Sampling scheme

[Byrd Sellers, Johnita]

Hello All,

 

I am attempting to find data on several different sites.  Some of the sites are going to be smaller than others.  I wanted to create a sampling scheme so that those smaller sites are well represented.  For instance, could some type of “weighting” be performed on the data.  Here is an idea of what we have:

Say we have 30 locations and from each location we want to get about 100 cases.  In some of the locations we may have 50 cases so we get information on all 50 of these cases.  In other locations, we may have 150 cases so we get all of the 100 cases necessary.  In one instance, we have 100% of that site represented and in the other cases, we have 100 of the cases necessary but we may only have about 65% of that site represented.  What is the best way to make sure that all sites are represented effectively?  I’m looking for other options besides a random sampling.

 

Any suggestions would be greatly appreciated.  Thank you in advance.

 

Sincerely,

Johnita

[John F Hall]

Johnita

 

I've taken the liberty of forwarding this to some colleagues outside the SPSS list who do sampling stuff for a living.

 

John Hall
[hidden email]
http://surveyresearch.weebly.com

----- Original Message -----

From: [hidden email]

To: [hidden email]

Sent: Monday, October 04, 2010 3:52 PM

Subject: Sampling scheme

Hello All,

 

I am attempting to find data on several different sites.  Some of the sites are going to be smaller than others.  I wanted to create a sampling scheme so that those smaller sites are well represented.  For instance, could some type of “weighting” be performed on the data.  Here is an idea of what we have:

Say we have 30 locations and from each location we want to get about 100 cases.  In some of the locations we may have 50 cases so we get information on all 50 of these cases.  In other locations, we may have 150 cases so we get all of the 100 cases necessary.  In one instance, we have 100% of that site represented and in the other cases, we have 100 of the cases necessary but we may only have about 65% of that site represented.  What is the best way to make sure that all sites are represented effectively?  I’m looking for other options besides a random sampling.

 

Any suggestions would be greatly appreciated.  Thank you in advance.

 

Sincerely,

Johnita

[Jon K Peck]

The SPSS Statistics Complex Samples option lets you draw stratified, multistage samples according to a wide variety of sampling schemes, and it includes a bunch of statistical procedures that take the sampling scheme into account.

HTH,
Jon Peck

-----"SPSSX(r) Discussion" [hidden email] wrote: -----

To: [hidden email]
From: "Byrd Sellers, Johnita" [hidden email]
Sent by: "SPSSX(r) Discussion" [hidden email]
Date: 10/04/2010 07:52AM
Subject: [SPSSX-L] Sampling scheme

Hello All,

 

I am attempting to find data on several different sites.  Some of the sites are going to be smaller than others.  I wanted to create a sampling scheme so that those smaller sites are well represented.  For instance, could some type of “weighting” be performed on the data.  Here is an idea of what we have:

Say we have 30 locations and from each location we want to get about 100 cases.  In some of the locations we may have 50 cases so we get information on all 50 of these cases.  In other locations, we may have 150 cases so we get all of the 100 cases necessary.  In one instance, we have 100% of that site represented and in the other cases, we have 100 of the cases necessary but we may only have about 65% of that site represented.  What is the best way to make sure that all sites are represented effectively?  I’m looking for other options besides a random sampling.

 

Any suggestions would be greatly appreciated.  Thank you in advance.

 

Sincerely,

Johnita

====================To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD

[David Marso]

John,
I am sure these people will LOVE you for posting their email addresses in
the header of a message in a public forum!  Haven't you ever heard of BCC?

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD