"Joseph Gwinn" wrote in message
...
In article ,
"Ed Huntress" wrote:

wrote in message
...
On Apr 17, 10:55 am, "Ed Huntress" wrote:




Dammit, Dan, if I did that, you'd just say that, "no, you have to look
at
the individual states and see what percentages they're gaining or
losing,
because Eregon was talking about states." Admit it. That's exactly what
you'd do, right?

).

In other words, you don't care what the facts are, you just want to
speculate about something that fits your ideology, just like Eregon
did.
Not interested in finding out what you're talking about, you'd rather
live
on conservative economic bull**** and content yourself with that.

Too bad. By digging out the facts, you get all kinds of surprises, and
you
realize that the ideologies are all bull****. But it's not very
comforting
that way, so hang on to your speculations. You always can cook up a
good,
comforting narrative that way.

--
Ed Huntress

I really do not care. And I am not trying to cook up a comforting
narrative. You might note that I did not say anything about right to
work states being better or worse than non right to work states. Or
anything concerning the original statement or even your statements.

What I am saying is that you can not average percentages the way you
did and get meaningful data.

Yes you can. What you get is confirmation that the RTW does not improve a
state's chance of limiting firings and layoffs. That's the only question
that was at issue.

You say that the average job loss
percentage for the Northeastern states is -5.71 And that is
incorrect.

No, it's exactly correct. The average job loss percentage for a union
state
in the Northeast (which is all of them) is -5.71%. Period.

I don't have an opinion on the politics, but the math is a different
matter:

Averaging percentages is incorrect unless the populations from which the
percentages were computed are equal.

Let's take a small example: We have two jars of colored beads, A and B:

Jar A contains 1000 beads, of which 100 are black and 900 are white, so
100/(100+900)= 10% are black.

Jar B contains 100 beads, 50 are black and 50 are white, so 50% are black.

If I pour the contents of Jars A and B into empty Jar C, what percentage
of the
beads in Jar C are black?

If we average the percentages, we get (10%+50%)/2= 30%.

If we instead count beads, we get (100+50)/(1000+100)= 14.1%.


I submit that 14.1% is correct, not 30%.


Joe Gwinn

Joe, it isn't a matter of math. It's a matter of English.

What I said, and what my data said, is that RTW laws in Southern states have
resulted in a percentage of job losses, in the average state, that is higher
than the average percentage of job losses in the average Northeastern, union
state. That's what the data above show; that's what the average was all
about.

You're talking about a different question. Your question asks whether the
average of the *total number* of job losses in one group of states is higher
or lower, as a percentage of employment, in one group of states versus the
other. You want the aggregate numbers.

I recall this distinction being a big issue in 5th or 6th grade math. They
made very certain that we knew the distinction. g

The answer to your question is that the one-year manufacturing job losses in
those Southern states with RTW laws, as of March 2010,
was 6.40% of the total manufacturing employment in those states one year
ago. In the Northeastern states, it was 5.99%. In other words, it's the same
conclusion: Right-to-work laws not only don't provide a hedge against
layoffs and firings. They do the opposite.

Eregon spewed forth the standard conservative bromide, that RTW states don't
experience the percentage of layoffs in a downturn that union states do.
That's horse pucky, and every real economist has known it for decades. It's
a piece of ideological nonsense that hangs on year after year, despite the
fact that the truth about it has been readily available since I was doing
economic reports in the metalworking press.

The important facts about right-to-work laws are that they are political
maneuvers intended to lure manufacturing plants in from other places. In
exchange for placing a plant in Alabama, a manufacturer gets cheap labor and
a free hand to fire and lay people off with no serious impediments (plus tax
breaks, etc.). They can give a dirt-poor economy a big shot in the arm but
then they wind up producing volatility in the job market.

The really important part, though, is that they don't reduce overall
unemployment. For a while, unemployment goes down, and then people move in
to take up the new jobs. With the volatility that the job market now
experiences, unemployment can shoot up. Right now, against a national
average of 9.7% unemployment (March), NC is 11.1%; GA is 10.6%; MS is 11.5%;
SC is 12.2%; and FL is 12.3%.

Right-to-work laws, in other words, come around to bite a state in the ass.
They wind up with lower wages, volatile employment, and high overall
unemployment rates in a downturn. In any case, they are no better off in
almost any way than states with union shops.

--
Ed Huntress

wrote in message
...
On Apr 17, 4:19 pm, "Ed Huntress" wrote:

Yes you can. What you get is confirmation that the RTW does not improve a
state's chance of limiting firings and layoffs. That's the only question
that was at issue.

That number has no meaning.

It answers the question of whether, on the average, individual states are
experiencing higher or lower layoffs in manufacturing if they have a RTW
law...which was the question in question.


You say that the average job loss
percentage for the Northeastern states is -5.71 And that is
incorrect.

No, it's exactly correct. The average job loss percentage for a union
state
in the Northeast (which is all of them) is -5.71%. Period.


Wrong.

No, RIGHT. It's the average of the rates of loss -- which, if you read
carefully, is what I've said about four times now.


Instead of getting all bent out of shape and assuming that I am trying
to come up with something to fit my ideology, you would do better to
cool down and read what I am saying. You get three figure numbers
that do not mean a thing.

I did. And I'm getting annoyed with you not reading what was being
claimed.

There is nothing wrong with the data I dug up and presented. You're just
fishing for some way to confound it from an angle that was not claimed.



I am just pointing out that your use of data is incorrect.

Nope.

'Tell you what Explain precisely what you want to know, and I'll compile
the
data. I'm not too lazy to run some easily obtainable numbers in an Excel
spreadsheet.

Just tell us what you want to see.

--
Ed Huntress

As I said previously, I am not really interested in the whole
subject. I was just trying to educate you, but you obviously not
interested. So just forget the whole thing.

Trying to educate me, eh? Thanks, Dan. We'll have a statistics face-off
sometime, to see if you get it. d8-)

BTW, if you want the aggregate figures, it's 6.40% manufacturing job losses
over the past year for Southern RTW states, and 5.99% for Northeastern union
states.

All you had to do was ask -- or look for yourself. d8-)

--
Ed Huntress

"rangerssuck" wrote in message
...
On Apr 17, 3:17 pm, Ignoramus25624 ignoramus25...@NOSPAM.
25624.invalid wrote:
On 2010-04-17, wrote:





On Apr 17, 10:55?am, "Ed Huntress" wrote:

Dammit, Dan, if I did that, you'd just say that, "no, you have to look
at
the individual states and see what percentages they're gaining or
losing,
because Eregon was talking about states." Admit it. That's exactly
what
you'd do, right?

).

In other words, you don't care what the facts are, you just want to
speculate about something that fits your ideology, just like Eregon
did.
Not interested in finding out what you're talking about, you'd rather
live
on conservative economic bull**** and content yourself with that.

Too bad. By digging out the facts, you get all kinds of surprises, and
you
realize that the ideologies are all bull****. But it's not very
comforting
that way, so hang on to your speculations. You always can cook up a
good,
comforting narrative that way.

I really do not care. And I am not trying to cook up a comforting
narrative. You might note that I did not say anything about right to
work states being better or worse than non right to work states. Or
anything concerning the original statement or even your statements.

What I am saying is that you can not average percentages the way you
did and get meaningful data. You say that the average job loss
percentage for the Northeastern states is -5.71 And that is
incorrect.

Instead of getting all bent out of shape and assuming that I am trying
to come up with something to fit my ideology, you would do better to
cool down and read what I am saying. You get three figure numbers
that do not mean a thing.

I am just pointing out that your use of data is incorrect.

Populations of states are very easy to get, and calculating the
population average for a set of state is a 20 minute
exercise. Relistically, I would not expect the population weighted
average to be significantly different from unweighted average.

Practically speaking, the main problem of conservatives is that they
listen to conservative entertainment programs. Those programs are all
about money, keeping the audience so that advertisers can sell them
various junk. They never wanted to be truthful or even to approach
reailty. Conservatives swallow this hook, line and sinker and are a
prey to advertisers.

I have a radio in the shed to keep pests away. It has a conservative
host. What is kind of shocking is selection of advertisers. I have
heard only the following advertised:

o Impotence, Weight loss and balding cures (the ads sound
like the products are likely to be snake oil)
o IRA scams related to high fee gold IRA accounts
o Other investment scams

This really does not fit the image of conservatives that I had in my
mind. Though I no longer consider myself such, I generally would hope
that they, on average, should in some ways be above the level that the
ads imply.

But, possibly, this is a radio program only intended for a subset of
conservatives. That radio program is all about ideology, never
carefully considering anything. So I am not surprised that its
listeners are basically suckers, and get ads intended for suckers.
Non-suckers just would not listen to this. I can try to find out what
it is. I do not listen to Rush Limbaugh, but would love to know what
is advertised on his programs.

The kinds of conservatives that I used to associate with, before the
war against Iraq, would be the sort of people who read Wall Street
Journal editorials, Milton Friedman, etc. I still susbcribe to most
of his ideas, but know that in reality they are usually perverted and
used to steal money from wherever they are applied.

i

When my blood pressure is low, I listen to WABC radio out of New York.
An alarming number of advertisers are for tax dodgers "I owed the IRS
$45,000, but this company got it settled for $1500." and credit
dodgers "I owed $45,000 on my credit cards, but this company got it
settled for $1500."

For a bunch of conservatives, they have racked up some serious debt.
For a bunch of "family values" folks, they do seem to be in an awful
hurry to shirk their responsibilities.

A lot like Gunner.


The liberals I used to associate with used to buy radios with both liberal
and conservative hosts. Of course, being the liberals that they were, their
dials were always stuck on the conservative stations and all they could do
was bitch. They never dared change the channel for fear they would miss some
wisdom. Some things never change I guess. The "one station" radio in your
shed seems less complicated than the older ones were.

In article ,
"Ed Huntress" wrote:

"Joseph Gwinn" wrote in message
...
In article ,
"Ed Huntress" wrote:

wrote in message
...
On Apr 17, 10:55 am, "Ed Huntress" wrote:




Dammit, Dan, if I did that, you'd just say that, "no, you have to look
at
the individual states and see what percentages they're gaining or
losing,
because Eregon was talking about states." Admit it. That's exactly what
you'd do, right?

).

In other words, you don't care what the facts are, you just want to
speculate about something that fits your ideology, just like Eregon
did.
Not interested in finding out what you're talking about, you'd rather
live
on conservative economic bull**** and content yourself with that.

Too bad. By digging out the facts, you get all kinds of surprises, and
you
realize that the ideologies are all bull****. But it's not very
comforting
that way, so hang on to your speculations. You always can cook up a
good,
comforting narrative that way.

--
Ed Huntress

I really do not care. And I am not trying to cook up a comforting
narrative. You might note that I did not say anything about right to
work states being better or worse than non right to work states. Or
anything concerning the original statement or even your statements.

What I am saying is that you can not average percentages the way you
did and get meaningful data.

Yes you can. What you get is confirmation that the RTW does not improve a
state's chance of limiting firings and layoffs. That's the only question
that was at issue.

You say that the average job loss
percentage for the Northeastern states is -5.71 And that is
incorrect.

No, it's exactly correct. The average job loss percentage for a union
state
in the Northeast (which is all of them) is -5.71%. Period.

I don't have an opinion on the politics, but the math is a different
matter:

Averaging percentages is incorrect unless the populations from which the
percentages were computed are equal.

Let's take a small example: We have two jars of colored beads, A and B:

Jar A contains 1000 beads, of which 100 are black and 900 are white, so
100/(100+900)= 10% are black.

Jar B contains 100 beads, 50 are black and 50 are white, so 50% are black.

If I pour the contents of Jars A and B into empty Jar C, what percentage
of the
beads in Jar C are black?

If we average the percentages, we get (10%+50%)/2= 30%.

If we instead count beads, we get (100+50)/(1000+100)= 14.1%.


I submit that 14.1% is correct, not 30%.


Joe Gwinn

Joe, it isn't a matter of math. It's a matter of English.

But, one ends up with a number upon which we are asked to base some conclusion.
If this number and the resulting conclusion are to be correct, the laws of math
must govern.


What I said, and what my data said, is that RTW laws in Southern states have
resulted in a percentage of job losses, in the average state, that is higher
than the average percentage of job losses in the average Northeastern, union
state. That's what the data above show; that's what the average was all
about.

The average of percentages is not that number. The problem with "the average
state" is that states are not average, in particular varying widely in
population, and the average of percentages will give undue influence to small
states.

There are lots of ways to combine statistics, so one must always go back to
purpose to know which ways make sense and which do not, given the stated purpose.


You're talking about a different question. Your question asks whether the
average of the *total number* of job losses in one group of states is higher
or lower, as a percentage of employment, in one group of states versus the
other. You want the aggregate numbers.

Yes. Just for curiosity, what do the aggregate by-population numbers show?


I recall this distinction being a big issue in 5th or 6th grade math. They
made very certain that we knew the distinction. g

They did at that, but the lesson didn't always stick. The average of averages
has like problems. I see this kind of error all the time in engineering
analyses, of all places. Noise calculations are particularly opaque and
confusing, and the standard way to untangle such things is to go back to basic
physical variables like voltage and current, and in electro optics to count the
individual photons and electrons.


Joe Gwinn

On Apr 18, 1:56*am, "Chief Egalitarian" wrote:
"rangerssuck" wrote in message

...





On Apr 17, 3:17 pm, Ignoramus25624 ignoramus25...@NOSPAM.
25624.invalid wrote:
On 2010-04-17, wrote:

On Apr 17, 10:55?am, "Ed Huntress" wrote:

Dammit, Dan, if I did that, you'd just say that, "no, you have to look
at
the individual states and see what percentages they're gaining or
losing,
because Eregon was talking about states." Admit it. That's exactly
what
you'd do, right?

).

In other words, you don't care what the facts are, you just want to
speculate about something that fits your ideology, just like Eregon
did.
Not interested in finding out what you're talking about, you'd rather
live
on conservative economic bull**** and content yourself with that.

Too bad. By digging out the facts, you get all kinds of surprises, and
you
realize that the ideologies are all bull****. But it's not very
comforting
that way, so hang on to your speculations. You always can cook up a
good,
comforting narrative that way.

I really do not care. And I am not trying to cook up a comforting
narrative. You might note that I did not say anything about right to
work states being better or worse than non right to work states. Or
anything concerning the original statement or even your statements.

What I am saying is that you can not average percentages the way you
did and get meaningful data. You say that the average job loss
percentage for the Northeastern states is -5.71 And that is
incorrect.

Instead of getting all bent out of shape and assuming that I am trying
to come up with something to fit my ideology, you would do better to
cool down and read what I am saying. You get three figure numbers
that do not mean a thing.

I am just pointing out that your use of data is incorrect.

Populations of states are very easy to get, and calculating the
population average for a set of state is a 20 minute
exercise. Relistically, I would not expect the population weighted
average to be significantly different from unweighted average.

Practically speaking, the main problem of conservatives is that they
listen to conservative entertainment programs. Those programs are all
about money, keeping the audience so that advertisers can sell them
various junk. They never wanted to be truthful or even to approach
reailty. Conservatives swallow this hook, line and sinker and are a
prey to advertisers.

I have a radio in the shed to keep pests away. It has a conservative
host. What is kind of shocking is selection of advertisers. I have
heard only the following advertised:

o Impotence, Weight loss and balding cures (the ads sound
like the products are likely to be snake oil)
o IRA scams related to high fee gold IRA accounts
o Other investment scams

This really does not fit the image of conservatives that I had in my
mind. Though I no longer consider myself such, I generally would hope
that they, on average, should in some ways be above the level that the
ads imply.

But, possibly, this is a radio program only intended for a subset of
conservatives. That radio program is all about ideology, never
carefully considering anything. So I am not surprised that its
listeners are basically suckers, and get ads intended for suckers.
Non-suckers just would not listen to this. I can try to find out what
it is. I do not listen to Rush Limbaugh, but would love to know what
is advertised on his programs.

The kinds of conservatives that I used to associate with, before the
war against Iraq, would be the sort of people who read Wall Street
Journal editorials, Milton Friedman, etc. I still susbcribe to most
of his ideas, but know that in reality they are usually perverted and
used to steal money from wherever they are applied.

i

When my blood pressure is low, I listen to WABC radio out of New York.
An alarming number of advertisers are for tax dodgers "I owed the IRS
$45,000, but this company got it settled for $1500." and credit
dodgers "I owed $45,000 on my credit cards, but this company got it
settled for $1500."

For a bunch of conservatives, they have racked up some serious debt.
For a bunch of "family values" folks, they do seem to be in an awful
hurry to shirk their responsibilities.

A lot like Gunner.

The liberals I used to associate with used to buy radios with both liberal
and conservative hosts. Of course, being the liberals that they were, their
dials were always stuck on the conservative stations and all they could do
was bitch. They never dared change the channel for fear they would miss some
wisdom. Some things never change I guess. The "one station" radio in your
shed seems less complicated than the older ones were.

How about addressing the irresponsibility of the conservatives, as
witnessed by the advertisers on their radio programs?

"rangerssuck" wrote in message
...
On Apr 18, 1:56 am, "Chief Egalitarian" wrote:
"rangerssuck" wrote in message

...





On Apr 17, 3:17 pm, Ignoramus25624 ignoramus25...@NOSPAM.
25624.invalid wrote:
On 2010-04-17, wrote:

On Apr 17, 10:55?am, "Ed Huntress" wrote:

Dammit, Dan, if I did that, you'd just say that, "no, you have to
look
at
the individual states and see what percentages they're gaining or
losing,
because Eregon was talking about states." Admit it. That's exactly
what
you'd do, right?

).

In other words, you don't care what the facts are, you just want to
speculate about something that fits your ideology, just like Eregon
did.
Not interested in finding out what you're talking about, you'd
rather
live
on conservative economic bull**** and content yourself with that.

Too bad. By digging out the facts, you get all kinds of surprises,
and
you
realize that the ideologies are all bull****. But it's not very
comforting
that way, so hang on to your speculations. You always can cook up a
good,
comforting narrative that way.

I really do not care. And I am not trying to cook up a comforting
narrative. You might note that I did not say anything about right to
work states being better or worse than non right to work states. Or
anything concerning the original statement or even your statements.

What I am saying is that you can not average percentages the way you
did and get meaningful data. You say that the average job loss
percentage for the Northeastern states is -5.71 And that is
incorrect.

Instead of getting all bent out of shape and assuming that I am
trying
to come up with something to fit my ideology, you would do better to
cool down and read what I am saying. You get three figure numbers
that do not mean a thing.

I am just pointing out that your use of data is incorrect.

Populations of states are very easy to get, and calculating the
population average for a set of state is a 20 minute
exercise. Relistically, I would not expect the population weighted
average to be significantly different from unweighted average.

Practically speaking, the main problem of conservatives is that they
listen to conservative entertainment programs. Those programs are all
about money, keeping the audience so that advertisers can sell them
various junk. They never wanted to be truthful or even to approach
reailty. Conservatives swallow this hook, line and sinker and are a
prey to advertisers.

I have a radio in the shed to keep pests away. It has a conservative
host. What is kind of shocking is selection of advertisers. I have
heard only the following advertised:

o Impotence, Weight loss and balding cures (the ads sound
like the products are likely to be snake oil)
o IRA scams related to high fee gold IRA accounts
o Other investment scams

This really does not fit the image of conservatives that I had in my
mind. Though I no longer consider myself such, I generally would hope
that they, on average, should in some ways be above the level that the
ads imply.

But, possibly, this is a radio program only intended for a subset of
conservatives. That radio program is all about ideology, never
carefully considering anything. So I am not surprised that its
listeners are basically suckers, and get ads intended for suckers.
Non-suckers just would not listen to this. I can try to find out what
it is. I do not listen to Rush Limbaugh, but would love to know what
is advertised on his programs.

The kinds of conservatives that I used to associate with, before the
war against Iraq, would be the sort of people who read Wall Street
Journal editorials, Milton Friedman, etc. I still susbcribe to most
of his ideas, but know that in reality they are usually perverted and
used to steal money from wherever they are applied.

i

When my blood pressure is low, I listen to WABC radio out of New York.
An alarming number of advertisers are for tax dodgers "I owed the IRS
$45,000, but this company got it settled for $1500." and credit
dodgers "I owed $45,000 on my credit cards, but this company got it
settled for $1500."

For a bunch of conservatives, they have racked up some serious debt.
For a bunch of "family values" folks, they do seem to be in an awful
hurry to shirk their responsibilities.

A lot like Gunner.

The liberals I used to associate with used to buy radios with both
liberal
and conservative hosts. Of course, being the liberals that they were,
their
dials were always stuck on the conservative stations and all they could
do
was bitch. They never dared change the channel for fear they would miss
some
wisdom. Some things never change I guess. The "one station" radio in your
shed seems less complicated than the older ones were.

How about addressing the irresponsibility of the conservatives, as
witnessed by the advertisers on their radio programs?

Those ads are targeted at the liberals who are unable to turn their dials. A
fool and his money are easily parted.

"Joseph Gwinn" wrote in message
...
In article ,
"Ed Huntress" wrote:

"Joseph Gwinn" wrote in message
...
In article ,
"Ed Huntress" wrote:

wrote in message
...
On Apr 17, 10:55 am, "Ed Huntress" wrote:




Dammit, Dan, if I did that, you'd just say that, "no, you have to
look
at
the individual states and see what percentages they're gaining or
losing,
because Eregon was talking about states." Admit it. That's exactly
what
you'd do, right?

).

In other words, you don't care what the facts are, you just want to
speculate about something that fits your ideology, just like Eregon
did.
Not interested in finding out what you're talking about, you'd
rather
live
on conservative economic bull**** and content yourself with that.

Too bad. By digging out the facts, you get all kinds of surprises,
and
you
realize that the ideologies are all bull****. But it's not very
comforting
that way, so hang on to your speculations. You always can cook up a
good,
comforting narrative that way.

--
Ed Huntress

I really do not care. And I am not trying to cook up a comforting
narrative. You might note that I did not say anything about right to
work states being better or worse than non right to work states. Or
anything concerning the original statement or even your statements.

What I am saying is that you can not average percentages the way you
did and get meaningful data.

Yes you can. What you get is confirmation that the RTW does not
improve a
state's chance of limiting firings and layoffs. That's the only
question
that was at issue.

You say that the average job loss
percentage for the Northeastern states is -5.71 And that is
incorrect.

No, it's exactly correct. The average job loss percentage for a union
state
in the Northeast (which is all of them) is -5.71%. Period.

I don't have an opinion on the politics, but the math is a different
matter:

Averaging percentages is incorrect unless the populations from which
the
percentages were computed are equal.

Let's take a small example: We have two jars of colored beads, A and
B:

Jar A contains 1000 beads, of which 100 are black and 900 are white, so
100/(100+900)= 10% are black.

Jar B contains 100 beads, 50 are black and 50 are white, so 50% are
black.

If I pour the contents of Jars A and B into empty Jar C, what
percentage
of the
beads in Jar C are black?

If we average the percentages, we get (10%+50%)/2= 30%.

If we instead count beads, we get (100+50)/(1000+100)= 14.1%.


I submit that 14.1% is correct, not 30%.


Joe Gwinn

Joe, it isn't a matter of math. It's a matter of English.

But, one ends up with a number upon which we are asked to base some
conclusion.
If this number and the resulting conclusion are to be correct, the laws of
math
must govern.

It's hard for me to project what other people think when I write something
like that, but what I meant is this, and I was responding to the issue that
I felt Eregon raised: I wasn't thinking about overall numbers, because there
are so many variables, from state populations to other job opportunities in
given states. I was thinking about the question of whether states with RTW
laws tended to have greater or lesser immunity from recessions. The subject
was the recovery, and the year-over-year figures show a continued loss for
practically all (maybe all) states. And the RTW states show the same trend,
to a greater degree than the Northeastern union states.



What I said, and what my data said, is that RTW laws in Southern states
have
resulted in a percentage of job losses, in the average state, that is
higher
than the average percentage of job losses in the average Northeastern,
union
state. That's what the data above show; that's what the average was all
about.

The average of percentages is not that number. The problem with "the
average
state" is that states are not average, in particular varying widely in
population, and the average of percentages will give undue influence to
small
states.

Well, see above. If the question is what it's doing to the *states*, then
use my original numbers. If you want to know what it's done to the working
population in a region, then use the aggregate.

I'm well aware of the distinction, Joe. It's something that comes up often
in my work with US economic and other data and I'm usually careful to
clarify which it is I'm talking about. If I failed to do so here, I
apologize. But the aggregate numbers, as I've shown, indicate exactly the
same thing. That's something anyone would realize if they took even a quick
glance at the numbers in the table I linked to, and I suggested this to Dan
in an earlier message.


There are lots of ways to combine statistics, so one must always go back
to
purpose to know which ways make sense and which do not, given the stated
purpose.


You're talking about a different question. Your question asks whether the
average of the *total number* of job losses in one group of states is
higher
or lower, as a percentage of employment, in one group of states versus
the
other. You want the aggregate numbers.

Yes. Just for curiosity, what do the aggregate by-population numbers
show?

You must have missed it in my last post to you. g It's 6.40% average job
loss in the Southern RTW states taken together, versus 5.99% loss in the
Northeastern states.

That's year-over-year job losses, ending in March 2010, for manufacturing
jobs.



I recall this distinction being a big issue in 5th or 6th grade math.
They
made very certain that we knew the distinction. g

They did at that, but the lesson didn't always stick. The average of
averages
has like problems. I see this kind of error all the time in engineering
analyses, of all places. Noise calculations are particularly opaque and
confusing, and the standard way to untangle such things is to go back to
basic
physical variables like voltage and current, and in electro optics to
count the
individual photons and electrons.


Joe Gwinn

Well, then you're familiar with the problem. Given your experience, I accept
your criticism of the way I worded my statement. d8-)

--
Ed Huntress

In article ,
"Ed Huntress" wrote:

"Joseph Gwinn" wrote in message
...
In article ,
"Ed Huntress" wrote:

"Joseph Gwinn" wrote in message
...
In article ,
"Ed Huntress" wrote:

wrote in message
...
On Apr 17, 10:55 am, "Ed Huntress" wrote:




Dammit, Dan, if I did that, you'd just say that, "no, you have to
look
at
the individual states and see what percentages they're gaining or
losing,
because Eregon was talking about states." Admit it. That's exactly
what
you'd do, right?

).

In other words, you don't care what the facts are, you just want to
speculate about something that fits your ideology, just like Eregon
did.
Not interested in finding out what you're talking about, you'd
rather
live
on conservative economic bull**** and content yourself with that.

Too bad. By digging out the facts, you get all kinds of surprises,
and
you
realize that the ideologies are all bull****. But it's not very
comforting
that way, so hang on to your speculations. You always can cook up a
good,
comforting narrative that way.

--
Ed Huntress

I really do not care. And I am not trying to cook up a comforting
narrative. You might note that I did not say anything about right to
work states being better or worse than non right to work states. Or
anything concerning the original statement or even your statements.

What I am saying is that you can not average percentages the way you
did and get meaningful data.

Yes you can. What you get is confirmation that the RTW does not
improve a
state's chance of limiting firings and layoffs. That's the only
question
that was at issue.

You say that the average job loss
percentage for the Northeastern states is -5.71 And that is
incorrect.

No, it's exactly correct. The average job loss percentage for a union
state
in the Northeast (which is all of them) is -5.71%. Period.

I don't have an opinion on the politics, but the math is a different
matter:

Averaging percentages is incorrect unless the populations from which
the percentages were computed are equal.

Let's take a small example: We have two jars of colored beads, A and
B:

Jar A contains 1000 beads, of which 100 are black and 900 are white, so
100/(100+900)= 10% are black.

Jar B contains 100 beads, 50 are black and 50 are white, so 50% are
black.

If I pour the contents of Jars A and B into empty Jar C, what
percentage of the beads in Jar C are black?

If we average the percentages, we get (10%+50%)/2= 30%.

If we instead count beads, we get (100+50)/(1000+100)= 14.1%.


I submit that 14.1% is correct, not 30%.


Joe Gwinn

Joe, it isn't a matter of math. It's a matter of English.

But, one ends up with a number upon which we are asked to base some
conclusion.
If this number and the resulting conclusion are to be correct, the laws of
math must govern.

It's hard for me to project what other people think when I write something
like that,

Isn't projecting what most people will think upon reading something at the core
of the writer's tradcraft?


but what I meant is this, and I was responding to the issue that
I felt Eregon raised: I wasn't thinking about overall numbers, because there
are so many variables, from state populations to other job opportunities in
given states.

This variability is true regardless of which approach to combining statistics
one chooses.


I was thinking about the question of whether states with RTW
laws tended to have greater or lesser immunity from recessions. The subject
was the recovery, and the year-over-year figures show a continued loss for
practically all (maybe all) states. And the RTW states show the same trend,
to a greater degree than the Northeastern union states.

So this is the question to be answered. Now, I would submit that most people
thinking of such a question ultimately are wondering how the "average worker"
fares in RTW (Right-To-Work) versus non-RTW states. Now one cannot address both
average-state and average-worker in a single number, unless all states have
about the same population, which is not the case, so one must either choose or
provide both. More below.


What I said, and what my data said, is that RTW laws in Southern states have
resulted in a percentage of job losses, in the average state, that is higher
than the average percentage of job losses in the average Northeastern, union
state. That's what the data above show; that's what the average was all
about.

The average of percentages is not that number. The problem with "the average
state" is that states are not average, in particular varying widely in
population, and the average of percentages will give undue influence to
small states.

Well, see above. If the question is what it's doing to the *states*, then
use my original numbers. If you want to know what it's done to the working
population in a region, then use the aggregate.

Yes, and I think that people are more interested in how the average worker will
fare. The readers (the Governor excepted) really couldn't care less how the
state fares. More to the point, people identify with the average worker, not
with the Governor or the State. Or the legislators.


I'm well aware of the distinction, Joe. It's something that comes up often
in my work with US economic and other data and I'm usually careful to
clarify which it is I'm talking about. If I failed to do so here, I
apologize. But the aggregate numbers, as I've shown, indicate exactly the
same thing. That's something anyone would realize if they took even a quick
glance at the numbers in the table I linked to, and I suggested this to Dan
in an earlier message.

If aggregate numbers (by population) give the same numbers as by-state, then for
rhetorical reasons alone it's best to use the by-population averages.

Or, do it both ways, and show that it yields the same answer regardless.


There are lots of ways to combine statistics, so one must always go back
to purpose to know which ways make sense and which do not, given the stated
purpose.


You're talking about a different question. Your question asks whether the
average of the *total number* of job losses in one group of states is
higher or lower, as a percentage of employment, in one group of states versus
the other. You want the aggregate numbers.

Yes. Just for curiosity, what do the aggregate by-population numbers
show?

You must have missed it in my last post to you. g It's 6.40% average job
loss in the Southern RTW states taken together, versus 5.99% loss in the
Northeastern states.

That's year-over-year job losses, ending in March 2010, for manufacturing
jobs.

Well, I did slide over the political argument, and will continue to do so. But
I will suggest a mathematical approach to answering the question.

We have a collection of entities of varying populations. These entities come in
two varieties, which we will call red and cyan (not-red). The question is if
the average weight of the red population differs significantly from that of the
the cyan population.

Approach: Pool the populations of all the red entities into one large red
population. Compute the mean height and the standard deviation of height of
this pooled red population. Do likewise for the cyan entities.

We now have a mean and standard deviation for the red population, and another
mean and standard deviation for the cyan population.

Plot each mean with limits at plus and minus one standard deviation, with red
and cyan on parallel lines. If the difference in means is much less than the
overlap region of the +/- one standard deviation extents, it will be hard to
conclude that red versus cyan much matters - the difference being "lost in the
noise". The fancy way to express this is to say that we "failed to reject the
null hypothesis", the null hypothesis being simply that there is no detectable
effect. So, there are three kinds of answer possible: Red better, no
significant difference, cyan better.

There is a more precise and fancy way to calculate this, yielding confidence
intervals, which are beloved of statisticians, but the above visual method is
pretty effective, and far easier to understand and explain. What also works
well is to plot the two gaussian distributions (bell curves) on the same graph,
and look to see how well separated the two curves are.

Now weight is not employment, but the parallel should be clear.


I recall this distinction being a big issue in 5th or 6th grade math.
They made very certain that we knew the distinction. g

They did at that, but the lesson didn't always stick. The average of averages
has like problems. I see this kind of error all the time in engineering
analyses, of all places. Noise calculations are particularly opaque and
confusing, and the standard way to untangle such things is to go back to
basic physical variables like voltage and current, and in electro optics to
count the individual photons and electrons.


Joe Gwinn

Well, then you're familiar with the problem. Given your experience, I accept
your criticism of the way I worded my statement. d8-)

OK,

Joe Gwinn