Sampling theory

Question 1

Explain what the following quality control procedure are:

a. Control of equipment
b. Duplicate
c. Blank
d. Comparison of sampler

Answer 1

a. Send a sample with MilliQ to lab and compare with sample of MilliQ that has run through the equipment to see if there are any errors.
b. Take two samples of the same thing and see if the method is working well without any contaminations or errors.
c. Doing the sampling as you would but without the analyte of interest, for example using MilliQ water instead of lake water. You try out your method and see if there’s pollutants in any stage.
d. Compare samples taken by different samplers in lab to see if there are any errors.

Question 2

Describe one situation where knowledge about extreme concentrations is necessary.

Answer 2

When you want to see if the concentrations go above the legal “gränsvärden”. It could decide whether a company is legal or not. For example industrial company with a maximum level of pollutant release into recipient.

Question 3

Describe one situation where knowledge about average concentrations is necessary.

Answer 3

When you want to see if a piece of land is contaminated, you don’t want the concentration from just one spot then. When you want general knowledge of an area.

Question 4

What is the difference between single and composite samples? How are they generated?

Answer 4

A single sample is only one piece of material taken from the target.
A composite sample is several single targets put together.

Question 5

What are the pro and the cons for each of the ways regarding single and composite samples?

Answer 5

A single target has the pro that is shows spatial or time variations. The cons are that there is a risk of extreme values and that many single samples are required for you to get a true picture of reality.
A composite sample gives a more reliable estimation of the mean. It is also cost efficient. Cons are that it cannot be used for volatile subjects, that there’s an increased risk of cross-contamination and that trends and patterns are not visible.

Question 6

Explain the concept of sampling dimension. Which “dimension is the best?

Answer 6

The best dimension is called the 0-dimension, in which samples are in individual units (for example small heaps of material) and every unit has an equal chance of being selected for analysing. The 3-dimensional is the worst since this only allows you to do sampling of the outside.

Question 7

Why is the concept of sampling dimension a privileged situation? Use the principle of sampling theory to explain.

Answer 7

In reality there’s often a lot of stuff (roads or pipes for example) that make it difficult or impossible to use a low sampling dimension.

Question 8

Explain the advantages and drawbacks of using: Random or stratified (partially) stratified sampling

Answer 8

Random sampling would give areas that have no sampling at all and therefore important information will be missed. Partially random sampling is better because then you divide the area into sections and then randomize a number of samples in each section, this makes sure that the whole area at least gets covered.

Question 9

Explain the advantages and drawbacks of using:

Systematic sampling

Answer 9

Easy to do since samples are taken over a grid and/or on given time intervals. The drawback is that there could be issues with missing important patterns/trends. Gives a good overview of an area.

Question 10

Explain the advantages and drawbacks of using: Systematic random sampling

Answer 10

This is when you make a grid of the area and take random samples in the grid. You get the perk of not missing any patters/trends and it also covers the whole area. Has the highest accuracy of all sampling geometries.

Question 11

How much material is “really” analyzed when addressing VOC?

Answer 11

Around 5 g

Question 12

How much material is “really” analyzed when addressing metals?

Answer 12

Less than 1 g

Question 13

When is it not recommendable to prepare composite samples?

Answer 13

When you analyze VOC since they will evaporate.

Question 14

Describe the splitting methods/instruments called:

Coning and quartering

Answer 14

Make a pyramid –> flatten –> quartering –> make smaller pyramids –> flatten …

Question 15

Describe the splitting methods/instruments called: Fractional splitting

Answer 15

Separating the target into a number of smaller heaps. Fastest way to separate bigger heaps of material and gives a rather good result.

Question 16

Describe the splitting methods/instruments called: riffler splitter

Answer 16

A small device in which you can pour the target and it will divide between two smaller boxes. Then you can take one of those and pour in again to get an even smaller amount.

Question 17

Describe and give an example of a systematic error resp. a random error.

Answer 17

A systematic error is an error that affects every sample taken, it could for example be a wrongly calibrated instrument. A random error is a sample that shows an extremely low or high value, this happens due to the normal distribution phenomena. If you only take for example two single samples, one might show a very high concentration compared to the other one but in reality it is all part of the normal distibution curve. If more samples are taken you will get close to the mean value.

Question 18

What is the fundamental error? How can you address it?

Answer 18

Fundamental errors are due to taking too small samples, if that is the case then the sample won’t be representative of the target. It can be addressed by taking larger samples (or reducing particle size)

Question 19

What is the segregation error? How can you address it?

Answer 19

Segregation errors are due to the different properties of the different subjects in a target (for example density), which can lead to one subject being in the top layer and one in the bottom layer for example. This can be addressed by making sure to take samples through the whole target, with a pipe for example.