RobertNiles.com
About Robert
Reporting Help
Finding Data on the Internet
Questions for Candidates
Stats Lessons
Mean
Median
Percent
Per capita
Standard Deviation
Margin of Error
Data Analysis
Sample Sizes
Stats Tests
Help Board
Bookstore


Statistics for the Utterly Confused

This book offers a super-accessible approach to the much-misunderstood subject of statistics.
More information
-->

Home > Statistics Every Writer Should Know > The Stats Board > Discusssion

drug data analysis, thai style
Message posted by Phil Verhoef on May 16, 2000 at 12:00 AM (ET)

ok, i have chunk of data that was generated by a small hospital in rural thailand, and i've been given the task of analyzing it.
a group of 90 rice farmers were given an herbal tea OR a placebo tea on Day 0, and specific lab tests were down: enzyme levels, blood counts, etc. then, on day 7, all these parameters were again measured to determine if the herbal tea had had a measurable effect on certain enzyme levels.

Now, this is where it becomes more complicated. 1 of the measured enzymes is lowered in the presence of pesticides. BUT, not all of the farmers had lowered levels of this enzyme. so, i did the analysis on all the farmers that received the treatment. But, is it justified/can i do the analysis focusing only on the farmers who had "below normal" levels of this enzyme on Day 0? I did this analysis, and interestingly, the tea seemed to work much better in the farmers with below normal enzyme levels, compared with the whole group.
any thoughts?
thank you.
sorry no capitals. broke my fingers on my left hand, am typing one-handed. :)
-Phil


READERS RESPOND:
(In chronological order. Most recent at the bottom.)

Re: drug data analysis, thai style
Message posted by Chad Allen on May 18, 2000 at 12:00 AM (ET)

Sorry to hear about your accident.
Now, if you are fairly certain that the pesticide level significantly effects enzyme level, you need to establish a new variable. This variable should be if the farmer came into contact w/the pesticide vs. no pesticide contact. Analyze this variable in conjunction with the group variable (i.e. experiment vs. placebo group). Might I suggest this analysis be performed via a 2-way ANOVA/MANOVA. However, it is imperative that you use both variables in the analysis, as not to inflate familywise error rate and increas the likelihood of causing Type I errors.

To be certain that pesticide level effects enzyme level, you could either review the published literature or perform an analysis to confirm this.

Hope I was of some assistance, please let me know if I can do anything else.
Chad


Re: drug data analysis, thai style
Message posted by Phil Verhoef on May 18, 2000 at 12:00 AM (ET)

thanks, chad. (sigh) that's as i suspected... the stats don't scare me... it's just that when they conducted the sampling, nobody controlled for pesticide levels, or measured them, and the idea that this drug affects them is really just an idea... and no research has been done that i've found. thanks again! i think i can do some minimal analysis with some of the other data that won't, for instance, be altered by pesticide levels, but... another study or 10 should be done...
-phil



Your $5 contribution helps cover part the $500 annual cost of keeping this site online.

Niles Online Sites:RobertNiles.comTheme Park InsiderViolinist.com

RobertNiles.com™, the site, content and services 咀opyright 1996-2002, Robert Niles.
All rights reserved. Questions? Comments? Read my Privacy Policy, or E-mail me!