27 November 2008

Teknik Pengambilan Sampel

Beberapa minggu yang lalu saya mempelajari tentang Sampling dari Prof. dr. Hakimi, SpOG, PhD. Berikut ini peta atau mind map dari apa yang saya pelajari tersebut. Double click untuk mendapatkan gambar yang lebih besar dan klik kanan untuk menyimpan. Semoga bermanfaat.

Ini adalah format outlinenya

SAMPLING

I. Terminology

A. method type

1. probability

a) use random selection

b) each unit have prob of being selected

2. Non probability

a) sistematic selection/haphazard

b) not random

B. group

= where can we go wrong

1. the theoretical population

2. the study population

= what population can u get acces to?

3. the sampling frame

= How can you get access to them?

4. the sample

= who is in your study

II. Probability

A. key concept

1. statistical terms in sampling

a) variable

b) statistic

c) parameter

2. sampling error

a)

b) standard error in normal curve

c) confidence level

d) confidence interval

3. the sampling distribution

= the dist of a statistic across an infinite number of samples

III. Design

A. probability

1. simple random

a) objective

select n from N, every NCn has equal chanc

b) procedures

table of random numbers

computer random generator

mechanical device

c) can sample with or without replacement

d) f=n/N is the sampling fraction

2. stratified sampling

a) proportional/quota random sampling

b) objective

pop divided into non overlapping data

then do simple random sampling in each strata

c) purpose

insure representation

administrative convenience

sampling problem may differ in each strata

increase precision (lower variance)

d) type

proportionate

= equal sampling fraction

disproportionate

= unequal sampling fracture in each stratum

3. systematic sampling

a) procedure

N/n=k the interval siza

random select number drom 1-k

take every kth unit

b) assumption

population is randomly ordered

c) advantages

easy

more precise than simple random

4. cluster (area) sampling

a) procedure

divide into cluster

randomly sam[le cluster

measure all unit within sampled cluster

b) advantage

administratively useful, esp wide geo area

5. multistage sampling

a) multi stage cluster or

b) any combination of single stage

B. Non Probability

= likely to misinterpret population, difficult/impossible to be detected

1. accidental/haphazard/convenience

= problem: representativeness

a) man on the street

b) available/accessible clients

c) volunteer sample

2. modal instance

= problem: not represent modal group proportionately

a) sample for the typical case

3. purposive

= problem

= proportionality

= need theory to sample the extreme

a) predefined group

b) extreme group

4. expert

= problem: expert may be wrong

a) panel judgement of representativeness

b) advantage

the judgement support the sampling

5. quota

a) proportional

= problem

= how do u pick the cx

= how do you know the proportion

objective

represent major characteristic of pop

by sampling a proportional amount of each

b) non proportional

6. heterogenity

a) make sure you include all sectors

b) use when one or more people are a good proxy for the group

c) for instance, when brainstorming issues across stakeholder groups


Salam Sukses

0 komentar:

AddThis

Share |

  © Free Blogger Templates 'Greenery' by Ourblogtemplates.com 2008

Back to TOP