Statistical analysis
Statistics are one of the most important tools for a scientist. They allow scientists to test hypotheses. Perhaps you would like to know if a treatment makes your favorite organism grow better. Statistical analysis would let you determine the probability that the treatment increases growth. The below video shows you how to perform statistical analysis (t-test) in Excel.
Choice of t-test (paired vs two sample equal variance vs two sample unequal variance)
Ideally statistical analysis should be planned before the experiment is setup. Paired t-test has more statistical power than the other two types of t-test because it helps minimize the effect of nuisance factors that confound the experiment results. An example of how to setup an experiment to take advantage of paired t-test in plants is as follows. Control and treatment for each replicate would be planted side by side so that the replicate was in a block. Doing this minimizes the effect that different growing positions have on each replicate. A paired t-test takes this into consideration and assumes while the absolute measurement values might change for different replicates, the mean difference between control and treatment for each replicate should be conserved. The best part of using a paired t-test or other blocking strategy, is that replicates do not need to be completed at the same time. This way there is no real limit to the number of replicates for an experiment.