biostats Flashcards
What is a sampling error?
when the entire sample is not relevant to the entire population
Three examples of non-sampling errors
poor sampling method, measurement/questionnaire error, behavioral effects.
Statistical population
all the individuals for which we have an interest concerning some property
Sample
piece of a population
Census
data collected from the entire population
Descriptive statistics
values that we calculate based on the sample. ex: mean, mode, S.D, etc.
Inferential statistics
applying probability to sample values to understand something about the population. ex: 95% CI
Statistic vs. Parameter
statistic=sample
parameter=population
Accuracy
closeness of a measurement to its correct value
Precision
closeness of repeated measurements to one another
Bias
when a sampling method tends to produce high or low values relative to the correct value
Experimental Design
Goal is to answer the research question as clearly and as efficiently as possible.
Design must ensure that
1. the right data are collected
2. the sample size is sufficient (n=replicates used)
3. the highest power of the test possible (statistical test that can detect differences when there are differences)
Statistical bias
any instance that creates a difference between an expected value and the true value of a population characteristic being estimated, leading to inaccurate results.
Noise
natural variation or variability caused by the experimenter, or more simply put, any unexplained variability within a data sample
3 ways to decrease noise
- increase sample size
- standardize protocols
- control environmental factors
What is the difference between a replicate and a trial?
Replicate: an observation in a treatment. n = number of replicates per treatment. N = number of replicates in an experiment.
Trial: an exact repeat of an entire experiment
