Go Back   Honda-Tech > Community Forums > General Discussion and Debate
Sign in using an external account
Register Forgot Password?
Search


Welcome to Honda-Tech!
Welcome to Honda-Tech.com.

You are currently viewing our forums as a guest, which gives you limited access to view most discussions and access our other features. By joining our community you will have access to post topics, communicate privately with other members (PM), respond to polls, upload content and access many other special features. Registration is fast, simple and absolutely free so please, join the Honda-Tech community today!


Reply
 
 
 
submit to reddit
 
Thread Tools Search this Thread
Old 02-09-2005, 05:01 AM   #1
fredtoast
Garage is empty, add now
 
Join Date: Nov 2002
Location: Traveling, tn
Posts: 513
iTrader Rating: (0)
Default Example of deceiving statistical argument.

A while back I read a long story in the paper about how people earning minimum wage could not afford the average rent on an apartment.

http://www.detnews.com/2003/re...0.htm

Does anyone see the huge mistake in this argument? The AVERAGE rent for an apartment is going to be greater than about half of the units on the market. MINIMUM wage should allow you to afford MINIMUM rent. AVERAGE wages should allow you to afford AVERAGE rent.

Anyone who has dealt with me here before can probably guess that I am in favor of raising the minimum wage. However when I see blatant deception like this I will call it out no matter who is using it.
__________________
Not even the noxious fumes from the flatulence of aroused simians could penetrate the Hellman's-sealed Samsonite luggage barricade. Just ask Joe Louis's corpse.
fredtoast is offline   Reply With Quote
Old 02-09-2005, 05:16 AM   #2
Filthy Focker!
Garage is empty, add now
 
Join Date: Feb 2004
Location: Gatineau, Quebec
Posts: 1,024
iTrader Rating: (0)
Default Re: Example of deceiving statistical argument. (fredtoast)

The average has nothing to do with the midpoint.

fs
Filthy Focker! is offline   Reply With Quote
Old 02-09-2005, 05:17 AM   #3
zer0gauge
Garage is empty, add now
 
zer0gauge's Avatar
 
Join Date: Oct 2003
Location: San Antonio, TX, 78232
Posts: 25
iTrader Rating: (0)
Default Re: Example of deceiving statistical argument. (fredtoast)

Quote:
Originally Posted by fredtoast
A while back I read a long story in the paper about how people earning minimum wage could not afford the average rent on an apartment.

http://www.detnews.com/2003/re...0.htm

Does anyone see the huge mistake in this argument? The AVERAGE rent for an apartment is going to be greater than about half of the units on the market. MINIMUM wage should allow you to afford MINIMUM rent. AVERAGE wages should allow you to afford AVERAGE rent.

Anyone who has dealt with me here before can probably guess that I am in favor of raising the minimum wage. However when I see blatant deception like this I will call it out no matter who is using it.
your argument would hold water if they were referring to the MEDIAN rent for the market, not the AVERAGE. Median is the exact middle point, where average takes the combined sum and divides it by the number of units involved.
For example:
3 ppl live on 1 road. Respectively they 25k, 37k, and 93k a year. The median income for the street is 37k. The average would be 51.6k. See the difference?
__________________
"Every normal man must be tempted at times to spit upon his hands, hoist the black flag, and begin slitting throats."
-Mencken

[quote=nismoserturbo]
I don't know why I even bother trying to post on H-T, its impossible not to get flamed[/quote]
zer0gauge is offline   Reply With Quote
Old 02-09-2005, 05:19 AM   #4
AP-Signworks1
Junior Member
Garage is empty, add now
 
Join Date: Jan 2001
Location: Atlanta, GA, US
Posts: 192
iTrader Rating: (0)
Default Re: Example of deceiving statistical argument. (filthy_shovel)

No, it isn't the median. But, he is right on the point. Why should someone making minimum wage be shopping for "average" priced apartments. ie, apartments in my area go for anywhere from $400-$1200 for a 1 bedroom depending on section of town. The average is around $750. Someone making minimum wage is obviously going to be living in a $400 unit, but if we used that statistic here, it would make it seem that people making minimum wage are homeless and have nowhere to live when that isn't the truth.

Quote:
Originally Posted by filthy_shovel
The average has nothing to do with the midpoint.

fs
AP-Signworks1 is offline   Reply With Quote
Old 02-09-2005, 05:24 AM   #5
ricodemus
Honda-Tech Member
Garage is empty, add now
 
ricodemus's Avatar
 
Join Date: Apr 2002
Location: The First State
Posts: 5,244
iTrader Rating: (0)
Default Re: Example of deceiving statistical argument. (zer0gauge)

Quote:
Originally Posted by zer0gauge

your argument would hold water if they were referring to the MEDIAN rent for the market, not the AVERAGE. Median is the exact middle point, where average takes the combined sum and divides it by the number of units involved.
For example:
3 ppl live on 1 road. Respectively they 25k, 37k, and 93k a year. The median income for the street is 37k. The average would be 51.6k. See the difference?
Bad example to discredit because that average (51.6k) certainly isn't affordable to the person on the minimum on that street. He's saying that if you use average you get an inflated number like the 51.6k and for sure that person in the minimum category couldn't afford that average price.
__________________
90 Civic Si
93 Integra LS-S - wrecked and stripped
01 Maxima 20th Ann. Ed. - traded in
05 Civic EX
ricodemus is offline   Reply With Quote
Old 02-09-2005, 05:27 AM   #6
zer0gauge
Garage is empty, add now
 
zer0gauge's Avatar
 
Join Date: Oct 2003
Location: San Antonio, TX, 78232
Posts: 25
iTrader Rating: (0)
Default Re: Example of deceiving statistical argument. (ricodemus)

Quote:
Originally Posted by ricodemus

Bad example to discredit because that average (51.6k) certainly isn't affordable to the person on the minimum on that street. He's saying that if you use average you get an inflated number like the 51.6k and for sure that person in the minimum category couldn't afford that average price.
the example was only to show the difference between median and average, therefore the example was perfectly apt, especially considering his statement of:
<TABLE WIDTH="90%" CELLSPACING=0 CELLPADDING=0 ALIGN=CENTER><TR><TD>Quote &raquo;</TD></TR><TR><TD CLASS="quote">
The AVERAGE rent for an apartment is going to be greater than about half of the units on the market.</TD></TR></TABLE>.
__________________
"Every normal man must be tempted at times to spit upon his hands, hoist the black flag, and begin slitting throats."
-Mencken

[quote=nismoserturbo]
I don't know why I even bother trying to post on H-T, its impossible not to get flamed[/quote]
zer0gauge is offline   Reply With Quote
Old 02-09-2005, 05:35 AM   #7
ricodemus
Honda-Tech Member
Garage is empty, add now
 
ricodemus's Avatar
 
Join Date: Apr 2002
Location: The First State
Posts: 5,244
iTrader Rating: (0)
Default Re: Example of deceiving statistical argument. (zer0gauge)

Quote:
Originally Posted by zer0gauge
the example was only to show the difference between median and average, therefore the example was perfectly apt, especially considering his statement of:
.
well you said that his argument didn't hold water when your example proves that it does. He's pointing out the flaw in the report by not using the median and just the logic that minimum wage probably wouldnt be able to afford the avg apartment, even if the median and the avg were the same.
__________________
90 Civic Si
93 Integra LS-S - wrecked and stripped
01 Maxima 20th Ann. Ed. - traded in
05 Civic EX
ricodemus is offline   Reply With Quote
Old 02-09-2005, 06:43 AM   #8
fredtoast
Garage is empty, add now
 
Join Date: Nov 2002
Location: Traveling, tn
Posts: 513
iTrader Rating: (0)
Default Re: Example of deceiving statistical argument. (filthy_shovel)

Quote:
Originally Posted by filthy_shovel
The average has nothing to do with the midpoint.
Actually with a large sample like "all apartments" the average will usually be very close to the median. That is why I put the word "about" in my statement. Either way my point is still valid.
__________________
Not even the noxious fumes from the flatulence of aroused simians could penetrate the Hellman's-sealed Samsonite luggage barricade. Just ask Joe Louis's corpse.
fredtoast is offline   Reply With Quote
Old 02-09-2005, 06:58 AM   #9
Splat
Collector-o-Guns and Drinker
Garage is empty, add now
 
Splat's Avatar
 
Join Date: Jul 2000
Location: Armed
Posts: 9,239
iTrader Rating: (0)
Default

"Liars figure, and figures lie."

My dad is fond of that expression, having spent most of his professional life as an auditor.

The more current equivalent we use around here is "Torture the data long enough, and it will tell you anything you want to know."
__________________
Pat http://images.honda-tech.com/set1/smile/emdgust.gif
'99 S70 T5M (1 of 197), '02 S40 1.9T, '83 245DL Wagon, 244GL Sedan (busted), '76 IH Scout II

I believe she's going down,
I believe we're gonna die die die!
Splat is offline   Reply With Quote
Old 02-09-2005, 07:04 AM   #10
purplegsr
Member
Garage is empty, add now
 
purplegsr's Avatar
 
Join Date: Oct 2003
Location: WeHo, CA, USA
Posts: 1,912
iTrader Rating: (0)
Default Re: Example of deceiving statistical argument. (fredtoast)

I am a health policy researcher/statistical programmer, so I know a bit about the subject. There are too many methods questions that can't be answered by this article, so I will try to find the original publication and let you know what I think.

That being said, There are a few issues:

Median and Mean can both be referred to as the 'Average'. Usually, when I do an average, I report the Mean with a 95% confidence interval (which is based on the standard error and degrees of freedom). Based on this short blurb, I don't know whether they used an average mean or average median for their reported 'average'.

Second, it is very very very difficult to get media coverage that understands what the hell you are saying most times. I have been interviewed for tv, print, and radio news and they always want me to boil it down to the most salient, easiest to understand points. If you talk about the limitations of your study, the methods used, the % error, etc they will completely ignore it and try to get you to say something like 'Rent costs $791 and poor people make $237 so they can't afford it.' Unless the reporter really understands the issue, they often try to simplify what you say. So, until I see the actual publication, how they dealt with outliers, sample size, and their data sources for income and home prices, I would assume that the study's authors aren't trying to decieve, but that they probably did a bad job of reporting their findings.
purplegsr is offline   Reply With Quote
Old 02-09-2005, 07:16 AM   #11
purplegsr
Member
Garage is empty, add now
 
purplegsr's Avatar
 
Join Date: Oct 2003
Location: WeHo, CA, USA
Posts: 1,912
iTrader Rating: (0)
Default Re: Example of deceiving statistical argument. (purplegsr (Exospeed))

Here is the actual report, although I think this one is newer.
http://www.nlihc.org/pubs/uaw04/UpAgainstaWall.pdf

NOTE: The average reported in the above article is actually a MEDIAN value for rent, they do not calculate a mean in the report (at first glance).

They use a Census Dept. survey, the American Community Survey (ACS), to calculate the median pricing for rent and median income in order to make their assessments. I have heard good things about the survey, and they supposedly have enough sample to make local estimates. The one problem I do have with the study is that there is no multivariate analysis, it is all descriptive and there is no measure of statistical significance. This is usually what happens when a trade organization uses data like this - although their findings are probably relatively accurate, they don't test it, and then they leave those findings in question. As a researcher in an academic setting, we always do more analysis (chi-square, t-test, and regression) to make sure we can make certain statements.

All in all, I think the report is ok, but I think the article is a little bit misleading about how they reached the findings.
purplegsr is offline   Reply With Quote
Old 02-09-2005, 07:35 AM   #12
fredtoast
Garage is empty, add now
 
Join Date: Nov 2002
Location: Traveling, tn
Posts: 513
iTrader Rating: (0)
Default Re: Example of deceiving statistical argument. (purplegsr (Exospeed))

Quote:
Originally Posted by purplegsr (Exospeed)
.All in all, I think the report is ok, but I think the article is a little bit misleading about how they reached the findings.
I'm not saying that there is ANYTHING wrong with the numbers. I'm just saying that the CONCLUSION is incorrect. A person making minimum wage should not expect to afford an "average" apartment. It doesn't matter how you define "average". A person making minimum wage should be able to afford minimum rent.

I was afraid that this thread would dissolve into a discussion about minimum wage. I never imagined that we would get into an in depth analysis of statistical methodology.
__________________
Not even the noxious fumes from the flatulence of aroused simians could penetrate the Hellman's-sealed Samsonite luggage barricade. Just ask Joe Louis's corpse.
fredtoast is offline   Reply With Quote
Old 02-09-2005, 08:02 AM   #13
purplegsr
Member
Garage is empty, add now
 
purplegsr's Avatar
 
Join Date: Oct 2003
Location: WeHo, CA, USA
Posts: 1,912
iTrader Rating: (0)
Default Re: Example of deceiving statistical argument. (fredtoast)

Yeah, I agree, I remember about a month ago I was discussing So. Cal home prices with my girlfriend and another friend - there was another report that came out in CA about median home prices and they had a quote that '80% of people in CA could not afford a home'.

However, they actually meant that 80% of Californians couldn't afford a home if it was priced at the median, which is something like $470,000 now. I argued the same point that you are, it's an unfair comparison because it doesn't mean 80% of people are renters, it means that a fair majority of people buy cheaper houses that are not in urban areas. As someone who does this for a living, I always question 'research' findings that overstate their claims like that.

I think academic researchers are doing better at stating their biases, limitations of their numbers, and not overstating their findings just for a headline. But there are still quite a few trade associations and advocacy groups that tend to find a number and illogically use it for their own goals, or try to. You should read Deborah Stone's 'Policy Paradox', it's a very good book about how research is never objective, and that transparency of methods and dialogue is the only true way to further your research and actually learn anything.
purplegsr is offline   Reply With Quote
 
 
 
submit to reddit
Reply

Tags
argument, arguments, deceiving, examples, information, numbers, sample, statistical

Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Forum Jump



All times are GMT -8. The time now is 12:16 PM.



2008 Copyright, InternetBrands Inc.
Powered by vBulletin® Version 3.7.4
Copyright ©2000 - 2014, Jelsoft Enterprises Ltd.

Honda and the Honda marquee are registered trademarks of the American Honda Motor Company, Inc. Neither American Honda Motor Company nor its subsidiaries or affiliates shall bear any responsibility for Honda-Tech.com content, comments, or advertising. Honda-Tech.com is not affiliated with American Honda Motor Company in any way. American Honda Motor Company does not sponsor, support, or endorse Honda-Tech.com in any way. Copyright/trademark/sales mark infringements are not intended or implied.
Emails & Contact Details