Data Validation Quotes

You don't want to come out with anything that's wrong, of course, in a scientific, you know, a major scientific announcement, and so you're being so careful trying to check, well maybe it's this, maybe it's that, you're looking at every possible thing.

No one believes an hypothesis except its originator but everyone believes an experiment except the experimenter. Most people are ready to believe something based on experiment but the experimenter knows the many little things that could have gone wrong in the experiment. For this reason the discoverer of a new fact seldom feels quite so confident of it as others do.

My goal was never to withhold information; my goal was to make sure everything was accurate, and frequently it wasn't.

When developing intelligence assessments, initial tactical reports often require additional collection and validation.

The data stream has been corrupted, return to first principles.

I thought I'd better check this third plate, which is another date, see if there's an image there in the right place that would be consistent with the images on the other plates. That was the final proof.

Now we have a whole separate supplier of data for Able Danger who's verifying that same information.

Okay, are you really 'experimenting'? How do you know if you're really experimenting? You're working on it methodically and you're publishing the results! It's not an experiment if you don't publish the results in some verifiable and falsifiable form, okay?

If the data do not prove that indexing wins, well, the data are wrong.

Rob Engle and I are concerned with extracting useful implications from economic data, and so the properties of the data are of particular importance.

If I see something that seems out of sync with what's already known, the first thing I do is try to find out what's wrong with the data. Once you've done that, and it still seems wrong, that's when things get interesting. It means you've found something new to understand.

In science consensus is irrelevant. What is relevant is reproducible results.

A theory is only as good as its assumptions. If the premises are false, the theory has no real scientific value. The only scientific criterion for judging the validity of a scientific theory is a confrontation with the data of experience.

You've got to be happy if they get your facts right. Since January I don't think I've recognized a damned thing that I've filed. I just pour everything out of the boot. Otherwise you get a phone call at three in the morning asking why you left out that the candidate had his teeth drilled that morning.

On two occasions I have been asked, 'Pray, Mr. Babbage, if you put into the machine wrong figures, will the right answers come out?' I am not able rightly to apprehend the kind of confusion of ideas that could provoke such a question.

To no circumstance is the wide diffusion of error in the world more owing than to our habit of adopting conclusions from insufficiently established data. An indispensable preliminary, then, in every investigation, is to get at facts. Until these are arrived at, every opinion, theory, or system, however ingeniously framed, must necessarily rest upon an uncertain basis.

I cannot understand for the life of me why DOE is going forward with this licensing procedure when we do not know whether or not the scientific documentation upon which you are basing your decisions is, in fact, flawed.

It doesn't matter how beautiful your theory is, it doesn't matter how smart you are. If it doesn't agree with experiment, it's wrong.

If a columnist writes that something happened on a certain date, or that the government spent a certain amount of money on something, or that a specific number of people have died in the war in Iraq, to pick a few examples, it is his or her responsibility to make certain that information is correct.

Get in the habit of vetting your research as you go - particularly research conducted online. Verify facts from multiple reputable sources before you record them.

Still, it is an error to argue in front of your data. You find yourself insensibly twisting them round to fit your theories.

One of the distinguishing features of anything that aspires to the name of science is the reproducibility of experimental results.

Careful economic analysis does have one important benefit, which is that it can help kill ideas that are completely logically inconsistent or wildly at variance with the data. This insight covers at least 90 percent of proposed economic policies

If I get that thing down there in that area, that's 67% lifetime. If you don't believe me, Google it. I'm on the Internet.

When reputable scientists correct flaws in an experiment that produced fantastic results, then fail to get those results when they repeat the test with flaws corrected, they withdraw their original claims. They do not defend them by arguing irrelevantly that the failed replication was successful in some other way, or by making intemperate attacks on whomever dares to criticize their competence.