What is a meta-analysis (in the context of medical research)?

618 views

What is a meta-analysis (in the context of medical research)?

In: Other

2 Answers

Anonymous 0 Comments

Science is fluid. This means that what we think is right today may be flipped completely upside down tomorrow.

At its core, science is nothing but a checklist: a series of steps we take to learn about the natural world. We call this series of steps “the scientific method.”

To learn about the world via the scientific method, we conduct studies to answer the questions we pose. Because the world is complex, no single study will ever be an absolute truth. Any study you see has important limitations, which is why it is important to approach the same question through different routes to get a better picture. We do studies over and over and over until we have more confidence in our explanation.

A meta analysis in turn is an analysis of analyses. A researcher combs the literature to see all the studies that examined phenomenon X, and uses specific statistical procedures to yield an overall effect size, which tells you how big a particular difference is. It’s the most accurate picture we have to what the effect is like in the real world.

That’s because a lab and the real world are two different planets. The lab is a very quiet room in which any slight noise is enough to startle you. But the real world is noisy, and you probably won’t even sense, yet alone be startled by the very same sound. A meta analysis can help us judge whether the sound that spooks in the lab actually spooks on the street.

Now let’s use this analogy: a digital photograph of a cat.

Each study is a pixel that yields a color that we think paints a cat. A meta analysis is gathering all of these pixels into one image to see if a cat is indeed there.

Anonymous 0 Comments

A meta-analysis looks across several different independent pieces of research and assesses the implication of the totality of their conclusions.

Basically, if lots of different researchers looked at the same problem in similar ways, you may be able to make conclusions about the results they found with more certainty than you would be able to with any one of them on its own.