Sampling Techniques,Ppt

7/31/2019 Sampling Techniques,Ppt

1/26


2/26

A research population is generally a large collection of

individuals or objects that is the main focus of ascientific query.

Sample is a group of people, things, or places where data

are collected.

Sample is a part which represents a population

(A sample is simply a subset of the population.)


3/26

Sample

Thenumber

Themethods


4/26

The minimum number of sample:

Population

< 100

101-500 501-1000

>1000

Sample

50%

30-50% 20-30%

15-20%


5/26

Simple

RandomSampling

Systematic

randomSampling

StratifiedRandomsampling


6/26

1List all names of population

2

Write all names in small pieces of papers, thenroll them

3

Put the rolled papers into a box and shake the

box, so the rolled papers will be mixed.

4

Take the rolled papers one by one until you getthe number of needed sample

A. Raffling


7/26


8/26

1 List all names of population

2

Prepare the table of random numbers

3 Close your eyes and tick one number using pencil.

4

Check whether the number you choose is available in thelist of population.

5

Move your pencil up/down/left/right until you get thenumber of needed sample


9/26


10/26

1List all names of population

2

Determine the interval by dividing thepopulation with the sample needed

3

Close your eyes and choose one number asyour first sample

4

Determine the next samples based on theinterval which has been counted before.

It is used when the population is arrangedsystematically


11/26

It is used when there are stratification in the

population

e.g. Population of the stratification in senior high

school: grade X, XI, and XII

*grade X, XI, and XII are called

Subpopulation

There are two types of stratified random sampling

A. Proportional Stratified Sampling

B. Disproportional Stratified Sampling


12/26

A. Proportional Stratified Sampling

In determining the sample, we shouldconcern the ratio of each stratum

B. Disproportional Stratified Sampling

In determining the sample, we notconcern about the difference of thenumber of subpopulations member


13/26

In a senior high school, there are 222 studentsof grade X,333 students of grade IX, and444 students of grade IIX, so the total

number of students is 999.

Sample: 30% x 999 = 298


14/26

a. Proportional :

Grade X : 30% x 222 = 67 students

Grade IX : 30% x 333 = 100 studentsGrade IIX : 30% x 444 = 133 students

b. Disproportional :

298/ 3 class = 99 students of each grade


15/26

Random Sampling

- Define the population

- Select the sample

The benefit of random sampling:

It limits the probability that you choose abiased sample.


16/26

Defining the population

- It refers to the establishment ofboundary conditions that specify whoshall be included in or excluded from the

population


17/26

Establishing specification for stratifiedrandom sampling

The use of stratified random sampling will permit

you to include parameters of special interest and

to control for internal validity in terms of selectionfactors trough the use of moderatororcontrol

variable.

Stratification represents a good operationalstrategy forscreening membersof the population

into and out of the study and forreducing the

variability of the sample.


18/26

25%

75%

Two-year colleges in

the U.S.A

Private

Public

10%

15%

15%60%

Two-year colleges in the U.S.A

private, rural

private, urban

Public, rural

public, urban

two-year colleges in the

U.S.A

two- yearcolleges in the

U.S.A


19/26

1%

9%

5%

10%

3%

12%48%

12%

two-year colleges in the U.S.A

private, rural, large

private, rural, small

private, urban, largeprivate, urban, small

public, rural, large

public, rural, small

public, urban, large

public, urban, small


20/26

REMEMBER:

In stratified random sampling, you have more than one subpopulation.

Each subpopulation or stratum have random basis, however all should

be represented in the sample.

E.g. If in the population 65% are male, so in the sample taken should

be 65% male.

Each stratification parameter represent a control variable, that is, a

potential source of error or extraneous influence that may provide an

alternative explanation for the outcome of study.


21/26

Determining sample size

How large a sample should I employ??

Use as small a sample as possible for reason of time and

cost, while keeping it large enough to ensure its

representativeness.


22/26

Representativeness of the sample can be established at an

acceptable level of probability/ confidence level(z).

z is usually set between 90 99:

90 = 90% chance or representativeness (0.10 level) z = 1.65

95 = 95% chance or representativeness (0.5 level) z = 1.9699 = 99% chance or representativeness (0.01 level) z = 2.58


23/26

It is desirable to minimize sampling error in order to maximize

sample representativeness. So that, a researcher should

maintain the same proportion in stratified sampling.

e.g. 50% male and 50% female in the population should be

represented by 50% male and 50% female in the sample.


24/26

Formula:

N= (z/e)2 (p) (1

p)

N = sample size

z = confidence levele = proportion of sampling error in a given situation

p = the estimated proportion


25/26

E. g.

How to count the sample size of Private two-year colleges

account for 25 percent of all two-year colleges, when you

want 95% of confident level (z=1.96) with a tolerable amount

of error no greater than plus minus 10%.

Answer:

N= (z/e)2 (p) (1 p)

N= (1.96/0.10) 2 (0,25) (1 0,25)

N= (19,6) 2 (0,25) (0,75)

N= (384,16) (0,1875)

N= 72,03 => 72


26/26

Thank You

Sampling Techniques,Ppt

Documents