|
|
| Author |
Message |
candori
science forum beginner
Joined: 04 Jun 2006
Posts: 2
|
 Posted: Sun Jun 04, 2006 2:13 am Post subject:
Question from a non-statistician re pawnshops
|
|
|
Is there a way to analyze the following?
1. Pawnbrokers send detailed records by computer to the police of all
items that are taken in pawn. These records are very detailed and
contain many numbers (serial no. model no.) and letters (first name,
last name, make of article, type of article, etc). I believe probablity
dictates that the pawnbroker will inevitably make mistakes in typing.
Is there some way of calculating what the normal range (%) of mistakes
might be?
2. An honest pawnbroker will still, inadvertently, accept a certain
number of stolen items (given they're not psychic). I don't see how one
could measure probability on this, since it would depend so much on
instinct and experience, but I wonder nonetheless.
Thank you in advance. |
|
| Back to top |
|
 |
Pavel314
science forum addict
Joined: 29 Apr 2005
Posts: 78
|
| Posted: Sun Jun 04, 2006 12:17 pm Post subject:
Re: Question from a non-statistician re pawnshops
|
|
|
"candori" <can_dori@yahoo.com> wrote in message
news:1149387202.727685.128840@i39g2000cwa.googlegroups.com...
| Quote: | Is there a way to analyze the following?
1. Pawnbrokers send detailed records by computer to the police of all
items that are taken in pawn. These records are very detailed and
contain many numbers (serial no. model no.) and letters (first name,
last name, make of article, type of article, etc). I believe probablity
dictates that the pawnbroker will inevitably make mistakes in typing.
Is there some way of calculating what the normal range (%) of mistakes
might be?
|
I googled "typing errors as percent of keystrokes" and was led to the
folowing site:
http://panko.cba.hawaii.edu/HumanErr/Basic.htm
Reviewing their list of error rates, this one seemed to be the most
relevent:
Melchers & Harrington [1982]
Students performing calculator tasks and table lookup tasks. Per
multipart calculation. Per table lookup. Etc.
1%-2%
| Quote: | 2. An honest pawnbroker will still, inadvertently, accept a certain
number of stolen items (given they're not psychic). I don't see how one
could measure probability on this, since it would depend so much on
instinct and experience, but I wonder nonetheless.
|
Less than one-half of one percent, according to the National Pawnbrokers
Association. I googled "pawnbroker percentage of stolen goods"; the first
reference led me to:
http://www.nationalpawnbrokers.org/faqs.htm#5
However, the Alberta, Canada, website referenced below puts it at about 1%:
http://www.solgen.gov.ab.ca/policing/downloads/pawnbrokers_registration_act/recovery_of_stolen_property.htm
You could search a number of other sites from the Google search and take an
average but it looks like about 1% is a good estimate.
I think what you're looking for is the chance that a stolen item is not
identified as such because the pawn shop mistypes the serial number.
Assuming that all stolen goods have one identifying number and taking the
maximums in the above references, we have a 2% error rate on 1% of the goods
pawned resulting in a stolen good being unidentifiable. Hence, 2% * 1% =
..02 * .01 = .0002 = 0.02%. So two out of every ten thousand (or one out of
5,000) stolen items which are pawned will not be recoverable by tracing the
serial number because of typing errors.
If you allow more than one identifying code for some items, such as a
combination of manufacturer, size, model number and serial number, you'd
need to factor in the relative percentage of items with 1, 2, 3 or more id
numbers to get your estimated error rate. Or maybe the fact that the error
rate is for "multipart" calculations covers this. I guess it's a matter of
interpretation.
Paul |
|
| Back to top |
|
|
Duncan Smith
science forum beginner
Joined: 29 Apr 2005
Posts: 21
|
| Posted: Sun Jun 04, 2006 2:00 pm Post subject:
Re: Question from a non-statistician re pawnshops
|
|
|
Pavel314 wrote:
| Quote: | "candori" <can_dori@yahoo.com> wrote in message
news:1149387202.727685.128840@i39g2000cwa.googlegroups.com...
Is there a way to analyze the following?
1. Pawnbrokers send detailed records by computer to the police of all
items that are taken in pawn. These records are very detailed and
contain many numbers (serial no. model no.) and letters (first name,
last name, make of article, type of article, etc). I believe probablity
dictates that the pawnbroker will inevitably make mistakes in typing.
Is there some way of calculating what the normal range (%) of mistakes
might be?
I googled "typing errors as percent of keystrokes" and was led to the
folowing site:
http://panko.cba.hawaii.edu/HumanErr/Basic.htm
Reviewing their list of error rates, this one seemed to be the most
relevent:
Melchers & Harrington [1982]
Students performing calculator tasks and table lookup tasks. Per
multipart calculation. Per table lookup. Etc.
1%-2%
2. An honest pawnbroker will still, inadvertently, accept a certain
number of stolen items (given they're not psychic). I don't see how one
could measure probability on this, since it would depend so much on
instinct and experience, but I wonder nonetheless.
Less than one-half of one percent, according to the National Pawnbrokers
Association. I googled "pawnbroker percentage of stolen goods"; the first
reference led me to:
http://www.nationalpawnbrokers.org/faqs.htm#5
However, the Alberta, Canada, website referenced below puts it at about 1%:
http://www.solgen.gov.ab.ca/policing/downloads/pawnbrokers_registration_act/recovery_of_stolen_property.htm
You could search a number of other sites from the Google search and take an
average but it looks like about 1% is a good estimate.
I think what you're looking for is the chance that a stolen item is not
identified as such because the pawn shop mistypes the serial number.
Assuming that all stolen goods have one identifying number and taking the
maximums in the above references, we have a 2% error rate on 1% of the goods
pawned resulting in a stolen good being unidentifiable. Hence, 2% * 1% =
.02 * .01 = .0002 = 0.02%. So two out of every ten thousand (or one out of
5,000) stolen items which are pawned will not be recoverable by tracing the
serial number because of typing errors.
|
No, that would be 2% of stolen items which are pawned. Two out of every
ten thousand pawned items would be stolen and not be recoverable.
Duncan |
|
| Back to top |
|
|
candori
science forum beginner
Joined: 04 Jun 2006
Posts: 2
|
| Posted: Mon Jun 05, 2006 6:24 am Post subject:
Re: Question from a non-statistician re pawnshops
|
|
|
Pavel314 wrote:
| Quote: | "candori" <can_dori@yahoo.com> wrote in message
news:1149387202.727685.128840@i39g2000cwa.googlegroups.com...
2. An honest pawnbroker will still, inadvertently, accept a certain
number of stolen items (given they're not psychic). I don't see how one
could measure probability on this, since it would depend so much on
instinct and experience, but I wonder nonetheless.
I think what you're looking for is the chance that a stolen item is not
identified as such because the pawn shop mistypes the serial number.
|
Hi Paul,
First, thank you so much for your detailed reply. I will check out
those websites.
Second, the two questions I asked were not actually related. I know
that a certain number of stolen items will not be identified by the
police because the owners never reported the thefts to the police dept,
or the items don't have serial numbers and can't be identified. Police
will often look through a pawnshop's contents and records and in that
way suspicious items are seized and checked, and typos are identfied
and the owners sometimes brought to court. Of course, many stolen items
are disposed of through flea markets, corner stores, bars, hotels,
other types of stores, on the street, ebay etc.
However, some stolen items do go through pawnshops, and that's where my
second question comes in. Police will take items they are suspicious of
as well as those they have reports for. So, I figure there might be a
range in the number of stolen items that, let's say, an experienced
pawnbroker might inadvertently accept (the seller looked OK, they've
pawn items before that were OK, they have ID and the item is not one
that's usually stolen, etc). If the police (in a time period) find a
significant amount over that normal range, then one might question the
business practices of that particular shop.
I hope that explanation makes my questions clearer.
Thanks |
|
| Back to top |
|
|
Google
|
|
| Back to top |
|
|
|
|
The time now is Thu Sep 20, 2018 8:49 am | All times are GMT
|
|
Copyright © 2004-2005 DeniX Solutions SRL
|
|
Other DeniX Solutions sites:
Electronics forum |
Medicine forum |
Unix/Linux blog |
Unix/Linux documentation |
Unix/Linux forums |
send newsletters
|
| |
|
Powered by phpBB © 2001, 2005 phpBB Group
|
|
FAQ
Search
Memberlist
Usergroups
Profile
Log in to check your private messages
Log in
Forum index
»
Science and Technology
»
Math
»
Probability
