Linear Trendline Vs Common Sense

[Douglas Eckert]

I need to know why the amount of workload this year is down 9% so far, but
the linear trend line is up. True, work has increased in recent months, but
it is down for the fiscal year-to-date. The linear trend accounts for the
recent increases, but not for the fact that the total amount of workload this
year has declined when compared to the same time period last year. ‘Any
thoughts?

 

[James Silverton]

Douglas wrote on Thu, 7 Jun 2007 07:10:00 -0700:

DE> I need to know why the amount of workload this year is down
DE> 9% so far, but the linear trend line is up. True, work has
DE> increased in recent months, but it is down for the fiscal
DE> year-to-date. The linear trend accounts for the recent
DE> increases, but not for the fact that the total amount of
DE> workload this year has declined when compared to the same
DE> time period last year. ‘Any thoughts?

The trendline is a mathematical determination from the data and
essentially there would be roughly as many values below as above
the line, I guess. Perhaps a moving average might be more
helpful to you.

James Silverton
Potomac, Maryland

E-mail, with obvious alterations:
not.jim.silverton.at.verizon.not

 

[Jakob Austgulen]

If you now have so many orders that there is overtime to do them, is it any
help that tings was slow i january?
If orders again are turning down in january, then I would say that linear is
not the right thinking. Then You go to other matematical use.

 

[Harimau]

Linear trendlines shouldn't be used for this kind of analysis. If any
econometric or statistical analysis is going to be done, it needs to be done
in a proper (and extremely time consuming) way.

Just use your common sense. Your business intuition is probably going to be
better than what any statistician tells you anyway.

As for an explanation why the trend line is "up"... It could be because that
for the last few years, there's been an increasing trend, which is why the
trend line has a positive gradient. It's just even more reason not to use it
for this kind of analysis. The trendline isn't very aptly named... It's just
basically the best fit line to the data, which rarely (if ever) captures any
sort of causal relationship.

My guess (emphasize "GUESS") could be that you need to look into how
efficient your work is. There could be a productivity/efficiency improvement
from the last measurement date that explains the lower workload even though
work has increased.

 

[Douglas Eckert]

Thank you very much, Jakob Austgulen!

If you now have so many orders that there is overtime to do them, is it any
help that tings was slow i january?
If orders again are turning down in january, then I would say that linear is
not the right thinking. Then You go to other matematical use.

 

[Douglas Eckert]

Thank you very much, James Silverton!

 

[Douglas Eckert]

Thank you very much, Harimau!

Linear trendlines shouldn't be used for this kind of analysis. If any
econometric or statistical analysis is going to be done, it needs to be done
in a proper (and extremely time consuming) way.

Just use your common sense. Your business intuition is probably going to be
better than what any statistician tells you anyway.

As for an explanation why the trend line is "up"... It could be because that
for the last few years, there's been an increasing trend, which is why the
trend line has a positive gradient. It's just even more reason not to use it
for this kind of analysis. The trendline isn't very aptly named... It's just
basically the best fit line to the data, which rarely (if ever) captures any
sort of causal relationship.

My guess (emphasize "GUESS") could be that you need to look into how
efficient your work is. There could be a productivity/efficiency improvement
from the last measurement date that explains the lower workload even though
work has increased.