Overview
Data is essential for informed decision-making and understanding trends in our society.
This guide provides fundamental knowledge to help you better navigate, interpret, and use official statistics responsibly. It explains where data comes from, how to analyse it correctly, how to communicate findings responsibly, and why seasonal adjustments matter.
How to use data correctly
Common errors in statistics
Some common errors that you may encounter include:
Correlation is a statistical measure that indicates the extent to which the value of two or more variables move in relation to each other. Positively correlated variables tend to move in the same direction, while negatively correlated variables tend to move in opposite directions with one another.
However, it may not necessarily be the case that the change in one variable causes the change in the other. On the other hand, causation means that the change in one variable causes the other variable to change.
Example: Hot sunny weather would cause an ice-cream to melt and cause sunburn (with prolonged sun exposure). Melting ice-cream and getting a sunburn are correlated, where they tend to occur together in the hot sunny weather. If the presence of the hot sunny weather was ignored, it would be wrongly concluded that melting ice-cream causes sunburn!
Beware of results from small sample sizes, or polls
When testing out a hypothesis, it may not always be possible to collect data for the entire population due to logistical or financial reasons (e.g. research budget). Hence, an option for researchers would be to use a smaller group, which is known as a sample (Figure 10).
Population vs Sample
However, small sample sizes could affect the reliability of the results. One reason is because small sample sizes decrease the statistical power of a study, which means that there is a lower likelihood of detecting a true effect that exists in the entire group, via the study. Another reason could be that the sample is not representative of the population, like online polls, where only people who feel strongly about a subject would respond to the polls. This means the results are skewed towards this group of people, when the majority could be neutral about the subject. As such, robust statistical reporting or research typically requires a large enough sample size. To circumvent non-representativeness, one way is to conduct simple random sampling, where samples are chosen strictly by chance, so that all members of the population have the same chance of being selected for the study.
Concluding Remarks
As statistics is a broad field, the content above above serves as a brief and simple introduction to the different types of data, ways to analyse data, and the common pitfalls of using statistics. With this new-found knowledge, enjoy exploring and working with data to gain useful insights.