{"id":98946,"date":"2019-12-03T12:23:41","date_gmt":"2019-12-03T11:23:41","guid":{"rendered":"https:\/\/www.scribbr.nl\/?p=98946"},"modified":"2023-06-21T14:08:40","modified_gmt":"2023-06-21T12:08:40","slug":"experimental-design","status":"publish","type":"post","link":"https:\/\/www.scribbr.com\/methodology\/experimental-design\/","title":{"rendered":"Guide to Experimental Design | Overview, Steps, & Examples"},"content":{"rendered":"

Experiments<\/strong> are used to study causal relationships<\/a>. You manipulate one or more independent variables<\/a> and measure their effect on one or more dependent variables.<\/p>\n

Experimental design<\/strong> create a set of procedures to systematically test a hypothesis<\/a>. A good experimental design requires a strong understanding of the system you are studying.<\/p>\n

There are five key steps in designing an experiment:<\/p>\n

    \n
  1. Consider your variables<\/strong> and how they are related<\/li>\n
  2. Write a specific, testable hypothesis<\/a><\/strong><\/li>\n
  3. Design experimental treatments to manipulate your independent variable<\/strong><\/li>\n
  4. Assign subjects to groups, either between-subjects<\/strong> or within-subjects<\/strong><\/li>\n
  5. Plan how you will measure your dependent variable<\/strong><\/li>\n<\/ol>\n

    For valid conclusions, you also need to select a representative sample<\/a> and control any\u00a0extraneous variables<\/a> that might influence your results. If random assignment<\/a> of participants to control and treatment groups is impossible, unethical, or highly difficult, consider an observational study<\/a> instead. This minimizes several types of research bias, particularly sampling bias<\/a>, survivorship bias<\/a>, and attrition bias<\/a> as time passes.<\/p>\n

    <\/p>\n