But first, here's a little background on the study before we dive into the pros and cons of each sampling method (and the WINNING sampling procedure). My hope in this study is to discover the perceptions early and middle adolescent girls have regarding the effects their single parent homes have on their school performance (which includes grades, attendance, involvement in activities). The purpose of this study is to not generalize that every adolescent girl has a certain perception regarding their home life's influence on their school life. However, the intent is to gain knowledge and understanding of the beliefs these girls have pertaining to the relationship of their single parent homes and school performance to better support these students by either supplying interventions or resources. Ok, so now really, a-sampling we will go...
Simple Random Sampling: Simple random sampling is a probability sampling method used to give every member of the population the same chance of being selected. To do this type of sampling, I would have to make a list of every middle and high school female student who lives in a single parent home in all of America. Although it would give a vast representation of the sampling population, it would not be a wise use of time (or money...if I was getting paid to do this) to develop the list alone--and then contact all of the randomly selected participants. On top of that, I can't even begin to imagine to interview (the planned methodology) all the participants. Simple random sampling, you are not this researcher's friend.
Systematic Sampling: Like simple random sampling, systematic sampling is another probability sampling. The difference between this procedure and the previous one is that every nth member of the list is selected. Again, what a spectrum of responses that could lead to a broad understanding of exactly what adolescent girls from single parent homes need in the school to support them, BUT the time it would take to develop the list and then contact every nth name from the list of all middle school and high school girl students whose home life consists of one parents would be unfathomable. Thank you systematic sampling for existing...just not for this research question.
Stratified Sampling: Like the name denotes, this method of probability sampling involves identified subgroups of the population. If I wanted to analyze the difference between early ( approximately middle school aged) and middle (approximately high school aged) adolescent girls, then this sampling technique would be useful in ensuring that an adequate number from both groups were selected for sampling (proportional stratified sampling). For the purposes of this particular study, that isn't a variable I am considering, so it wouldn't be that useful. Also, since stratified sampling is an probability sampling, the big ol' list would have to still be made encompassing the population of all adolescent girls hailing from single parent homes. Thanks, but no thanks, stratified sampling.
Cluster Sampling: Cluster sampling is our last look at probability sampling. The catch of cluster sampling is that it is used when a list of individuals doesn't exist by selecting naturally occurring groups (i.e. middle and high schools---DING DING DING!!!). Though a list of all middle school and high school females from single parent homes COULD possibly exist, the reality of it ever coming into fruition is VERY slim to none. This gives cluster sampling a leg up on the others. It would be MUCH easier (and wiser with time and money) to choose a state (like Virginia, for instance), and list all the school districts, then all the middle and high schools from the school districts, and randomly selecting one of each to contact. After contacting them, a list of the names given will be created, and this is where the final random selection will happen. SHEESH! That still seems like a lot of work, though it could give a comprehensive understanding. Cluster sampling, you feel so right, but you're still so wrong!
Convenience Sampling: This is our first look at non-probability sampling. Convenience sampling is a less time-consuming and costly procedure, allowing the researcher to choose an available sample of participants. Anything with the word "convenient" in the name is something I tend to be a fan of. Unfortunately, for this study, a convenience sample wouldn't allow for the specific criteria (adolescent female, single parent home) to be addressed. Convenience sampling=just not that convenient for this study.
Quota Sampling: The second type of probability sampling is quota sampling. This involves the researcher selecting a sample that would represent an entire population. I like the method because instead of the dreaded "list" (see above), it would allow for a geographic area, e.g., Richmond City or Henrico County, to be selected, and to obtain names from chosen schools in those districts. However, the representation of the entire population isn't necessarily a goal of this research study, so it would save time (and again, potentially money) if I went with...(DRUMROLL)...
Purposive Sampling: This is the chosen sampling procedure I would utilize for this research study. Purposive sampling, the final non-probability method, gives the researcher the freedom to select the participants knowing that they would be particularly informative. Since the criteria of the participants gets specific and can be a somewhat sensitive topic for adolescent females to talk about, utilizing purposive sampling makes the most since for time purposes and to receive a better result. So congratulations, purposive sampling--you do serve a purpose in this research study!
Good thought process here...very practical!
ReplyDeleteQuick question--what's the biggest problem with systematic random sampling?
Totally forgot to mention that systematic sampling arranges list in a systematic way, so there can tend to be a cyclical pattern in every "nth" participant. This doesn't allow for the data gathered to be very representative of the sampled population.
Delete