Lecture+3+ +Principles+of+Sampling
Lecture+3+ +Principles+of+Sampling
PRINCIPLES OF SAMPLING
OUTLINE
Definitions Advantages and Disadvantages of Sampling III. Reasons for Sampling IV. Where Does Sampling Occur? V. Goals When Sampling VI. Types of Quality Characteristics VII. Selecting the Sample VIII.Sampling Tools
I. II.
I.
DEFINITIONS
I.
D. Unit smallest quantity being inspected a representative sample of a fresh commodity, an ingredient or a single package of food E. Sampling A method for obtaining information from a small portion of a lot when it is too expensive, time consuming, impossible/preferred to measuring the total population
DEFINITIONS (cont.)
I.
DEFINITIONS (cont.)
F. Inspection measuring, examining or testing a sample of ingredients or product units to determine if items are acceptable and meet specifications.
B. Choose samples that represent the population minimize error that will naturally occur due to the fact that only a portion of the population is used to estimate population parameters. Representatives samples depend on: 1. Size of lot 2. Purpose of control 3. Product variability 4. Regulation
Attributes
Quality characteristics of either a product or process Which may be tabulated as either: -present or absent -satisfactory or non-satisfactory - within limit or outside limit
Variables - Can be measured and expressed in increments of inches, mg, degrees, %, etc.
- Can be measured by use of some scales which can theoretically be divided into infinite sub-units
Attributes Vs Variables - Attribute characteristics are tabulated while Variables are measured - Tabulating attributes as the basis for quality control is generally less expensive than measuring variables BUT - Attributes require many more observations to obtain the same information than can be obtained from a small number of measurements of variables
-chipped off
- dirty
B. VARIABLES based on : sample statistic of average on standard deviation and the type of frequency distribution e.g - thickness of bottles, cans - weight of bottles, cans - headspace * will effect fill weight of products.
VII.
A.
Non-random sampling bias; disturb product acceptance decisions Sampling from same location Selecting only those products that appear defective (or non defective) Ignoring portions of lot inconvenient to sample Deciding on the pattern of stratification with limited knowledge
a. b. c.
d.
VII.
B.
1.
Random sampling
Simple random sample
VII.
i.
K = N/n = # product unit in lot / # samples desired ii. Pick every kth product
VII.
Example : You have 2000 1-gallon containers of icecream and you need 20 samples K = 2000/20 = 100, or each 100 gallons is a sub lot Therefore, if the first gallon chosen (randomly) off the line was the 14th product from the first sub lot, then the 14th product from each of the remaining 19 sub lots must be chosen.
VII.
3. Stratified random sample Used when the lots are known to come from different machines, production shifts, operators, etc
4. Composite sample Combination of two or more random samples from a uniform flow of solid or liquid
VII.
5. Skip lot sampling Used to greatly reduce the amount of sampling when desired quality is uniform and items are guaranteed as with ingredients or finished foods manufactured the TQM way.
VIII.
SAMPLING TOOLS
VIII.
C. Rifle or Divider A vibrating, subdividing sampling system with numerous sample pockets for use with a free flowing particulate powder.