Introduction to Food

Analysis:
Sampling
FST 507
Spring 2013

Why Do Analysis?

Who is it for?
Research (Your Research/)
Product Development
Quality Assurance/Process Control
Food Safety (FSMA)
Food Contaminants
Toxicology
Food Defense

Health

Nutrition

Types of Analysis

Qualitative - provides information on the

chemical identity of the analytes in the
sample

Yes/no
Good/Bad
Acceptable/Not Acceptable
Pass/Fail

Quantitative - numerical information on

the relative amounts of one or more of
analytes present.
Values (Numeric/Significant Figures)

Semi-Quantitative

Pass/Fail set on a level of detection or level of

safety

Analysis

Wet Chemistry
Gravimetric Analysis (weight)
Stoichiometric Analysis (molar equivalents)

Analytical Chemistry
Instrumentation

Rheology
Physical Chemistry
Density, Viscosity, Particle
Size/Distribution, Texture

Sample Selection

Determine the characteristics of a (large) quantity of food

material to be sampled:

To analyze every part of the material and obtain an accurate

measure of the property of interest, but in most cases this
is practically impossible. (Why?)

the contents of a truck arriving at a factory

a days worth of production (lot)
products stored in a warehouse

Most analytical techniques destroy the sample.

analytical techniques are time consuming
expensive or labor intensive, costly

It is therefore normal practice to select a fraction of the

whole material for analysis, and to assume that its
properties are representative of the whole material.
Selection of an appropriate fraction of the original (whole)
material is one of the most important stages of food
analysis and can lead to large errors when not carried out
correctly.

Population, Sample, Laboratory

Sample

Population: The whole of the material whose

properties we are trying to obtain an estimate of is
usually referred to as the population.
Define the population: location, date,

Sample: A fraction of the population is usually

selected for analysis, which is referred to as the
sample. The sample may be comprised of one or
more sub-samples selected from different regions
within the population.
Laboratory Sample: The sample may be too large
to conveniently analyze using a laboratory procedure
and so only a fraction of it is actually used in the final
laboratory analysis.

Sampling Objective

To ensure that the properties of the laboratory

sample are representative of the properties of the
population, otherwise erroneous results will be
obtained.
Selection of a limited number of samples for
analysis is of great benefit because it allows a
reduction in time, expense and personnel required
to carry out the analytical procedure, while still
providing useful information about the properties
of the population.
Always be aware that analysis of a limited number
of samples can only give an estimate of the true
value of the whole population.

Sample Plan

To ensure that the estimated value obtained from the

laboratory sample is a good representation of the true
value of the population it is necessary to develop a
sampling plan.
A sampling plan should be a clearly written document that
contains precise details that an analyst uses to decide the:

sample size
the locations
the method used to collect the sample
the method used to preserve them prior to analysis

Documentation of procedures carried out during the

sampling process (ie. An SOP or a laboratory notebook)
For certain products and types of populations sampling
plans have already been developed and documented by
various organizations which authorize official methods
the Association of Official Analytical Chemists (AOAC).

Sample Types

Official Samples: Legal (Claims,

Safety/Sampling Plan Required)
Raw Materials: Acceptable
Process Control Samples: To
ensure we are in control of the
process (In process samples)/QA
Finished Product: Labeling/Claims
R&D Samples: Variables/Stability

Measured Property

Attribute: A property that a product either does or does not

have. (examples)
Foreign material in the sample (eg. Metal or glass)
Property of a Material (yellow color or a banana; the activity
measurement of enzyme)

Variable: some property that can be measured on a

continuous scale
The fat content in a sample/the color of a sample

Seriousness of the Sampling Plan: If the laboratory sample

does not represent those of the population, then the
seriousness of the outcome, if a mistake is made in the
sampling, is much greater than if the property measured is a
quality parameter. Consequently, the sampling plan has to be
much more rigorous for detection of a potentially harmful
substances than for quantification of quality parameters.

Nature of the Population

Finite: Definite Size (Truckload of Oil)

finite population usually provides information

about the properties of the population.

Infinite: No Definite Size/Change.

(conveyor belt of material)

infinite population usually provides information

about the properties of the process.

To facilitate the development of a sampling

plan it is usually convenient to divide an
"infinite" population into a number of finite
populations, e.g., all the products
produced by one shift of workers (Lot), or
all the samples produced in one day
(Production Code).

Nature of the Population

Continuous: one in which there is no

physical separation between the different
parts of the sample
Liquid Milk

Compartmentalized: one that is split

into a number of separate sub-units
Boxes of product in truck (sub-units)
The number and size of the individual subunits determines the choice of a particular
sampling plan

Nature of the Population

Homogenous: one in which the properties of the

individual samples are the same at every location
within the material
Perfect (1 sample; ie: water?)

Heterogeneous: one in which the properties of

the individual samples vary with location.
In practice, most populations are heterogeneous
and we must carefully select a number of
individual samples from different locations within
the population to obtain an indication of the
properties of the total population.

Test Method

The test method will also be a factor in the

sampling
Speed of Analysis
Precision, Accuracy, Specificity
Cost per analysis
Everyone wants is cheap, fast and accurate!

Technique is destructive or non-destructive.

It is convenient and advantageous to analyze the
properties of many samples if the analytical
technique used is capable of rapid, low cost,
nondestructive and accurate measurements.

Problems with Sampling

Sampling Bias
Poor Sample Storage
Contamination
Stability
Mislabeling of samples

Sampling

Actual Sampling
A minimum (correct and needed) number of subsamples that needs to be analyzed to obtain an
accurate representation of the population
sequential sampling: sub-samples selected from
the population are examined sequentially until
the results are sufficiently definite from a
statistical viewpoint. For example, sub-samples
are analyzed until the ratio of good ones to bad
ones falls within some statistically predefined
value that enables one to confidently reject or
accept the population.

Sampling/Location

Location
In homogeneous populations it does
not matter where the sample is
taken from because all the subsamples have the same properties.
In heterogeneous populations the
location from which the sub-samples
are selected is extremely important.

Sampling/Location

Random Sampling
from any location
avoids human bias/Statically based

Systematic Sampling - drawn

systematically with location or time
Judgment Sampling - drawn from the
whole population using the judgment
and experience of the analyst

Sample Preparation

Preparing a heterogeneous sample

inter-unit variation differences from unit
to unit
Ex: Box of oranges. Some good and some bad

intra-unit variation difference within the

unit itself
Ex. Nutritional value of chicken. The skin is
different than the muscle, and the bones have
little to no nutritional value.

Sample Preparation

Making a sample homogenous from a

heterogeneous sample
Blender
Grinder
Mortar & Pestle

Methods for reducing solid or semiDry foods- Grind

powder, then mix
solidto
foods
thoroughly with a spoon or spatula.

Moist solid foods- (meat products)Homogenize by chopping rather than

mincing
Oils and Fats- (cheese, butter,
margarine)-Prepared by gentle
warming and mixing

Preventing changes in Sample

Microbial Growth and Contamination.
Lipid Oxidation Protection.
Physical Changes.
Chemical Changes.
Enzymatic Inactivation

Refrigerate/Freeze/Room Temperature
Stability

Sample Preparation

Sample Identification

Laboratory samples should always be labeled carefully so that if

any problem develops its origin can easily be identified.

Information used to identify

a) Sample description
b) Time sample was taken
c) Location sample was taken from
d) Person who took the sample
e) Method used to select the sample.
The analyst/laboratory should always keep a detailed
notebook clearly documenting the sample selection and
preparation procedures performed and recording the results of any
analytical procedures carried out on each sample. Each sample
should be marked with a code on its label that can be correlated to
the notebook. Thus if any problem arises, it can easily be
identified.
GLP/GMP (Good Laboratory Practices/Good Manufacturing Practices.

Sampling/Statistically Based

How do you know you have enough

sample(s) to determine if the result(s) is
(are) correct?
The sampling is based on the data
acquired.
Normal Distribution
Confidence Interval (Confidence Level)
Outliers (Dixons Q-test)
Run Chart

Assumption

At this point, our method is good.

One sample analyzed multiple times
only tells us how good our method or
analyst reproduced the results
waste of time.
Multiple samples analyzed will lead
us to whether our result represents
the population. (We have some work
to do! How much work?)

Normal Distribution

Normal Distribution

Standard Deviation
1 = 68% of all data
2 = 95% of all data
3 = 99% of all data

Central Tendency

The mean ( x ) is the central point

(line) of our distribution.

Standard Deviation (s, )

N = number of samples
xi = each individual result
Standard deviation is also known as
the variance.

Coefficient of Variation

Same as %RSD (percent relative standard

deviation)

This is how close the data is in context to the

mean. It is good information but does not
answer the question whether we have sampled
correctly.
But, a rule of thumb for the food industry, 5%
RSD or less is showing good data. The FDA and
USDA allow up to 20% CV on Package Label
Claims (Facts Panel).

Confidence Level (Interval)

Confidence Level/Confidence
Interval/Confidence Testing
CL/CI
90%, 95%, 99%, 99.7% Confidence
How is this applied?

Confidence Assumptions

Again, we assume the test method is

giving correct results.
There is usually a minimal amount of
samples needed and there is usually
a maximum amount of sample that if
one more sample is taken it wont
affect the confidence.

Confidence Test

For a small population (<30 Samples)

n = number of samples
Degrees of Freedom (df): for n samples
s = standard deviation
t = /2 (two sided test)
Are the values within the range of the from the mean.
This calculation sets the deviation at some CI (ie. 90%, 95%,
99%).

What is 1 CI (as a decimal fraction)

/2
90% CI1-0.90 = 0.10
/2 = 0.10/2 = 0.05

95% CI1-0.95 = 0.05

/2 = 0.05/2 = 0.025

99% CI1-0.99 = 0.01

/2 = 0.01/2 = 0.005

99.9% CI1-0.999 = 0.001

/2 = 0.001/2 = 0.0005

t table

t table (continued)

Sources of Errors

Systematic (determinate) Error

A pipette that consistently delivers the wrong
volume
A mechanical error

Random (indeterminate) Error

Judging the endpoint change in a titration

Gross Error (blunder)

Adding wrong reagent

Significant Figures
1.

Zeros after a decimal point are significant

figures.
64.720 and 64.700 have 5 sig figs

2.

Zeros before a decimal point with no other

preceding digits are not significant.
0.6472 has 4 sig figs
0.000456 has 3 sig figs

Significant Figures
3.

Zeros after a decimal point are not significant if

there are no digits before the decimal point.
0.0069 has 2 sig figs
1.0059 has 5 sig figs

4. Trailing zeros (those at the end) are significant

only if the number contains a decimal point.
5.640 has 4 sig figs
120000 has 2 sig figs

Significant Figures
Rules for addition/subtraction problems
Find the quantity with the fewest digits to the right
of the decimal point.
5.

7.939 + 6.26 +11.1 = 25.299

Round answer to 1 sig fig or 25.3.

Rules for multiplication/division problems

Find the quantity with the fewest digits.
6.

(27.2 x 16.63) 1.846 = 230.3011918

Round answer to 3 sig fig or 230.

Precision and Accuracy

Accuracy: This is how close the data

is to the target value.
Precision: This is how close the data
is to each other.

Good Precision/Good Accuracy

Poor Precision/Poor Accuracy

Outliers (How do you deal with them?

Outliers

Are the values we calculated outliers and can we

eliminate them before we do Confidence Interval
(CI)?
Q (calculated) [Dixon Q Value]

Q = |Suspect Value Nearest Value|/|Largest Value Smallest Value|

Q table (Q critical at P)
n (n = number of samples)
If Qcalc < Q critical, then good value

Critical Q Values

Q Critical Table
Table of critical values of Q

Statistical Sampling Plan

From this information, were we able

to come up with a number of
samples to collect before we did the
analysis?
No..but
If our Standard Deviation is
acceptable (low %RSD) then our
data will start to be within the range
of our confidence interval (CI)

Other Sampling Techniques

Universal Mean/Student T-Test

Statistical Process Control
2 Std. Dev. Of the Mean (95%
Confidence)
Military Testing
ICMSF (microbiological sampling
See Dr. Cole)

Student T-Test

= mean (x) in a small population

is universal mean

T is calculate and compared against a table

of T values at a specific confidence
probability (P=95%,) and the degrees of
freedom (n 1)
Has this helped us determine sample size?
Maybe! Does T pass?
Important: Need your data and the universe
(2 sets of data)

Statistical Process Control

SPE/Control Chart
Average, 1 and 2
standard deviations
2 Standard deviations is
95% of the Normal
Distribution (UCL, LCL)

xp

0.800

1.28

0.900

1.64

0.950

1.96

0.990

2.58

0.995

2.81

0.999

3.29

Large Populations (Z value)

Military Sampling Plan

MIL Std 105: Sampling for attributes

AQL: Acceptable Quality Level
LTPD: Lot Tolerance Percent Defect
Enter Parameters: Lot Size (?)
Usually requires a lot of samples (becomes
expensive). Not Practical.
More of a Validation Tool.
Not a day to day sampling tool.
QA
Primarily a Microbiological Method/Plan