Null hypothesis and Inferential stats Flashcards
What should be considered for descriptive stats?
- decide measure of central tendency - mean/median
- decide measure of variability is most appropriate, SD, SEM, Confidence interval, ranfe or IQR
What should you do when reporting IQR?
- put in parentheses the lower and upper quartiles
- don’t forget to include the units for any measurements if appropriate
What are the steps for null hyp stats testing?
1.) Establish your null hypothesis and alternative hypothesis
2.decide on a critical significance level a (0.05)
3.choose your test, calculate your test stats (t stat, U stat etc) and therefore your P value
4.accept or reject the null hypothesis
How would you write H0?
- There is no difference in [what you are measuring] between [the categories]
How to calculate your P value
A. Decide what sort of data you have (nominal, ordinal, ratio)
B. Decide whether you’re looking for differences or relationships
C. Work out whether you satisfy certain assumptions
D. Identify the appropriate test
E. Calculate the P value- decide if it less than the critical significance level
How does experimental design influence statistical analysis?
a.) stats power
b.) types of replicates - biological and technical
c.) related or unrelated
What do technical replicates show?
- they show variation due to measurement error, random variation
What do biological replicates show?
- Show true biological variation and this is often quite signifcant and relevant
What are examples of replicates?
- UPM = urban particulate matter
- Microglial-like cultured cells loaded with fluorescent dye to measure mitochondrial membrane potential
- CTCF= corrected total cell fluorescence
What are examples of biological replicates?
- means of 3 technical replicates
- data are expressed as mean +/- SD
How do you create good graphs and figures?
- clearly state in the figure legend:
- sample size
- P value
- summary statistic (i.e. mean, standard deviation or median and IQR)
