I am part of a research team looking at the outcomes of a homework help program and would like to make a few suggestions of how to generate data.

I am thinking of using a sampling method, but am not certain which one (cluster, simple random?)

Details:

1. Students are attending a homework help program offered at a community center

2. These students come from 30 different schools: in some cases, schools send 1 or 2 students to the homework help, while others send 15 to 20

3. We need to get comparitive data from these schools to see if after X years, the students who have received homework help are performaning better than their non-participating peers.

4. we don't have the resources nor the time to contact 30 schools and get approval for student academic data from their 30 schools

I was thinking that a sampling method might be useful - but which one?

A cluster sample methose - where clustered are defined by

1. geographic area and/or

2. number of students from school participating in the program (so cluster those schools sending 1 or 2 student together, 5 to 10 etc..) but that doeasn't sound right.

OR a simple random sampling method - randomly select X schools (10 for example) to particpate in the analysis regardless of how many of their students are attending homework help.

OR a proportionate stratefication + simple random sample - stratefy the schools according to a range of students sent/attending program and then a simple random sample per strata (sample size of each strata would be proportinate to the strats's pop size).

Any suggestions (readings, articles, methods) would be greatly appreciated.

Thank you!

I am thinking of using a sampling method, but am not certain which one (cluster, simple random?)

Details:

1. Students are attending a homework help program offered at a community center

2. These students come from 30 different schools: in some cases, schools send 1 or 2 students to the homework help, while others send 15 to 20

3. We need to get comparitive data from these schools to see if after X years, the students who have received homework help are performaning better than their non-participating peers.

4. we don't have the resources nor the time to contact 30 schools and get approval for student academic data from their 30 schools

I was thinking that a sampling method might be useful - but which one?

A cluster sample methose - where clustered are defined by

1. geographic area and/or

2. number of students from school participating in the program (so cluster those schools sending 1 or 2 student together, 5 to 10 etc..) but that doeasn't sound right.

OR a simple random sampling method - randomly select X schools (10 for example) to particpate in the analysis regardless of how many of their students are attending homework help.

OR a proportionate stratefication + simple random sample - stratefy the schools according to a range of students sent/attending program and then a simple random sample per strata (sample size of each strata would be proportinate to the strats's pop size).

Any suggestions (readings, articles, methods) would be greatly appreciated.

Thank you!

Last edited: