Introduction Flashcards
What is rule 1?
State your objectives before you start your project
What is rule 2?
Many questions cannot be answered due to biological hierarchy e.g. decreasing knowledge from molecular level to biosphere
What is rule 3?
Just because a variable is measurable does not make it relevant to your study
What is rule 4?
- Make an effort to collect sufficient data for proper statistical analysis e.g. number of replicates
- For monitoring studies = trade off between # of sample sites and # of replicates
- For experiments = trade off between # of factors and levels and # of replicates
What is rule 5?
Balance sampling and experimental design if there is no reason not to i.e. strive for equal # of replicates for each sample location or treatment
What is rule 6?
ALWAYS use controls in experimental studies (i.e. non treatment)
What is rule 7?
- Get a representative unbiased (random) sample of the study population
- allows generalization (through inferential statistics) to the population of interest
What is rule 8?
Conduct a preliminary survey
What is rule 9?
Adjust the sample unit size to fit the organism sampled e.g.nested plot design
What is rule 10?
Use stratified sampling (based on homogeneity of geography and density of organisms) or use blocked experimental design (based on physical gradients)
What is rule 11?
- Test assumptions of parametric statistical analysis
- if the requirements CANNOT be met then: transform the data, use non parametric statistical analysis or re-sampling technique (e.g. Monte Carlo)
What are the assumptions of parametric statistical analysis?
- Independence
- Homogeneity of Variance
- Normality
- No bias (randomness)
What is rule 12?
Use a variety of statistical tests until you get the significant result you need
What is rule 13?
Before starting decide on the number of significant figures to record (only applies when continuing an existing study)
How to determine the adequate amount of significant figures?
Using the Sokal and Rohlf Method
Example: Max value = 80 mm and Min value = 20 mm
- Determine the range = 60
- Divide the range by 300 and 30
ANSWER: Measure to nearest 0.2 or up to 2 mm
What is rule 14?
An ecological estimate needs measure of contributing errors e.g. mean +/- confidence interval
What is rule 15?
Critical value used on environmental science is usually 0.05 (alpha)
It is SUGGESTED boundary for rejecting null hypothesis
When is null hypothesis rejected or accepted (p-value)?
REJECTED when p-value LESS than 0.05
ACCEPTED when p-value GREATER than 0.05
What is rule 16?
Statistical significance does NOT always equal biological significance
What is rule 17?
Be aware and check mathematical model being applied
What is rule 18?
Garbage in, garbage out….sloppy work will get you every time when you decide to do a statistical analysis and decision errors can occur
What is a type 1 error?
Rejecting true NULL hypothesis (person not guilty but convicted)
What is a type 2 error?
Rejecting true ALTERNATE hypothesis (person is guilty but not convicted)