In scientific research, an experiment typically begins with a hypothesis, which is a proposed explanation based on limited evidence. Researchers design experiments to test this hypothesis by manipulating an independent variable while keeping other variables constant to measure its effect on a dependent variable.