1, completely random design. Also known as grouping design, there are two main forms: first, all homogeneous subjects are randomly assigned to various treatment groups by completely random grouping, and each group receives different treatments; The second is to randomly sample from different people, observe the same index and make a comparative study.
2. Pairing design. In medical research, there are two main forms of pairing design: one is allogeneic pairing, in which subjects are paired according to the same or similar conditions, and two subjects in each pair are randomly assigned to different treatment groups through stratified randomization; The second is self-pairing, in which the same subject uses two different detection methods or observes the same index in different parts of the same subject.
3. Random block design. Random block design, also known as compatibility group design, usually divides the subjects into several blocks (or compatibility groups) according to the same or similar properties, and randomly assigns the subjects in each block to different treatment groups for different treatments through stratified randomization.
4. Repeat the measurement design. Repeated measurements are often recorded in medical experiments, that is, the same index is measured many times at different time points. This design is called repeated measurement design, that is, after one or more treatments are given, the observed values of an index of the same subject are obtained at multiple time points.
Basic principles of experimental design
1. Control principle: 1. The influence of natural changes and non-therapeutic factors on the observation results can be excluded or controlled by control.
2. Experimental errors can be eliminated or reduced by comparison.
Second, the principle of balance:
Also known as the uniformity principle. In other words, experiment and control must abide by the principle of balance. In the experiment, subjects are required to observe some experimental factors, and all other conditions should be the same as much as possible. The conditions of non-experimental factors in each treatment group are required to be balanced and consistent to eliminate the influence of non-experimental factors on the experiment.
Third, the principle of randomness: 1, balance between groups.
2. The sample represents the whole population.
3. Various statistical analysis methods are based on randomization.
Fourth, the principle of repetition: repetition is another important means to eliminate the influence of non-processing factors. The degree of repetition is reflected in the number of experimental cases and the number of repetitions. The more experimental cases or repetitions, the more objective and true situation of opportunity variation can be reflected.