What is this study design?

August 4, 2011 8:41 PM

Help me design my research project: I'm collecting numerical data on a group of people compared to age-matched averages. Eg, I have a group of left-handed people and I want to know if their blood pressure differs from the average age-matched blood pressure. I don't have a control group, just data on blood pressure. What study design is this? Hint: it's not case control. I'm stuck.
posted by superfish to Science & Nature (6 answers total)

How do you know the population you're getting the averages from and the sample group you've selected are comparable?
posted by mandymanwasregistered at 11:59 PM on August 4, 2011

Response by poster: The study group is a small subset of the larger population, who happen to have characteristic A. I'm testing to see if characteristic B differs in the small group from in the general population. Does that make sense?

I just know it's not a case control trial, because I don't have a control group per se, rather data from the general population. But what is it?
posted by superfish at 12:43 AM on August 5, 2011

So, the concern is that the age-matched population data would include people with the trait you're trying to measure? I'm not sure why they wouldn't be considered controls from what you're describing, so am just trying to get a better idea of what keeping your from considering these controls in that sense.
posted by goggie at 7:31 AM on August 5, 2011

Best answer: It would be a cross-sectional study, as you are measuring the frequency of a particular exposure and/or outcomes (here, handed-ness is the exposure and the outcome is BP) in a defined population at a particular point in time. While you are looking to see if there is an association between the two, you won't be able to say much about causality with this study design.

Age-matching is done to remove the confounding effect of age, which is a risk factor for hypertension.

You can still describe your two groups as cases (left-handed) and controls (right-handed). And you can analyze the data using the same methods you would for a case-control if you want (and report an odds ratio).

As mandyman pointed out earlier, it's very important to ensure that your sample is representative of the population from which it is drawn. So if your sample of left handers is age 20-30 yrs, but the actual population ranges in age from 20-70, your sample is biased (and is also very unlikely to find much hypertension due to the relative youth). Same thing if, for example, the target population is 50/50 women/men but your sample is all men.

Good luck!
posted by lulu68 at 7:44 AM on August 5, 2011

I don't think there's a special name for when you have individual level data in only one group. The analysis you're doing is identical to a usual cohort cross-section where you matched on age and adjusted for nothing. All that enters into such an analysis is the averages.
posted by a robot made out of meat at 8:19 AM on August 5, 2011

Response by poster: Thanks all! It's not really on handedness and BP, that was just an example. It is MUCH MORE EXCITING than that. Now, to get some results!
posted by superfish at 3:27 PM on August 5, 2011