We explain what inferential statistics is and its different uses. Also, examples and descriptive statistics.
What is inferential statistics?
It is called inferential statistics or statistical inference to the branch of Statistics in charge of making deductions, that is, infer properties, conclusions and trends, from a sample of the set. Its role is to interpret, make projections and comparisons.
Inferential statistics usually employ mechanisms that allow you to carry out such deductions, such as point estimation tests (or confidence intervals), hypothesis tests, parametric tests (such as mean, difference of means, proportions, etc.) and non-parametric (like the chi-square test, etc.). Correlation and regression analysis, time series, analysis of variance, among others, are also useful.
Hence, inferential statistics is extremely useful in population analysis and trends, to get a possible idea of its actions and reactions in the face of specific conditions. This does not mean that they can be predicted faithfully, nor that we are in the presence of an exact science, but it does mean a possible approximation to the final result.
Examples of inferential statistics
Some examples of the application of inferential statistics are:
- Voting trend polls. Before an important election, various pollsters poll public opinion to collect relevant data and then, having the sample analyzed and broken down, infer trends: who is the favorite, who is second, etc.
- Market analysis. Companies often hire other specialized marketing companies to analyze their niche markets through various statistical and differential tools, such as surveys and focus groups, from which to deduce which products people prefer and in what context, etc.
- Medical epidemiology. Having specific data on the affectation of a certain population by one or more specific diseases, epidemiologists and public health specialists can reach conclusions regarding what public measures are necessary to prevent these diseases from spreading and contribute to their eradication.
Descriptive statistics
Unlike inferential, descriptive statistics does not care about conclusions, interpretations or hypotheses based on what is reflected by the sample, but rather on the ideal methods for organizing the information it contains and highlighting its essential characteristics.
In other words, it is about “objective” statistics, committed to presentation of data (textual, graphical or by tables) and the mathematical operations that can be applied to obtain greater data margins, new information or exact frequencies and variabilities.