In this article, we will discuss the conduction of research. Through this discussion, the readers should know how to design the research schedule, conduct experiment and analyse data.
I highly recommend PHD student to learn statistics very seriously and carefully. We must keep in mind statistics is not only a mathematics tool but also the research design tool.
The tasks of research design includes: the problem definition, the experiment design, the data analysis. In particular, student should pay more attention to “confidence interval”, “hypothesis test”, “regression” concepts of statistics. They are useful both to experiment design and data analysis.
Design an experiment should answer the following questions:
- What purpose of the experiment.
- How many samples are enough to the experiment.
- How many times of the experiment should take.
All these questions can be answered by statistics. Please do not consider them as engineer questions that I made this mistake before. They are statistical questions.
Again, statistics is very important to data analysis. As I mentioned before, students need pay more attention to “confidence interval”, “hypothesis test”, “regression” concepts.
Firstly, you should know which variables are covered in the data. Are they numerical or categorical? What’s the relationship between the variables.
Secondly, you should know how to present the data according to their properties.
Thirdly, you should model the relationship of the variables. Here you can utilize regression concept.
Fourthly, you should verify your proposed model by using “hypothesis test”
Nowadays, the statistics data analysis do not need the research to calculate in details. Researchers can utilize extensive statistics software to realize them. These software include: Matlab, Minitab, SPSS and SAS, etc.