There are many details involved in the design of a research project. Research designs are used to investigate a specific question or hypothesis. The hypothesis is a prediction of what the researcher expects to find and helps guide the researcher in choosing the types of methods that are used. Researchers generally choose from a few basic types of research designs or methods to find the one that best fits the question or topic they are investigating.


The goal of most research studies is to find the relationship between two things or variables. A hypothesis is a statement that makes a prediction about how one thing -- the independent variable -- will be related to another -- the dependent variable. Research methods reflect the best way to find this relationship, for instance, through an experiment or a descriptive study. Experiments are ways to control certain factors in order to look specifically at the relationship between two variables and can help the researcher to discover causality. In contrast, descriptive studies simply observe and describe a phenomenon as it is, without the process of experimenting.


A pre-experimental or non-experimental design is a method that looks at a single group, without the researcher comparing it to any other groups. In contrast, a quasi-experimental design is one in which different groups are compared, using an experiment; however, the groups are not randomized. Finally, an experimental research design is a method in which at least two groups are used, one of which is a control group. A control group is a group that is not given any treatment during the experiment. Also, for a true experimental design, subjects are selected at random and, therefore, the results can be generalized.


Pre-experimental designs are flawed in that they do not compare two groups and do not use control groups; that is, groups that do not receive the treatment. They can be good for learning about a phenomenon but cannot predict a causal relationship. Quasi-experimental designs have greater validity because they do employ comparison groups, and can show a difference between the groups. However, these methods do not use randomization and, therefore, the results cannot be generalized. The use of a true experimental design is the only way that a prediction about or relationship between two variables can be truly confirmed or disproved.


While experimental research methods, also known as quantitative research methods, are best in making a prediction about the relationship between two variables, there are qualitative research methods that can also be beneficial for some types of studies. These methods are good for openly exploring a topic about which little is known, and understanding subjective information, such as the individual experiences of people. Qualitative research methods include case studies, participant observation, interviewing and survey research.