Post on 02-Jan-2016
Sampling of Animal Sampling of Animal PopulationsPopulations
Learning Objectives:Learning Objectives: define & differentiate samplingdefine & differentiate sampling advantages/disadvantages advantages/disadvantages
sampling methodsampling method select sampling strategyselect sampling strategy
Data SourcesData Sources
For epidemiological For epidemiological analyses, must analyses, must be:be:
CompletenessCompleteness
ValidityValidity
RepresentativeneRepresentativenessss
Accuracy, Refinement, Precision, Accuracy, Refinement, Precision, Reliability and Validity Reliability and Validity (Thrusfield, 1986)(Thrusfield, 1986)
Accuracy: investigation or measurement Accuracy: investigation or measurement conforms to the truthconforms to the truth
Refinement: eq, 13 kg and 13,781 kg, both Refinement: eq, 13 kg and 13,781 kg, both represent accurate, but the second is more represent accurate, but the second is more refined than the first. Another eq., otitis refined than the first. Another eq., otitis externa, otitis externa by bacteria externa, otitis externa by bacteria
Precision: as a synonym of refinement and Precision: as a synonym of refinement and to indicate the concistency of a series to indicate the concistency of a series measurements (repeated sampling)measurements (repeated sampling)
Accuracy, Refinement……….Accuracy, Refinement……….
Reliability (= reproducibility): Reliability (= reproducibility): produces similar results when its produces similar results when its repeatedrepeated
Validity: measure what its supposed Validity: measure what its supposed to measure, its long term to measure, its long term characteristic of tehniquecharacteristic of tehnique
Data can be collected as Data can be collected as part ofpart of
Routine data collection: laboratory Routine data collection: laboratory submissions, disease surveillance submissions, disease surveillance programmes, industry/farm data programmes, industry/farm data recording systemrecording system
Structured data collection: regular Structured data collection: regular monitoring of disease/productionmonitoring of disease/production
Epidemiological studiesEpidemiological studies
Data Collection ProcessData Collection Process
Whole population interest Whole population interest (=cencus)(=cencus)
It can be restricted to a It can be restricted to a samplesample : : obtained more quickly, less obtained more quickly, less expensive to collect, more expensive to collect, more accurate, more efficientaccurate, more efficient
SamplesSamples
Probability Samples : random Probability Samples : random selectionsselections
Non-Probability Samples: a Non-Probability Samples: a convenience sample, a convenience sample, a purposive or judgmental purposive or judgmental samplesample
Probability SamplingProbability Sampling
Simple Random Simple Random SamplingSampling
Systematic SamplingSystematic Sampling Stratified SamplingStratified Sampling Cluster SamplingCluster Sampling Multistage SamplingMultistage Sampling
Simple Random Sampling
Each individual has an equal probability of selection
An individual’s selection doesn’t depend on others being selected, homogenous population
Disadvantage : may result large variation, thereby requiring larger sample sizes
Eq: flipping a coin, using random number tables
Ex: catlle on farm
Systematic Random Sampling
The n sampling units are selected from the sampling frame at the regular intervals
The starting point in the first interval is selected on a formal random basis
A practical way to obtain a representative sample
It ensures that sampling units are distributed evenly over the entire population
Stratified Random Sampling
The sampling frame is divided into strata, then a simple random or systematic random sample is selected within each stratum
To be effective at reducing variation, for example: milk production in population of dairy cows of the Jersey and Holstein breeds. Genetic differences affecting milk volume between the two breeds.
Definite strata, but homogenous within it
Cluster Sampling
Sampling is applied at an aggregated level (=group) of individual units
Groups or clusters such litters, pens, herds, artificial groupings (geographic areas)
Can be selected by simple, systematic, or stratified random methods
Groups with similar characteristics, but heterogenous within groups
Multistage Sampling
Similar to cluster except that sampling takes place at all stages/at different hiererchical levels og aggregated units of interest
Subsampling within the primary units (litters, pens, herds). A sampling of secondary units (e.q., animals) would be selected
Often used as part of epidemiological studies Ex: cattle in region, to be sampled to
determine TBC or mastitis prevalence
Sampling to detect Sampling to detect diseasedisease
Finite populations:Finite populations:
n = [1-(1-n = [1-(1-ββ) ] [ ( N- d/2) + ½]) ] [ ( N- d/2) + ½]
Infinite populations (> 1000)Infinite populations (> 1000)
n = [ log (1-n = [ log (1-ββ)] / [ log ( 1- d/N) ])] / [ log ( 1- d/N) ]