sampling s 03

8/2/2019 Sampling s 03

1/16

I NTRODUCTI ON

TO SURVEY SAMPLI NG

Febru ary 26, 2003

Karen Foote Retzer

Survey Research Laborat ory

University of I ll inois at Chicago

www.srl.uic.edu


2/16

I nt roduction to Survey Sampling Page 2 of 16

Census or Sample?Census:

Gathering information about every

ind ividual in a popu lation

Sample:

Selection of a small subset of a population


3/16


Why Sample instead of t aking a

Census?

Less expensive

Less time-consuming

More accurate

Some samples can lead to statistical inference

about the entire population


4/16


Probabil it y Sample

Generalize to the entire population

Unbiased results

Non-Probabil it y Sample

Exploratory research

Convenience


5/16


Target Populat ion

Definition: The popu lation to which w e want to

generalize our find ings.

Unit of analysis: Individu al/ Household/ City

Geography: State of Illinois/ Cook County/

Chicago

Age/Gender

Other variables


6/16


Examples of Target Populat ions

Population of adults (18+) in Cook County

UIC faculty, staff, students

Kids und er 18 in Cook County


7/16


Sampling Frame

A complete list of all units, at the first stage of

samp ling, from which a sample is drawn

Examples:

Lists

Phone numbers in specific area codes

Maps


8/16


Sampling Frames

Example 1:

Population: Adults (18+) in Cook Cou nty

Possible Frame: list of phone numbers, list of block maps

Example 2:

Population: Females age 4060 in Chicago

Possible Frame: list of phone numbers, list of block maps

Example 3:

Population: Kids und er 18 in Cook County

Possible Frame: List of schools


9/16


Sample Designs for Probabil it y

Samples

Simple Random Samples

Systematic Samples

Stratified Samples

Cluster


10/16


Simple Random Sampling

Definition: Every element has the same probability of

selection andevery combination of

elements has the same probability of

selection.

Probab ility of selection:n/N,

where n=sample size; N=population size

Use Random N umb er tables, software packages

to generate rand om num bers

Most precision estimates assume SRS.


11/16


Systemat ic Sampling

Definition: Every element has the same probability of

selection, but not every combination can be

selected.

Use when draw ing SRS is difficult List of elements is long & not comp uterized

Procedure Determine Population size N & sample size n

Calculate Sampling Interval (N/n) Pick random start between 1 & Sampling Interval

Take every ith case.

Problem of Periodicity


12/16


St rat if ied Sampling:

Proportionate

To ensure samp le resemb les some aspect of

population

Popu lation is divided into sub groups (strata)

Students by year in school

Faculty by gender

Simple Random Samp le (with same probab ility of

selection) taken from each stratum.


13/16


St rat if ied Sampling:

Disproportionate

Major u se is comparison of subgroups

Popu lation is d ivided into subgroups (strata)

Compare girls & boys who p lay Little League

Compare seniors & freshmen who live in d orms

Probability of selection needs to be higher for

smaller stratum (girls & seniors) to be able tocomp are subgroup s.

Post-stratification weights


14/16


Cluster Sampling

Typically u sed in face-to-face surveys

Popu lation d ivided into clusters

Schools (earlier example) Blocks

Reasons for cluster sampling

Reduction in cost

No satisfactory sampling fram e available.


15/16


Determ ining Sample Size: SRS

Need to consider Precision

Variation in su bject of interest

Formula Sample size no = CI2 *(pq)

Precision

For examp le: no=1.962

* (.5*.5).052

Sample size not dependent on popu lation size.


16/16


Sample Size: Other I ssues

Finite Population Correction n=n o/ (1+no/ N)

Design effects

Analysis of subgroups

Increase size to accommodate non-response

Cost

sampling s 03

Documents