Blog post #3- Sampling

My current research problem is “How do the Directed Reading-Thinking Activity and Story Mapping comprehension strategies utilized throughout one school year retain their effectiveness in terms of usage over a 2-year span with elementary students of varying reading ability levels?  Also, how is students’ motivation about comprehension tasks in the 2-year follow up?"

My participants for this study will consist of 5 different school districts in the central Virginia area including Richmond City, Henrico, Hanover, Chesterfield and Goochland.  I will choose one elementary school in each district that I believe is representative of the district as a whole in terms of socioeconomic status.  From those 5 elementary schools I will then ask the principal to suggest one 2nd grade class that consists of students with varying reading ability levels.  Therefore, I will have five 2nd grade classes of students as participants in this research which could be generalizable to all 2nd graders in central Virginia.  The students will be participating in the study during their 2nd grade school year and a follow-up will be completed 2 years later when the participants are in 4th grade.

Below are the pros and cons of using different types of probability sampling methods in this research study:

Probability Sampling	Pros	Cons
Random Sampling	-Every 2nd grader has an equal chance of being selected. -A large participant group could be used	-Too many variables would be involved -It might be harder to follow up 2 years later with a random sample -Might not be representative of the population -Would require numbering each participant in the population
Systematic Sampling	-Simple to select participants -Every 2nd grader has an equal chance of being selected	-Students selected could all have similar characteristics, such as a high reading level -Might not be representative of the population -It might be harder to follow up 2 years later with these students
Stratified Sampling	-More representative of the population than random or systematic sampling -A smaller sample could be chosen -Able to narrow down which participants the researcher wants to study from the subgroups	-It may be difficult to prepare and gather data about the subgroups in the population -I would have to assign proper weighting to each subgroup in order to represent the population -Students would still be chosen randomly from each subgroup which could result in students with all similar characteristics
Cluster Sampling	-Would work well with a large population -Saves time and money -Would be easy to select a school as a naturally occuring group	-The results are less accurate than other  probability sampling methods -Involves all random selection which may result in a sample that does not represent the population -Would be difficult to follow students in a longitudinal study since everything is random

Below are the pros and cons of using different types of non-probability sampling methods in this research study:

Non-Probability Sampling	Pros	Cons
Convenience Sampling	-Easy access to participants -Follow-up with participants would be simple -Less time-consuming -High participation rate -Would work well if studying very specific topic on very specific participants	-I could not generalize the convenience sample to the population -May result in bias such as participants in only one socioeconomic status
Purposive Sampling	-Easy to answer research problem with useful subjects -Less time-consuming -Helpful in qualitative research in order to understand the problem in depth -Selected participants will provide the best information	-May be difficult to generalize to the population as a whole -Results depend on the characteristics of the subjects -May be difficult to find the exact subjects to participate.
Quota Sampling	-More representative of the entire population -Generalizable to other similar subjects -Researcher is able to set guidelines of who they choose as participants	-More time-consuming because the researcher must make decisions about selecting the sample -Results will depend on the characteristics of the participants

I believe the sampling procedure that fits best with my research problem is quota sampling.  My target population is central Virginia, but I would like to have some control over who my participants are in order for the study to be generalizable to all 2nd grade students in central Virginia.  I selected the five school districts and I will non-randomly decide on one school in each district that is representative of that population.  At each school I will be asking the principal to suggest a class that consists of 2nd grade students with varying reading abilities to participate in the study.  Therefore I am not handpicking each participant, but the sample is also non-random.  If I chose any of the probability sampling methods there would be too many variables involved such as SES, student achievement and my ability to follow these students 2 years later.  The convenience sampling method would not allow me to generalize my results to all 2nd grade students in central Virginia.  I think the purposive sampling technique would also be too specific in trying to generalize to all 2nd grade students.  I am trying to get a picture of whether these comprehension strategies are effective with students of varying ability levels and living in varying SES neighborhoods.  I think by narrowing my sample the way I have I will be able to achieve these results.  The quota sampling technique seems to fit best with setting guidelines for my participants but not being too specific.