Before most scientific research is published, it goes through a process called peer review.

Peer review means that other experts in the same field carefully examine the research before it is published.

These experts are called “peers” because they have similar knowledge and training as the researchers who wrote the study.

What happens during peer review?

When a scientist submits a research paper to a journal, the journal sends it to several independent experts.

These reviewers check things like:

• Was the study designed properly?

• Are the methods appropriate?

• Do the data actually support the conclusions?

• Are there mistakes, bias, or missing information?

The reviewers then send feedback to the journal editor.

What the reviewers can recommend

After evaluating the paper, reviewers usually recommend one of four outcomes:

1. Accept – The paper is strong and can be published.
2. Minor revisions – Small changes are needed before publication.
3. Major revisions – The paper needs significant improvement.
4. Reject – The study has serious problems and should not be published.

Most papers actually go through multiple rounds of revision before they are accepted.

Why peer review matters

Peer review acts as a quality control system for science.

It helps make sure that:

research methods are sound

conclusions are supported by evidence

mistakes are caught before publication

In other words, it helps maintain scientific credibility and reliability.

Peer review does not guarantee that a study is perfect or correct.

Mistakes can still slip through, and scientific understanding can change as new evidence emerges.

But peer review does mean that:

The research has been carefully evaluated by experts before being added to the scientific literature.

When scientists measure something using a sample (a smaller group taken from a larger population), the result they get is an estimate, not a perfect measurement.

The margin of error tells us how much the real value might differ from that estimate.

Think of it as the “wiggle room” around a result.

A simple example

Imagine a survey finds that:

60% of students prefer learning with comics rather than traditional textbooks.
The study reports a margin of error of ±4%.

This means the true value in the full population is likely somewhere between:

56% and 64%.

So the result is best interpreted as:

Around 60% of students prefer comics, give or take about 4%.

Why margin of error exists

Because researchers rarely study every single person in a population.

Instead, they study a sample and use statistics to estimate what the whole population probably looks like.

But samples naturally contain some random variation, so results are reported with a margin of error to reflect that uncertainty.

What affects the margin of error?

Several things influence how large or small it is:

1. Sample size
Larger samples give smaller margins of error because they better represent the population.

2. Population variability
If people's responses vary a lot, the margin of error tends to be larger.

3. Confidence level
Studies that aim for higher certainty (like 99% confidence instead of 95%) will have larger margins of error.

Important thing to remember

The margin of error does not mean the study is wrong.

It simply tells us:

How precise the estimate is likely to be.

A small margin of error = more precise estimate.
A large margin of error = less precision.

When scientists run an experiment, they want to know whether their results are real or if they might have happened just by chance.

The significance level is the rule scientists set to help decide that.

Think of it as a cut-off point for how much randomness we're willing to accept.

The most common threshold: p < 0.05

In many scientific studies, researchers use 0.05 (or 5%) as the significance level.

This means:

If the probability that the result happened by chance is less than 5%, scientists consider the result statistically significant.

In simple terms:

p < 0.05 → unlikely to be random → result is considered meaningful

p > 0.05 → could easily be random → result is not considered significant

Imagine a study testing whether using comics helps students learn science better than textbooks alone.

If the results give a p-value of 0.03, that means:

• There is a 3% probability the difference between the groups happened just by chance.

Because 3% is less than 5%, the result is considered statistically significant.

So researchers would say:

The comic-based learning method had a statistically significant effect on student learning.

Important thing to remember

Statistical significance does NOT mean something is large, important, or useful in the real world.

It only means the result is unlikely to be caused by random variation in the data.

That’s why scientists also look at effect size, sample size, and study design before drawing strong conclusions.

Sample Size:
A sample size is the number of people, items, or data points chosen from a larger population to take part in a study.

Why do scientists use a sample instead of studying everyone?
If you want to learn something about a large group, let’s say people who eat an apple a day, it’s unrealistic to try to study every single person in the world who does that. In research, the population refers to everyone who fits the criteria being studied. But studying an entire population is rarely possible because of practical limits like time, cost, access, and ethics.

Instead, researchers select a smaller group that represents that larger population. This is the sample. The goal is to choose it carefully so that what we learn from the sample can reasonably reflect what’s happening in the wider group.

Scientists don’t just pick a number at random, though. Sample sizes are usually calculated using statistical formulas that take into account things like the desired confidence level, the acceptable margin of error, and how much variation exists within the population.

In general, larger sample sizes tend to produce more reliable results. With more data, researchers can be more confident that the patterns they observe aren’t due to chance, and that their conclusions are more likely to apply to the broader population they’re trying to understand.

This is the maths bit. Everything about the results from an experiment is usually turned into numbers in some way, whether that’s tumour size, blood pressure levels, reaction times, test scores, or survey responses. Assigning numerical values to results allows scientists to measure change in a clear, structured way, making it easier to compare results between groups, across time, or under different conditions.

But statistics isn’t just about collecting numbers, it’s about helping researchers understand what those numbers actually mean. Natural variation exists in almost everything we measure. No two people, samples, or experiments will produce identical results every single time. Statistical analysis helps scientists determine whether a change they are seeing is likely due to the treatment or condition being tested, or whether it could simply be random chance or natural variation.

This often involves calculating averages, comparing groups, and using specific statistical tests designed to measure how confident researchers can be in their results. You might see references to things like significance levels, p-values, or confidence intervals. While these can sound intimidating, they are essentially tools used to answer one key question: Is this result likely to be real, or could it have happened by accident?

By using statistical analysis, scientists can move beyond simply observing patterns and instead provide evidence to support their conclusions. It helps ensure research findings are reliable, repeatable, and meaningful, which is essential when those findings might influence medical treatments, public health advice, education strategies, or policy decisions.

You’ve probably heard someone say something like, “Studies show that X is linked to Y,” and it can sound pretty convincing. But in science, the word linked carries a very specific meaning, and it doesn’t always mean one thing is causing the other. One of the most common and most misunderstood concepts in research is correlation. It sounds intimidating because it’s a statistics term, but at its core, it’s simply a way scientists measure how two things change alongside each other.

Correlation: 

This is technically a statistics word - oh no. 

But really, all correlation really is, is an actual calculated number that describes the relationship between two variables, indicating the extent to which the change in one variable is associated with the change seen in the other variable. 

But remember! Correlation does NOT equal causation. 

It just strongly implies it. 

Correlation is incredibly useful because it helps scientists spot patterns, generate new research questions, and identify relationships worth investigating further. But it’s only the starting point, not the final answer. Just because two things move together doesn’t mean one is responsible for the other — sometimes there’s coincidence, sometimes there’s a hidden third factor, and sometimes the relationship runs in the opposite direction. Understanding the difference helps us interpret research more critically and prevents us from jumping to conclusions based on patterns alone.