Sampling techniques 2

MATH 100SAMPLING TECHNIQUES

THE BASIC REASONS FOR THE USE OF SAMPLES

1. A sample allows us to obtain information with greater speed.

2. A sample allows us to obtain information with reduced cost.

3. A sample allows us to obtain information with greater accuracy.

PROBABILITY SAMPLING and NON-PROBABILITY SAMPLINGPROBABILITY SAMPLING

NON-PROBABILITY SAMPLING

• Any method of sampling that uses some form of random selection.

• Sets up some procedure that assures that diff. units in the population have equal probabilities of being included.

• Does not involve random selection

• Probabilities of selection are not specified for the individuals in the population.

NON – PROBABILITY METHODS OF SELECTION•QUOTA SAMPLING

•SNOW BALL SAMPLING

•CHUNK SAMPLING

•JUDGEMENT SAMPLING

#1 QUOTA SAMPLING

- Identify samples and stop as soon as completed.

#2 JUDGEMENT SAMPLING- Pick students who are in the first row

only.

- Pick students who are handsome or beautiful.

- Pick students who have long hair.

#3 SNOW BALL SAMPLING

- Start with “1” , then “1” will pick the next and so on.

#4 CHUNK SAMPLING

- Consider one area only.

METHODS OF PROBABILITY SAMPLING

SIMPLE RANDOM SAMPLING

- simplest form of random sampling - consists of choosing a sample from a set

of all possible samples of a pre-chosen size, giving each sample an equally chance of being the selected one.

Diff. techniques of drawing a sample in a simple random sampling:•LOTTERY OR FISHBOWL TECHNIQUE

- simply writing the names or numbers of all the individual members in a small-rolled piece of papers, then placed in a container.

•TABLE OF RANDOM NUMBERS- these tables show sets of random digits

arranged in groups.a. Direct Selection methodb. The remainder method

Illustration:

•Suppose a researcher wanted to draw a random sample of 8 students from a list of 24 fourth year students in research class.

Using the DIRECT SELECTION METHOD:1. Make a list of the students and give them

corresponding serial numbers from 01 to 24.

2. Pick a line number and a column.3. Read downward and record the next

digits4. Repeat step 4. Ignore repeated numbers

and the numbers greater than 24.5. If the column has been exhausted,

proceed to the next.6. Select the n individuals.

Using the REMAINDER METHOD- Used whenever the direct selection

method cannot be applied.

HOW TO CONDUCT THE METHOD:-the number taken from the Table of random numbers is subtracted from the upper limit within which this number falls, the remainder is the sample.

Example

•Using the first column of the table of random numbers, pick 10 sample units from a population of 123.

•Using the first column, pick 100 sample units from a population of 1150.

SYSTEMATIC RANDOM SAMPLING- Involves the selection of the desired sample

in a list by arranging them systematically.- in this method, a population is divided into

n groups of k individuals each. (every kth individual is selected)

- Method of selecting a sample by taking every kth unit in an ordered population where only the first unit being selected in random and the rest will be selected systematically.

sizesamplen

sizepopulationN

ervalsamplingkwheren

Nk

kforformula

int

:

ILLUSTRATION:- Suppose that the researcher desired to

draw five samples of household heads member in a row housing units consisting of 20 units in a given area for his study.

Steps:

1. Arrange the housing units in order from 1 to 20.

2. Determine the sampling interval.

3. Select a number at random from the interval.

4. The next housing unit is selected every kth unit thereafter.

STRATIFIED RANDOM SAMPLINGSTRATIFICATION- Method in which the population size N is

divided into a number L, of non-overlapping, so that the samples within the stratum are more or less homogeneous and samples among strata are most heterogeneous.

Illustration:- Suppose that the researcher wanted to

draw a sample(4 students from each stratum) of 16 students, 17 sophomores, 13 juniors and 10 senior college students.

Steps

1. Stratify the students strata.

2. Select random samples from each stratum.

CLUSTER SAMPLING

-used when the population is very large and widely spread out over a wide range of geographical area.

- The population is divided into M clusters, which may not be of same size.