EBVM Flashcards
this is when ALL animals in a population are examined to find something out about the population
Census
this is when SOME of the animals in a population are examined to find something out about the population
sample survey
this refers to the entire group of interest in a survey (‘all cows in UK’)
target population
this refers to the animals we can (if we wish) include in a survey
study population
this refers to the ‘thing’ we are surveying
sampling unit
this is the list of all the animals in a study population
sampling frame
this is a sample where we choose which individuals to include
non-probability sampling
this is when we choose individuals for a survey because it is convenient (biased - not a goof representation of the population)
convenience sampling
this is when we choose individuals for a survey, but try to represent the population (good theory but still biased)
purposive sampling
this gives a deliberately unbiased and random sample
purposive sampling
this is when a sampling frame is made and a random subset is selected
simple random sampling
this is when there is no list of all individuals in a sampling frame available, so you create clusters, randomly select some clusters, then create a list of animals in the cluster and randomly sample some of them (less precise)
cluster sampling
this is when characteristics of a population that might be important are identified, then random samples are taken from the groups with the characteristic in proportion to their occurrence in the population (more precise)
stratified sampling
this refers to the ability of a survey to detect effects/relationships present
power
this is a range of values within which we are confident that the true population value lies (usual at a 95% level)
confidence interval
the sample size required for surveys demonstrating disease freedom depends on this
endemic prevalence
this scale of measurement is unordered categories (eye color, sex, religion, etc)
nominal
this scale of measurement is ordered categories (small -> medium -> large)
ordinal
this scale of measurement is counts of things, round numbers only
discrete
this scale of measurement is a numerical scale, fractions are allowed
continuous
this type of continuous data is where 0 is arbitrary (temp - 0C is 32F)
interval data
this type of continuous data is where 0 means nothing/none (weight)
ratio data
this is the chance of seeing the effect we saw in trial data if the null hypothesis was true
p-value
if the p-value is – the null must —
low, go