraska's increasing wells decreased water before it reached Nebraska. With his defense ruined, the Nebraska attorney general denounced the economist to his department head in Lincoln, Nebraska. The attorney general valued loyalty to himself more than honesty to truth, while the economist valued loyalty to honesty.

Comment Security agencies blemish our view of government (Score 1) 470

Continuing revelations about U.S. security agencies (torture, forbidding free speech, spying on their citizens, promoting specific denominations of Christianity) blemishes all other government agencies (Commerce, Agriculture, Education). We envision these other benign government agencies' surveys spying on us, maybe even sharing information with government security agencies. Does U.S. extensive security represent a new necessity, obsessive employees, employees seeking promotions, or a cowardly and impotent population.

Comment Re:Use SDHC memory in a card reader-writer, set lo (Score 1) 116

I have now actually checked this.
I switched an SDHC to read-only, wrote a file to it on Linux, took the SDHC to another computer, and the file was indeed written.
So, the SDHC lock is no guarantee against writing, and is apparently useless.
I stand corrected, and thank Carnildo for ending my misadventure.

I prefer using read-only hardware to "chattr -i" immutability plus a Linux kernel enforcing this,
since the software approach is cumbersome and changes files' ctime attribute.
What is available?
The following in the alternate model AEPDDESUWP will not write to any memory it can read,
and outputs to either eSATA or USB computer ports,
I still need to put my operating system on flash memory before I insert it into such a read-only device.

Comment Americans prefer to buy the goods of Satan/China (Score 1) 1121

People are either Christian, or headed for hell and of Satan.
Most American goods come from China, so American Christians prefer to buy the goods of Satan.

So with science and engineering -- of Satan.
Yet virtually everything around us came through science and engineering's magic touch
-- blended shirts (banned in the bible), invisible electromagnetic waves everywhere to be interpreted, risen people in airplanes, medical prosthetics getting even amputees to walk.
In the last hundred years, name one "good" thing religion has given us.
Now name something good science has given us.

Comment Use SDHC memory in a card reader-writer, set lock (Score 2) 116

No matter what an intruder tries, if you put your operating system on read-only media, intrusion becomes limited.
Of course, installation and changes become more difficult because you must reboot with your media set to read-write, then reboot again to read-only. SDHC memory works well for this, since it has a read-write switch like the old floppy drives. Put the memory in a
      USB "card reader" for SD
(microSD doesn't appear to have a read-write switch).
You can insert the SDHC in something that looks like a flash drive, then insert the whole in a USB slot.

Or, you can use something like the Adonics eSATA/USB Digidrive
to connect to your computer's eSATA port (if you have such a port on the back of your computer),
which is probably more efficient (fewer waits) than a USB 3.0 connection.

In Linux, you might choose to put most of your operating system on SDHC switched to read-only,
then put a variable area on a regular disk drive for logs, although you can put logs into a memory area that disappears on reboot.
Or you might put your webpages on a separate SDHC,
so your webpages get no intrusion changes.
You could then unmount your webpage SDHC, switch to read-write, make changes, unmount, switch to read-only.

In Debian Linux, the foundation for most Linuxes (eg, Ubuntu), you can look at the "Securing Debian Manual",
Debian has a highly tailored Aide (like tripwire) that uses checksums to detect any file changes.
In Debian, "dar" Disk Archiver (like tar) makes backups on external disk drives, but dar probably requires some tailoring (I use dar).
For a firewall, you could use Debian's easily used Guarddog.
In some sense, Debian is the administrator's operating system -- for the serious.

Comment In experiments, correlation implies causation (Score 1) 223

Data arises from retrospective or prospective studies.
Retrospective data was created before a statistician could design an experiment.
Prospective data sees the statistician set various levels of a variable
to randomly selected experimental units (maybe people, maybe production machines).
In a (prospective) experiment, an observed correlation implies causation.

For example, in manufacturing plastic, keeping constant other variables (humidity and speed of production),
set the temperature sometimes at 100 degrees and sometimes at 200 degrees,
randomly choosing the order these temperatures get applied.
If the 200 degree temperature produces a stronger plastic (response or dependent variable),
then your positive correlation implies causation.
In the future, knowing that increased heat increases plastic strength, the manufacturer would raise the temperature.

But experiments consume time and money, so institutions not individuals usually perform them.
Million dollar clinical trials do determine whether a drug is effective.
While experimental economics can determine causation, most economics is retrospective, so conclusions become controversial.

Comment Mississippi State did this with GM (Score 1) 543

GM and Department of Energy have been sponsoing a competition, providing the same model vehicle that universities then convert.
Mississippi State University appeared on the Smithsonian Mall (Washington, DC) in the Folk Festival, June-July 2012.
While the vehicle wasn't an SUV, GM provided the same vehicle model merely to avoid variable results arising from model used.

Comment R -- programming plus math, graphics (Score 1) 355

Programming plus math, statistics, legendary graphics -- that's R.
Just a programming language is a barren language for most people.
Use a programming language that eases and magnifies another field/course.
R dominates numerous statistics departments, biogenetics, and financial quants,

R originated from AT&T (as S) at the same time as Unix and C.
Many people start an R session whenever they login, since R can even do the trivial as a calculator with history and help.
With vectors x and y, the following creates a good graph.
R has associative arrays (hashes) and parallel programming, native to R since the programmer often uses vector objects.


Westinghouse Electric Company was bought in 2005 for $5 billion by the Japanese company TOSHIBA,

And the U.S. has the British company BP drilling much of the U.S. oil.
Why is the U.S. excited about its self-sufficiency?
Why does the U.S. push foreign company energy projects more than its own projects?
With China producing 30 percent of the world's engineers, Russia 7 percent, and the U.S. only 3 percent (see this week's Science article); and
with 8 out of 9 of China's political leaders engineers;
what part of the world's engineering curve does the U.S. think it sits?

We all agree on the necessity of compromise. We just can't agree on when it's necessary to compromise. -- Larry Wall