SCH 2304 Analytical Chemistry
Sampling
This is the process of obtaining from the population a portion in such a way that the portion is a true
representation of the population to a given confidence level. If sampling is wrong, then the analytical
process is useless. Sampling is obtaining a small representative sample.
Objectives of sampling
- Economic; the value of material
- Legal; assure consumer material is of a certain quality or product safety
- Safety; exposure to analyst, eg radioactive material
The sample picked should relate to the whole consignment
Sampling scheme
Questions to ask:
- What is the accuracy required?
- How to take the sample?
- Size of the primary sample?
- Sample storage and treatment
Accuracy – the greater the accuracy, the higher the cost (generally)
How to obtain the sample
- Random sampling – ensures that all parts of the sample have equal chance of being
sampled. If on a conveyor belt, use random numbers. Random sampling is very difficult to
achieve practically.
- Systematic sampling – regular predetermined intervals. Caution: other aspects could be
changing with the sampling interval
Size of sampling increment.
Size of sample. Consider sampling tool and particle size. The sampling tool should collect the largest
particles.
Sampling error = sample mean (X) – population mean (u)
X–U
Sample SD = σ /n1/2 where σ is the real sd of the population and n is the number of samples
Sampling error (at a given conf level) Es =± ts/ n1/2 where s is the sd of the experiment.
You decide the error you can tolerate and then calculate n.
Sampling for specification acceptance
Specifications are limits for tolerance
You have to decide:
Sampling
This is the process of obtaining from the population a portion in such a way that the portion is a true
representation of the population to a given confidence level. If sampling is wrong, then the analytical
process is useless. Sampling is obtaining a small representative sample.
Objectives of sampling
- Economic; the value of material
- Legal; assure consumer material is of a certain quality or product safety
- Safety; exposure to analyst, eg radioactive material
The sample picked should relate to the whole consignment
Sampling scheme
Questions to ask:
- What is the accuracy required?
- How to take the sample?
- Size of the primary sample?
- Sample storage and treatment
Accuracy – the greater the accuracy, the higher the cost (generally)
How to obtain the sample
- Random sampling – ensures that all parts of the sample have equal chance of being
sampled. If on a conveyor belt, use random numbers. Random sampling is very difficult to
achieve practically.
- Systematic sampling – regular predetermined intervals. Caution: other aspects could be
changing with the sampling interval
Size of sampling increment.
Size of sample. Consider sampling tool and particle size. The sampling tool should collect the largest
particles.
Sampling error = sample mean (X) – population mean (u)
X–U
Sample SD = σ /n1/2 where σ is the real sd of the population and n is the number of samples
Sampling error (at a given conf level) Es =± ts/ n1/2 where s is the sd of the experiment.
You decide the error you can tolerate and then calculate n.
Sampling for specification acceptance
Specifications are limits for tolerance
You have to decide:
