Feds say amusement injuries are up and down

Exactly, carnivals and fairs are accidents waiting to happen. Markey should have his eye on this, not amusement parks.

I agree 100 percent

The sampling method used is invalid.  They are sampling hospital emergency rooms not amusement parks.  This is statistically invalid for the purposes of the study.  The error is calculated by the CPSC as +/-50%, and that is based on the false idea that the sampling is valid. 

The quality of the analysis is so poor that one of the 2 deaths listed occurred on a water slide even though the report specifically states that water slides are not included.  If these guys can't get it right on the deaths, how good do you think that the rest of the analysis is?

The total number of injuries actually recorded as verified fixed amnusement park ride injuries was less than 100.  Nearly half of these injuries came from 1 of the 100 hospitals in the survey.  From this data the report concludes that 6594 injuries occurred.

Ann Brown's statement that the number of deaths has increased is an outright falsehood.  Even her own agency's study showed no significant increase in the number of deaths.

For comic relief I guess, the report features an analysis to show the most amusemnt ride injuries occur in summer.  Did we really need a government study to tell us this?

Say, Jim, are you interested in writing an editorial on this?

-----------------
Jeff - Webmaster/Admin - CoasterBuzz.com
"From the global village... in the age of communication!"
Watch the grass grow!

I don't know about an editorial, but I do think that post should definitely be forwarded to whom it may concern...when the casual enthusiasts know more about the facts than those who study and crunch the figures do, something is definitely amiss.  Way to be, Jim!

Which leaves us to the interesting counterpoint of, if these figures are so WRONG, how are we to trust any other figures that Markey and his advocates use in defense of strict ride regulations?  I think the biggest problem with amusement park safety is the fact that these numbers can't be kept straight.  Parks report their accidents without exception, so the responsibility is out of their hands when it comes to the number crunching...

This simple and downright stupid error (for lack of a more viable description), in my opinion, is much graver and more in need of change than the current procedure governing amusement ride safety.  How hypocritical...how ludicrous...

-----------------
The Luv Monkey has spoken...
NITRO kicks it into high gear...only at Six Flags Great Adventure!
Track Record: 28 and counting...

Just to add to this ludicrous report.

What these people who make claim's like "increase in number of people sent to the emergency room" fail to state any increase or decrease in attendance and its effect. Regardless of the insane amount of error in this study, if injuries are up lets say 7% and attendance is up 10%, the percentage of people injured per visit actually went down.
*** This post was edited by Joe E. on 8/28/2001. ***

reminds me of the old Mark Twain quote "There are three kinds of lies: lies, damn lies, and statistics."

Hail Jim Fisher!- great logic ! Thank you!

The major problem here is that there is no accurate accounting of accidents. Contrary to your beliefs Monkey, the parks DO NOT have to report accidents.. look at Florida for example.. the big parks there are exempt ..they answer to no one(via special state legislation), and therefore do not have to even report a death on premisis..

I would like to see accurate data collected. I wish the parks would voluntarily do this. Just for a show of good faith.

Furthermore, the CPSC has NOTHING to do with Markey. You all are too into "conspiracy theories"!

*edited part*
I have actually read part of this report (not through it all yet, it is lengthy) and the news article on this topic has some data incorrect... this article is confusing and misleading. I really recommend you READ the report for YOURSELVES instead of relying on the media that you all (most) have repeatedly bashed, trashed and assasinated this summer.

You can access the report at this link:
http://www.cpsc.gov/library/amus2001.pdf

Please ....PLEASE read it for yourselves! And remember what Jim Fisher states on how they collected this data: FROM HOSPITALS not PARKS or Carnivals..this was a sampling taken from EMERGENCY ROOMS across the US...

Not the best way to collect data! Even so it shows the increase of injuries in fixed site parks! *** This post was edited by BB on 8/28/2001. ***

Yes, I read the report, and the article isn't misleading it all. And no, it does not show an increase in injuries at fixed-site parks. Perhaps you should read it again. Check the chart in figure 1.

Why does the CPSC have an agenda? They indicate a long-term increase in injuries ('93 to '00) but do not indicate the decrease this year at fixed-park installations. If they indicated a decrease, or God forbid pair the stats with attendance, they wouldn't be able to justify asking for more money from Congress.

What I find ironic is that the portion of the industry they do have regulation over actually saw an increase in injuries. That's a real kick in the twig and berries, isn't it?

Did anyone notice the substantial increase in injuries from inflatables, or the fact that more than 10% of the injuries overall come from inflatables? I think we need a federal commission to investigate the health risks of giant amusement balloons (GAB's) as they're obviously a threat to the well being of our children.

-----------------
Jeff - Webmaster/Admin - CoasterBuzz.com
"From the global village... in the age of communication!"
Watch the grass grow!

I dont trust these park's to give us accurate #'s, 2 people here say they don't believe the #'s from the report. I don't believe the #'s that come out of park's, until this year CA did not have to report injury's and Fl does not have to report accident's. I just want accurate #'s from both side's of the fence, both sides are making up #'s here. What is truly wrong with having park's report ALL accident's in park's, most company's including mine have to report all injury's at a work place.
Can someone give me the injury stat's for FL and CA from last year since they don't get reported, so we can't tell if the fixed park's injury's are up or down.
-------------
Army rangers lead the way
*** This post was edited by supermandl on 8/28/2001. ***

Sorry, Jeff, but I'm going to have to disagree with your interpretation of Figure 1 in the report.  While the 2000 fixed-site data show a reduction in the injury frequency vs. 1999, the 2000 number is still 72% higher than the 1993-1996 average.  Furthermore, looking at Table 2 and doing some further calculations, we see that the fixed-site injury rate for 1997-2000 is 21.3 per million, while for 1993-1996 it was 13.8.  This shows an increase of 54% in the injury rate.  These figures indicate, in my opinion, that the long-term trend has been for an increased injury rate.

Of course, all this analysis assumes that the figures reported are accurate estimates of the number of actual injuries that occurred.  I have not tried to assess the appropriateness of the study techniques employed by the authors for their intended purpose.

Of course, none of this proves that amusement park rides in general or roller coasters specifically are unsafe.  Neither does it prove that heightened federal regulation of amusement parks would lead to improved safety.  However, if you believe the data presented by the report are accurate, SOMETHING happened to increase the injury rate sometime between 1994 and 1998.  In my opinion, somebody should find out what happened and do something about it.  Maybe the reduced figure for 2000 indicates that somebody IS doing something about it.

I LOVE coasting.  Jeff, I would like to help you keep the feds out of places they don't belong.   I can help with the data analysis since I do it for a living (I'm an actuary).  Right now, though, the data we have indicate that the feds (or somebody) should intercede to stem the apparent rise in the observed injury rate.  Are there specific reasons we should disbelieve these data?

This is a great site, Jeff.  Sorry to have to disagree with this one.

I don't disagree with the long-term trend (well, actually I do because the CPSC has themselves admitted to an error margin of +/- 50%), what I'm saying is that they neglect to mention the decrease this year. As I said, it's easier to ask for more money when you can demonstrate need. Using long-term numbers, they can, but short-term, it's less of a case.

As Jim mentioned, their methodology is terrible and prone to incredible error.

-----------------
Jeff - Webmaster/Admin - CoasterBuzz.com
"From the global village... in the age of communication!"
Watch the grass grow!

OK, Jeff, I thought about this some more.  You're right about their methods; they ARE inappropriate, but not for the reasons you are naming.  In this analysis, the "why" is the important part.

The cited error in the estimate of the total annual injuries of around 50% (actually the standard deviation is more like 25% with the 95% confidence interval used) is not the problem itself.  An error level of this magnitude has been accounted for in the statistical tests used by the analysts writing the report (I think).  However, this level of error indicates that their sample size is quite small compared to the overall number of hospitals in the US.  Also, it seems that particular hospitals would tend to receive almost all of the fixed-site injury cases -- the one nearest the park where the injury occurred.  If these "near-amusement-park" hospitals are overrepresented in the sample, the resulting estimate of the total injuries will be excessive, and vice versa.  It appears here that the excessive case might be taking place, since half the reported injuries came from one hospital.  This hospital may be a "near-amusement-park" hospital.  With a sample of only 100 hospitals being used in the NEISS, there is huge volatility in the possible outcomes.  With no "near-amusement-park" hospitals included in NEISS, we may underreport the fixed-site injury level.  Just one "near-amusement-park" hospital being included in the 100 may be too many, grossly overstating the fixed-site level.

Also interesting is the fact that the fixed-site level appears to accelerate in 1997, the year in which the NEISS sampled hospital list changed.  OK, CPSC, where exactly is this "killer" hospital, and did it enter the NEISS in 1997?

This is why these kinds of analyses have to be looked at so carefully.  The data appear compelling, but there are serious questions regarding underlying method.  There needs to be further disclosure of the data underlying this study.

ACBLuke:

The sample is 1.5 to 2% of hospital emergency rooms.in a sample stratified for size with a special catagory for children's hospitals.  This is a good way to sample for hazards with many thousands of exposure points such as sutomobile accidents, baby cribs, or garden hoses.  it is not a valid way to sample for hazards with relatively few exposure points such as the 200 or so amusement parks.  You covered the reasons pretty well.

As near as I can tell from the limited information in the CPSC's reports for the last several years, the change in the sample group in 1997 was not the cause of the increase in accidents reported.  Rather, it appears that one hospital (#58) located in the northeast had a sudden upturn in the number of cases reported.  It seems that this hospital was in the sample prior to 1997.  Since I don't know the exact location of the hospital, I can't speculate on the exact cause for the the increase.  The increase in cases from this one hospital was the entire cause for the increase in estimated injuries.  The sum of injuries based on all 99 other hospitals combined has been essentially unchanged.

Just a comment to BB: maybe you misunderstood what I said.  I noted that parks are currently doing a very good job of reporting accidents, especially considering that in many states these things aren't necessary.  And as for those Florida parks: even though they're not required to report them, the press manages to find out, in this media savvy age of ours.  Long story short: 99.99% of the time, if an accident occurs, the public finds out.
-----------------
The Luv Monkey has spoken...
NITRO kicks it into high gear...only at Six Flags Great Adventure!
Track Record: 28 and counting...

Here's a statistic for the math minded:  Injuries per miles coasted or I/mc.  In the past few years of coastering, the only injuries I've seen  (minor ones) were from kids running in bare feet on concrete at water parks.  The question is, should we ban a) FEET, b) concrete, c) kids, d) water parks, or e)None/All of the Above - your choice.  I've always thought fixed-site parks are WAY safer than fairs and carnivals, and feel only more justified in that. 

And now, a third-grade word problem:  If the total number of injuries remains relatively unchanged while attendance is increasing, can a politician (or a news station) do basic math?

-----------------