Enjoy the benefits of a premium membership!
Upgrade Account
-
TalkBass has been independent since 1998. Add your voice.
Create a free account to reply to discussions, view embedded media, and browse with fewer display ads.
Join free Log in
Want zero display ads or expanded classifieds tools? Compare plans.
Get off my LAWN! "Grumpy Old Farts"
- Thread starter B-string
- Start date
- Status
A Scottish drinking song begs to be balanced with an Irish words of wisdom folk song. So you see.
Irish songs all contain wisdom, for those who can see it.
I'll stop there.
I'll stop there.
- Oct 22, 2015
- 21,164
- 256,806
- Disclosures
- I am an affiliate for Ari Cap.
There is a fine lady from Uisneach....
I'll stop there. Er. well, perhaps I should go on a wee bit. Saturday morning I open Friday's post on my way to the Growers' Market. Enclosed, a card bearing a representation of one often associated with Uisneach, with whom I share a common resonance. Purchased a frame Tuesday. It's too small because my measurement was guesswork. Return the frame today for a larger frame, and will place the framed piece in my practice area today.
@the harp unstrung Inspiration move me brightly
Dunno if this was mentioned. Is the body count normalized per movie? (I'm guessing Clint's a cowboy a lot more than he's Dirty Harry. But there may be automated weapons in the DH movies - I forget.)
- Oct 22, 2015
- 21,164
- 256,806
- Disclosures
- I am an affiliate for Ari Cap.
Limerick is just to the south.
A useful reference and the answer to all our questions.
He shot the most in the war movies. Just saying.
The statistics should take into account gross and net revenue, then correlate kill count to obtain a feel for what the public wants to see weighted by production cost. Of course, what was popular then may not be now. Maybe viewing over each year needs to be factored in.
This gets complicated really fast.
We usually have raw numbers (not always, alas). What we want out of the raw numbers is added in later. Modern statistics modeling packages are stunningly useful in this regard. The trick, I think, is how to communicate what those stats mean (both to those asking and answering the questions).
And what I mean by this is stats packages can be applied willy-nilly and blindly to data sets by those who neither know the data in question nor what stats mean. That's a huge danger. Another other danger is not being able to say clearly what the correlations mean to folks not skilled in the art/craft/science.
- Oct 22, 2015
- 21,164
- 256,806
- Disclosures
- I am an affiliate for Ari Cap.
It sounds suspiciously like my former profession, to which I plead guilty. I am sorry; this sort of stuff just comes out on occasion.
Some are more reputable than others...
With proper polling you can show whatever you want. Statistics can be like that, especially when outliers are strategically removed from the data.
- Oct 22, 2015
- 21,164
- 256,806
- Disclosures
- I am an affiliate for Ari Cap.
Quite true, tho' apply statistics well away from things human (e.g., politics) and have a large enough sample base, numbers can describe a boatload of things very, very well. (And, scarily enough, you can describe human behavior pretty well in a number of instances.)
The old saw: vis a vis humans statistics are descriptive, not prescriptive.
That's actually improper polling. A proper research study should, ideally, be designed with as little bias as possible, though it is usually not possible to remove it all. Since the general public doesn't know what that looks like, and since most people funding studies are looking for specific answers, it's rare that you see the effort being made outside of strict scientific work, and even there you often see pretty sloppy work being done.
Here is an industrial example.
A product is produced, samples are taken and analyzed. Statistics indicate that the product does not conform to specifications. They didn’t get the required result. Scrapping the batch would be too expensive. They resample to obtain a larger population. Still doesn’t conform. There are meetings. The data is massaged, some outliers are dropped, this can be a subjective process, how many standard deviations out will you accept.
Specifications have to meet the capabilities of the plant. With a new product, sometimes the target specs have to be relaxed. There’s a difference between desired specs and what can be produced. Now the product conforms. The paperwork is signed off on, the batch passes quality control and is released.
Thankfully, the process, controls, and methods are more refined today. In the past, they got away with things that they shouldn’t have. Sometimes they got caught and a recall was in order, sometimes they didn’t.
There have been recent examples in the auto industry that indicate that better regulation isn’t always a bad thing.
When a pesky pollster calls, many people lie to purposely skew the results.
Scientific studies should be more accurate. Unfortunately, fixing data to get more funding or material published is not unheard of in this competitive world. It doesn’t end well if they eventually get caught.
I
What you have described is human intervention in the process, and I agree, it happens way too much. Tho' it depends what you are making - if it's, say, a vital piece in something related to safety, what you've described is in my opinion criminal. If you are making, say, transistors/circuits you can 'bin out' the more poorly functioning chips and sell them as 'OK performing circuits' rather than 'Top notch, super duper circuits'. And that is stats working pretty well.
I also agree with you that better regulation is often called for.
What you have described is human intervention in the process, and I agree, it happens way too much. Tho' it depends what you are making - if it's, say, a vital piece in something related to safety, what you've described is in my opinion criminal. If you are making, say, transistors/circuits you can 'bin out' the more poorly functioning chips and sell them as 'OK performing circuits' rather than 'Top notch, super duper circuits'. And that is stats working pretty well.
I also agree with you that better regulation is often called for.
- Status
