What statistical checks and test should I do for blood tests (coursework)
Check for normal distrobution by seeing if the mode, median, and means are all similar
Parametric tests when:
Normally distrobuted results
Equality (homogenity) of varience
Linearity
Independence
If any of these are not correct don’t use parametric instead use non-paraetric or use methods to mainipulate data to be parametric.
Parametric tests assume they apply to the entire population whereas non-parametric don’t
P values is proability/likelyhood of obtaining an observed result if the null hypothesis is true (basically the chance that the results were insignificant)
The conventiion is to use an alpha value of 0.05 which is the cutoff for significance.
How many variables are in the experiment should be taken into account and how many comparisons are being made.
The comparisons between variables made should be biologically meaningful (comparing Men with alzheimers and women without would have too many variables and not be biologically meaningful.
Normallity tests can be used to generate a P value to indicate if the data deviates from a noral distrobution.
Wether the data is independent also needs to be determined. for example seperate groups of mice (wild and mutant type) with no interactions can be seen as independant. If repeat tests on those mice are done and they learn more about the test since day one then the day one and 2 results from the same mouse are not independant. If tests aren’t independant then statistical tests that adjust for that fact will need to be used.
Use Jamovi/spss for tests (avoid excel)
What statistical errors should I be aware of for my coursework?
Type 1 Errors occur when the null hypothesis is rejected but it is actually true
Type 2 Errors occur when the null hypothesis is accepted but in reality it is significantly different
Type 2 errors are generally prefered.
What are techinical and biological replicates?
A biological replicate are between distinct samples that show individul variation and represent a whole population. Each sample has the same treatment condition and all the outputs are used in the stats test
A technical replicate is repeeated measurements of the same sample to determine veritation in measuring equipment and protocals. An average value per sample is used.
