Hi everyone, I hope you are all fine! Unfortunately, I will not tell you what I’m doing at YNSECT because of the confidentiality. However, I wrote something that could interest you!
Today, I’m going to tell you how to design a good experiment. As scientists, we are all concerned by this topic. Thus, I think that it could be helpful to recall the good habits. When you are planning for an experiment, there are 4 things that you have to watch out:
Then, you have to be careful with the reproducibility of your experiment. When you are preparing samples, you have to make at least triplicates (if it’s really not possible, duplicate). It’s really important to have accurate results and to perform significant statistical analysis later. However, reproducibility doesn’t concerns only your own project. Your tests have to be reproducible by anyone. To do so, the protocol must be detailed. The material employed has to be clearly mentioned as well as the reactant used, and the settings of all the parameters. Something realy important is the control of your experiment. All the parameters implemented in your project have to be controlled (Temperature, pH, reactants concentration...).Basically, if you want to visualize the influence of a specific factor, vary this one and keep the others content. Also, don’t forget to set up negative and positive controls. Sometimes forgotten, this monitoring is really useful for response comparison. The final point to realize a good experiment is results treatment. To start, take into account all the errors incomes. When you are observing a parameters not correctly set, notice in the lab notebook. It could be an explanation for the results obtained. Then, you have to perform statistical analysis in order to give consistency to your results and to support your data. At least, standard deviation has to be calculated. If p-value, standard error of the mean or confidence intervals, is added, it will really give weight to your outcomes. To conclude, I really wanted to insist of the results interpretation. In fact, there is no bad results, and it’s not a good practices to discard the response that are not expected. What matters is to understand what you are doing. Although, it could be frustrating if results don’t match with your expectations, try to figure out what are the causes, and how to avoid it in the future. That’s it for this BlogPost! I hope that this text will be helpful and not seems too boring and repetitive, see you soon ;)! REFERENCES:
1 Comment
Cécile
2/12/2016 12:59:32 am
Hello!
Reply
Leave a Reply. |