A hairy depiction of a crazed scientist running around with exciting coloured liquids is, in a very small way, true (‘no running in the lab’ woops!). We do spend a disproportionate amount of time dealing with liquids, albeit normally colourless, depending on the discipline. Lab based research requires some tiresome patience and tedious repetitive motions – all too often involving pipetting the tiniest amounts of solutions from one tube to another – I mean way, way too often. After a couple days like this you really do feel like a little lab machine! In the end, hopefully, your eyes regain their sight and the bigger picture of your research will come into focus when you produce some meaningful results. That, in one way, is the big and the little of research.

The even bigger picture to discuss is study design. You can spend as many hours as you want in the lab until you get pipetting induced hand strain, but it will all be meaningless if you haven’t planned your experiment properly – and you can definitely save time by thinking things out before rather than later (begrudging advice from experience). For those with a general interest in science and wanting to be able to critically appraise the (please don’t do it) Daily Mail articles…study design on a whole weighs a huge bearing when appreciating research.

Science doesn’t come without its politics and there is a hierarchy to study design itself definitely worth noting. Moving away from actual experimental design and onto population-based research, at the bottom of the triangle are cross-sectional observational studies which analyse data at a certain point in time such as at an epidemiological level e.g. cancer and smoking. They are not longitudinal in design so no cause-effect can be established but they often provide a good starting point for further investigations and showing changing population trends.

Case-control studies generally use retrospective factors to try and identify contributors to a disease e.g. comparing people with a common outcome/disease of interest and those without, and looking back in time to see any influencing risk factors unique to the outcome/disease group.

Cohort studies are usually longitudinal and groups of people of interest are followed up prospectively with respect to a disease or outcome of interest to try and identify risk factors.

Randomised control trials are reaching for the stars top of the hierarchy as they test hypotheses created from preliminary observations. However, these studies are not always feasible – for example a trial on whether smoking caused cancer.

Finally, systematic reviews and meta-analyses are the most reliable sources of information as they pull together all the research conducted on a topic and amalgamate a single conclusion. Each study design has its merits and faults and The BMJ has a good resource for anyone really interested.

The type of study being discussed and the amount of volunteers involved are two things I would recommend noting before deciding that you need to drink 3 glasses of champagne every night ;). We will discuss more issues to recognise in upcoming posts :).

As always, please contact us if you would like to participate an article or have any requests – we’d love to hear from you.

Sophie

 

Advertisements