|
Links |
|
Links |
|||
|
|
See Also: Research Design |
|
|||
|
|
See Also: Consent & Confidentiality |
|
|||
|
|
The choice Participants or subjects of the Research & Sampling
Strategy
should addressed at the very beginning of the Research Design process but must be accomplished by this point |
|
|||
|
|
The UVaWise HIC requires the following information on Sampling &
Participants
a) Describe the subjects b) Sample size c) How will the subjects be selected for participation? d) Diagnostic criteria for inclusion e) Criteria for exclusion |
|
|||
|
|
Each will be discussed below after a discussion of general sampling issues |
|
|||
|
|
A Sample is a subset of some larger population that is studied
for the purposes of drawing conclusions about that larger population |
|
|||
|
|
In Sampling strategy, Probability Sampling
is preferable to Quota Sampling |
|
|||
|
|
The choice of Sampling Strategy is determined by FOUR factors
a. funds b. numerical & geographic scope c. availability of adequate sampling frame d. chosen method for collecting data |
|
|||
|
|
The basic way to obtain a representative sample is by a random draw
from
everyone in the population in question. |
|
|||
|
|
No matter how small your population, you almost always need 50 people
in
order to have a scientifically valid sample. |
|
|||
|
|
1,100 is enough for most major survey firms to survey the US |
|
|||
|
|
Thus if you divide your sample into different groups, you need even
more
people. |
|
|||
|
|
If you want to compare men & women: 50 of each
If you want to compare lower income, middle income & upper income men and women, how many do you need? (answer found below) |
|
|||
|
|
There are more than ELEVEN Types of Sampling,
and hybrids can be made by combining various methods |
|
|||
|
|
Cluster Sampling is the most widely used
A sample of groups ( clusters ) is drawn before individuals w/in them are identified & selected |
|
|||
|
|
Interval Selection: from a list, choose respondents at fixed intervals |
|
|||
|
|
Non-Randomized, Snow-Ball sampling methodology
based on access to the subjects at some location or organization |
|
|||
|
|
Observational sampling
Continuous sampling Time sampling Time-point sampling Time-interval sampling Individual sampling Event sampling Blanket or satuaration sampling |
|
|||
|
|
Continuous sampling
Observe subjects or a period of time, recording behaviors during the whole period May use video |
|
|||
|
|
Time sampling
Recording behaviors only at specific times |
|
|||
|
|
Time-point sampling
Record beh at specific pts in time |
|
|||
|
|
Time-interval sampling
Record beh during preselected intervals |
|
|||
|
|
Individual sampling
Record beh of 1 subject at a time, maybe for a specified time |
|
|||
|
|
Event sampling
Record observations during specific events |
|
|||
|
|
Blanket or satuaration sampling
Record beh & obs of all activity Combine w/ event sampling, where you focus on events that are most imp to your thesis |
|
|||
|
|
Survey Sampling is the process by which you chose who will take the survey |
|
|||
|
|
If you compare two or groups of people, survey equal numbers of each |
|
|||
|
|
In many cases, a sample of 50 gives one a scientifically valid sample for each variable examined |
|
|||
|
|
Nonresponse can ruin any sample |
|
|||
|
|
Whatever the sampling strategy, you must gain the cooperation
of as many members of the sample as possible |
|
|||
|
|
When is a sampling rate consider adequate? |
|
|||
|
|
No clear answer |
|
|||
|
|
Your Statistical Package will give you a number of options for dealing w/ non-responses |
|
|||
|
|
The UVaWise HIC requires the following information on Sampling &
Participants
b) Sample size c) How will the subjects be selected for participation? d) Diagnostic criteria for inclusion e) Criteria for exclusion |
|
|||
|
|
a) Describe the subjects: |
|
|||
|
|
Describe the subjects only to the extent that is necessary to address your Research Objectives |
|
|||
|
|
Example: Include gender if that is an
area of examination
Include any special quality you may need: e.g., firefighters, managers, victim, student, etc. |
|
|||
|
|
b) Sample size: |
|
|||
|
|
For the senior project, 30 subjects are required; for a survey,10 for interviews |
|
|||
|
|
The Researcher should mention that because of time & money limitations,
the sample size is not scientific, i.e., adequate to statistically support the thesis in full However, the sample size will be adequate to support one or two of the Research Objectives |
|
|||
|
|
c) How will the subjects be selected for participation? |
|
|||
|
|
Most Researchers will use a non-randomized, snow-ball sampling methodology
based on access to the subjects at some location or organization |
|
|||
|
|
d) Diagnostic criteria for inclusion |
|
|||
|
|
Most surveys have no diagnostic criteria for inclusion or exclusion,
other than access to the subjects at some location or organization |
|
|||
|
|
Criteria for inclusion is established on the survey:
Example: Are you a wildland firefighter? Are you a victim of ...? Have you ever...? |
|
|||
|
|
e) Criteria for exclusion. |
|
|||
|
|
Critical criteria for exclusion will be based primaryily on self-exclusion resulting from the Introductory letter |
|
|||
|
|
Noncritical criterial for exclusion will be based on meeting the criteria for the sample: ie male, female, etc. |
|
|||
|
|
Go to UVW's Webpage |
|
|
|
Go to Dr. W's Webpage |
|
|
|
Go to the Course Information Webpage |
|
|
|
Go to Course Resources |
|
|
|
Go to Course Supplements |
|
|
|
Go to In-Class Projects |
|
If you want to compare lower income, middle income & upper income men and women, how many do you need? 6 grps:
|