Hi

To embellish a bit on Rick's comments.  It is not just correlational
data.  It is correlational data over time.  The classic example of
spurious correlation is just such a relationship (i.e., the high r
between number of storks and number of babies born in Copenhagen).

So, has anything much changed since 1988, when the SSRIs were
introduced?  Here are a few candidates that I found searching around a
bit (although arguments might be made about the exact time of onset of
the decline).

unemployment rate
teen pregnancy rate
school dropout rates
use of alcohol by high school seniors

I was also curious about the raw data for the study (but have not
actually looked at the study, so reader beware!).  I found overall
suicide statistics at:

http://krusekronicle.typepad.com/.shared/image.html?/photos/uncategorized/07aas.gif


It is not completely obvious (to the very imperfect eyeball) whether
the decline started around 1988 or a decade earlier.  Suicide rates
through the 1970s were consistently 1% higher than during the 1980s,
after which the decline continued (accelerated?).  Browsing around some
of the other sites that popped up in google further complicated the
story.  Essentially, the time series data looks different as a function
of such factors as: different age ranges and means of suicide (decline
after 1988 clearest for death by motor vehicle exhaust ... perhaps SSRIs
have a highly specific effect?).

Certainly lots of room for improvement of the study, although
interesting data nonetheless.  Be interesting, for example, to see
comparable data from countries where newer SSRIs were introduced later
or earlier, or perhaps not at all.

Take care
Jim

James M. Clark
Professor of Psychology
204-786-9757
204-774-4134 Fax
[EMAIL PROTECTED]

>>> [EMAIL PROTECTED] 12-Dec-06 2:50:27 PM >>>
While I appreciate the dangers of using correlations to demonstrate a 
causal relationship I think that we should consider  these data 
carefully. If you are suggesting that it is only a spurious
relationship 
then what would you propose as a logical "third variable" to account
for 
the apparent connection?

-Don.

Paul Brandon wrote:

> At 8:02 AM -0500 12/12/06, [EMAIL PROTECTED] wrote:
>
>> Last weeks issue of Time Magazine, with cover date 14/4/06, has an 
>> article The year in medicine A to Z.  One of the entries is
>>
>> DEPRESSION
>>
>> Researchers still don't understand why severely depressed teenagers

>> are more likely than adults to commit suicide while taking 
>> antidepressant drugs like Paxil, but a major study out of UCLA 
>> concluded that the drugs do more good than harm. Starting in the 
>> early 1960s, the annual U.S. suicide rate held fairly steady at 12
to 
>> 14 instances per 100,000--until 1988, when the first of a new 
>> generation of antidepressants, the selective serotonin reuptake 
>> inhibitors, was introduced. The suicide rate has been falling ever 
>> since, to around 10 per 100,000. The investigators estimate that 
>> nearly 34,000 lives have been saved.
>>
>> Since much of the discussion has been focused on the fact that 
>> antidepressants don't work, how would you explain this result?
>
>
> Facetiously:
> Since the time period in question coincides with the birth of the 
> InterNet, one could just as easily say that an increase of virtual 
> suicide by internet addiction is responsible.
> In other words -- a correlation is still a weak argument.
> Were the authors receiving drug company support by any chance?


-- 
Don Allen
Department of Psychology
Langara College
Vancouver, B.C., Canada
V5Y 2Z6

604-323-5871


---
To make changes to your subscription go to:
http://acsun.frostburg.edu/cgi-bin/lyris.pl?enter=tips&text_mode=0&lang=english






---
To make changes to your subscription go to:
http://acsun.frostburg.edu/cgi-bin/lyris.pl?enter=tips&text_mode=0&lang=english

Reply via email to