analysis of the relationship between particulate air pollution and lung cancer

#1
Hi,

I am currently doing a study on Lung cancer mortality and air pollution my hypothsis is to see if there is a link.

i have 6 variables ( i will just copy and paste as it is easier) and have data for all

 Low skilled workers – The percentage of houseshold where teh head of the house was employed in in a social class 4 (semi skilled) or social class 5 (unskilled). This is used a indicator of the percentage of population with a low income.
 Ethnicity – Percentage of districts residents from the new comenweatlth and Pakistan. Tkane from the 1991 Census.
 Limiting longterm illness – Percentage of the district who report suffering froma long term illness, data from 1991 census. This is used as ann indicator of general health of teh local population.
 Population density – Persons per km2. Taken from teh Census.
 Particulate Pollution – Average population dose os particulate pollution (PM10) ug m-3 . PM10 particles are very fine particles produced by diesel engines.
 Smoking -percentage of people who smoke.

All of these induvidually correlated with lung cancer give this... in the order in which i said

.413 .326 .216 .449 .446 .645 (pearsons correlation)
17.07 10.64 4.64 20.18 19.87 41.58 (r^2)

clearly smoking is the most signf but im looking at air pollution which is half as much.

After Smoking, Population density and particulate pollution is the next most sig nif.....does this mean i should focus the rest on them? im getting confused when it comes to doing regression analysis so that i get a signif when other indicators are held contant. Could anyone maybe point me in the right direction? As in what to do for regression analysis and what it actually shows and proves...

Im using SPSS..

Thank you for reading that!! and hope you can help

p.s
Sorry for my spelling!
 

CB

Super Moderator
#2
im getting confused when it comes to doing regression analysis so that i get a signif when other indicators are held contant. Could anyone maybe point me in the right direction? As in what to do for regression analysis and what it actually shows and proves...
Since your dependent variable (lung cancer/no lung cancer) is dichotomous*, you're probably looking for logistic regression. As you indicate, this will allow you to look at the predictive value of each IV with all the other IV's held constant. In other words, to see the unique predictive value of each IV, with the others controlled for. The logistic regression will give you a B value for each IV (like a regression weight) and, more usefully, an Exp(B), which is basically an odds ratio. The Exp(B) indicates the increase in odds of cancer occurrence for each one unit increase of the given IV, with all other IV's held constant.

Depending on the nature of your study you may be able to then convert the Exp(B) / Odds Ratio figures into Relative Risk figures, which can be more useful. That's maybe something for later though.

EDIT: *Actually, is your DV dichotomous? It seems like you're doing these analyses at a district level or something like that - might be good if you could describe your DV data in a bit more detail :)
 
#3
thanks for the reply, im afraid stats are not my strong point, and have trouble explaining what the results even show, even what you said i am struggling to get a grip of :S

But yes sorry, the study i am doing is data from 402 districts from across england. and is an ecological study.

Thank you very much for your input it is greatly apreciated.

also i should have said im not really looking at lung cancer / no lung cancer, im lookina the Standerdised mortality reates of each district and to see how they are effected. an SMR of 100 being normal <100 less lung cancer >100 higher lung cancer.
 
Last edited:

Link

Ninja say what!?!
#4
It appears the analysis is more on a population level than an individual level. Do you have specific cases of lung cancer, or just counts of cases? If its the latter, then look into Poisson regression.