This chapter provides an overview of sampling theory, which addresses the challenge of creating representative subsets of a larger population. Sampling is essential in statistical research (and not only), as it allows for making accurate predictions and testing hypotheses without needing to collect data from the entire population. This chapter covers different sampling strategies and their practical applications. In many research projects, you may need to create a new dataset to answer specific research questions. The data acquisition process involves decisions about the type, amount, and number of samples required. Several challenges, such as bias, representativeness, and response rates, must be managed to ensure reliable data. For example, conducting surveys on difficult-to-reach populations like the homeless or designing long-term longitudinal studies on chronic diseases presents unique obstacles that must be addressed through careful planning and sampling techniques. The chapter begins by discussing the importance of formulating precise research questions (RQs) and hypotheses, which drive the data collection process. A well-defined RQ identifies the knowledge gap being addressed, while hypotheses provide predictions that can be tested using statistical methods. This step is critical to guide the design of the experiments and ensure that the correct data are collected. The chapter further introduces the key concepts of survey sampling, where data are gathered from a population sample rather than the entire group. The sampling process is divided into two primary categories: non-probability sampling, which relies on non-random methods such as convenience or judgement sampling, and probability sampling, where every member of the population has a known chance of being selected. Lastly, the chapter discusses more advanced techniques such as stratified sampling and cluster sampling, which are used to ensure that samples accurately represent key subgroups of the population. It also covers methods like random sampling with and without replacement, which ensure that the data collection process is unbiased.