I have some experimental results with my samples having different mutants and different treatment. I was performing student t test for analyzing whether the data is significant or not. I want to know that is it necessary to perform Two Way ANOVA with these results or student t test will do? Is there any software apart from Microsoft Excel which can help me better understand the statistics as I cannot perform Post Hoc correction in Excel.
Generally, it would not be wrong if you perform a two-way anova on all data, followed by classical t-tests between different treatments within the two experimental groups (=different mutants, if I understood you correctly). First, the two-way anova would tell you if there is a significant effect of the treatments, mutation or their interaction. Then you can either do post-tests (e.g. Tukey, Bonferroni...) or simply do t-tests in Excel. (Two-way) anova is a very robust method, taking in to consideration the variation of all data from all experimental groups, and it does not analyze within group significance (if, of course, you don't perform post-tests).
I would perform anova on data that is contextually linked to each other (e.g. different amounts of the same stressor on the same kind of organism, tested against the unstressed control), but would prefer doing t-tests between unconnected groups of data (e.g. treatment A, treatment B, treatment n, versus the untreated control).
Generally, it would not be wrong if you perform a two-way anova on all data, followed by classical t-tests between different treatments within the two experimental groups (=different mutants, if I understood you correctly). First, the two-way anova would tell you if there is a significant effect of the treatments, mutation or their interaction. Then you can either do post-tests (e.g. Tukey, Bonferroni...) or simply do t-tests in Excel. (Two-way) anova is a very robust method, taking in to consideration the variation of all data from all experimental groups, and it does not analyze within group significance (if, of course, you don't perform post-tests).
I would perform anova on data that is contextually linked to each other (e.g. different amounts of the same stressor on the same kind of organism, tested against the unstressed control), but would prefer doing t-tests between unconnected groups of data (e.g. treatment A, treatment B, treatment n, versus the untreated control).
A nice and user friendly statistical program I use is GraphPad Prism: http://www.graphpad.com/scientific-software/prism/. It's got a great help file, written by professional statisticians, from which you can learn a lot!
Thanks jaka your response was very informative. I will try graph pad prism.
Yes graph pad prism statistical program is easy to use and understand the results. I,m using it for statistical analysis.
Regarding your investigations, I guess that it is the better way to perform a two factor variance analysis in order to find out the interactions between your mutants and your different treatments.
t-test is not appropriate in your trial. Yes you should do an ANOVA. There are various software packages available the most common ones being SAS and R.
- Badges
- Science topic
- Similar topics
- Biological Science
- Botany
- Plant Biology
- Plant Biotechnology