All research and therapeutic program designs must include a data sampling component. Data sampling is the process researchers use to seek out a representative segment of the population they wish to study. The primary question a researcher needs to ask is, “How will I recruit, find, or locate participants?”
Depending on the research design, one may use quantitative sampling, such as connivance sampling, or qualitative sampling, such as the snowball effect. Researchers base their sampling procedures on the design and the criteria defined for including participants. Therefore, not all potential participants or the population at large would qualify for the study. Please be sure to review the audio interview on sampling issues in this unit’s first studies.
One essential aspect of addictions counseling is to understand counseling intervention and prevention theories, models and approaches to treating individuals and groups. Understanding these is essential in being able to integrate these models in treatment planning and therapeutic interventions.
To successfully complete this learning unit, you will be expected to:
- Evaluate any sampling concerns for your proposed project, and identify possible solutions.
Quantitative and Qualitative Research
Why is sampling important?
It is important for learners to know that sampling is a methodological process and this is the process that is driven by logic, where the researcher deliberately chooses the objects of their study. And when I say objects of the study, what I mean is that sampling is a process that can be used in social sciences research but it is used in other areas of research as well.
So it might be used in biological research or cardiovascular research and the objects could be anything from people to measures of heart beats to anything. When we sample we are always trying to get an approximation of what’s really happening out there.
When we talk about sampling in our methods in social science it is really to gather the who is being studied in the particular research project, but it is important for learners to know that objects of study may be anything and really they are going to vary by the discipline of the researcher.
Before a researcher begins a study they have to think through what their sampling method is and this has to be specified, the procedures have to be reliable—meaning that the person is going to do the same thing as they recruit different individuals or objects in to the study—so that they do not introduce some type of systematic bias. And there are many types of sampling strategies in research, but pretty much researchers will choose the strategies that they will use for any single enquiry based on the nature of the problem in a specific question being studied.
And to choose sampling method, researchers first must really identify their target population. The target population of the study is the conceptual group to which the researcher wants to relate their ultimate findings. So if I am studying clinical interventions I may want to generalize to clinical population. Or if I am studying organizational behavior I may want to generalize to a specific type of organization or a particular type of employee and specific types of organizations. Or if I am studying addictive behavior, I may want to study individuals with a certain level of severity of problems such as mild dependence or in contrast severe dependence. And really the choice will all depend on the research question.
In quantitative studies in particular this usually means that this is the population that the researcher wants to generalize their findings or the population for which the findings are to be relevant. So the sample is, again, who or what is chosen for the actual study while the population, which is another word that comes up in this area is really the larger universe from which the sample is drawn and for which the sample is a theoretical representation.
It is important to know that sampling is a useful strategy in that it can be used in research as a design element as well. What I mean by that is that it can be used to limit a study in a purposeful manner. So while researchers will use inclusion criteria to set parameters of who is needed to be in the study, they can also choose their sample using exclusion criteria to limit the presence of particular variables that they do not want in there.
So for instance if I want to study normal behavior among individuals I may use a variety of exclusion criteria to remove perspective participants who might have a particular characteristic. So, for instance, maybe I want to study people who have depression, but I don’t want people who have psychotic depression or active psychosis. Maybe I want to study people in an organization who are employed and been there at least five years, but I don’t want people who have a history of being laid-off for long periods of time.
So I might, again, use my sampling procedures to help me focus in and really get someone who is a bit more specific. Another one might be if I am studying educational practice among successful students I might want to include individuals who self-identify as being successful, but exclude any students from my sample who have ever received a grade lower than a particular grade like a C or a B, but anyway this would all be determined by the specific study.
Another aspect of how one is going to determine what the target population is, also relates to the unit of analysis in the study. So the unit of analysis is a phrase where we are clearly defining what type of objects the researcher is studying.
In psychology, this could be something like the individuals of a certain type; for instance, like students, patients, clients. The unit of analysis could also be pairs of individuals, so we might sample pairs such as co-workers, or couples in relationships, or student-teacher dyads. Then we can also sample group, so we might—instead of looking at individuals or dyads we might look at specific families, we will sample families or sample work teams or platoons or classrooms and then even bigger than that we could also sample organizations such as businesses, associations, branches of the armed services among other things. In fact, in quantitative research you can even sample studies that are published such as in procedures like meta-analysis.
So, again, what we are actually sampling for the focus of our study all depends on the nature of the study and the specific questions. Once the unit of analysis is done and the researcher is clear in on what types of objects they need from the universe that’s out there—the population—then they have to determine how many they need and this means that they will have a number of participants or objects that they are going to collect for their study to be complete.
In quantitative studies this is typically determined by a power analysis and sampling will proceed until a specific number of objects are collected or until a specific number of types of objects are collected. For instance, some studies will need an N of a certain number, a total number of people; but they can be any objects that meet inclusion and exclusion.
Other studies might require certain number of different types of objects; for instance, if the study was contrasting men and women and wanted equal groups, they might keep sampling until they got the number they needed for adequate statistical power, but they then might have to do something like oversample one group to get equal groups.
So oversampling is also a strategy that researchers might use at times and again this might be to achieve a particular proportion of persons or objects, again in their study, or to get equal groups or a number of other things.
Then, from here, I would go into the second question, which would be that specific types of sampling.
What are sampling procedures?
In terms of specific sampling approaches for quantitative studies there are many potential strategies available and some are more typical in social science research than others. They vary a really along the lines of how much time and expense each takes, and also have relationships to different types of research designs and statistics.
One method that’s commonly used is called convenient sampling. This is a strategy used in research where the researcher is doing relatively unsophisticated exploratory work in an area. Getting the sample is pretty easy or convenient and it is also much less expensive than other methods of sampling, generally with convenient sampling participants meet inclusion and exclusion criteria, and can be accessed through a variety of methods of contact that are typically unrelated to the overall design.
Treatment and Intervention Models
So the bottom line is we are just going to get number of people that we need through a variety of methods and we are really not going to pay attention to that we just want to get people who, kind of, fit certain characteristics or objects, if you will, that fit certain characteristics and the topic is really to be convenient and not terribly expensive.
In contrast, a more expensive type of sampling is random sampling and this is where each member of the target population has an equal chance of being selected, random sampling is used when a researcher wants to see what is happening in a population, but really can’t get every single person or object in that population to participate in the study.
So, the sample is determined by a random selection process and then hopefully the sample that is arrived at will be representative of that broader population from which it came. It’s important to note that random sampling and random assignment are not the same.
Random sampling applies to how we select objects or people for a study. Random assignment applies to what we do after the objects or people have been selected such as in randomly assigning individuals some sort of experimental treatment or control. It is also important to note that randomization is not a unitary process; there are really many ways to randomize or randomly select people.
So, when learners say that they will use randomization, they need to specify their methods. For instance, will a learner use a random numbers table to select participants, or a computer-generated list of individuals to select people, or will they toss a coin to assign someone to a condition, or something different like a process called earn randomization where assignment may be linked to a complex set of characteristics?
There are many, many methods, so we want to keep clear on exactly what we mean by randomization, especially when we talk about it in terms of sampling methods.
Another method that learners will hear about for sampling is called snowball sampling and snowball sampling is used when the population of interest may be hard to find or difficult to access and in such situations it can be very expensive and time intensive to locate prospective participants.
So, what a researcher might do is use referrals from people they already have in the study to help locate additional prospective participants. While this method can introduce some bias into a study because of the social connections and how they might affect, you know, who comes in to the study it is still a good lower cost method relative to random sampling in populations with rare characteristics.
So, snowball sampling is a method that is seen when people maybe are really like needles in a haystack, they are really, really tough to find which you need for your study and this may be a better way to locate people to reduce time and expense.
Treatment and Intervention Models
Two other methods frequently used are stratified sampling and quota sampling. Stratified sampling is used when a researcher is comparing groups; the groups are sometimes called strata, and when researchers are trying to compare groups they often want to try to limit sampling error.
So, strata the two groups or two or more groups that are being compared they are usually defined by specific characteristics. Any singular stratum is uniquely different from the others based on at least one characteristic. So, for instance, if we were doing an educational study on adult learners, we might separate people into groups by age or income in the past year, or if we are doing some type an organizational study in hospitals, we might separate individuals into stratums such as degreed administrators, degreed healthcare providers, or non-degreed support staffs.
Once the strata are determined objects meeting those criteria are then located and randomly selected and then this particular method is useful when one has a specific research question, where you want do comparison of groups. But literally the different groups that we pick the strata, they are very, very specifically defined and then we want each of them to have some random representation of people from that strata.
In contrast, a slightly different way of having groups that are compared, but may not be of the exact same size, is a related method that does not use random selection and it is called quota sampling.
In this method strata are identified and defined, but they are defined in terms of their relative proposition in a population as a whole. So for instance in the study of nurses we might like to have male and female nurses to represent a sample of nurses, but we know that female nurses represent like 90% of the strata overall of the population of nurses, so in this case our strata might be males and females, but we have set up the strata so that 90% of the samples are one stratum and defined as females and the other will be 10% defined as males.
We could find individuals meet inclusion and exclusion criteria and then instead of using some type of random sampling to put them into the groups, we do convenient sampling. So when you use quota sampling the issue is more of what is the relative preposition of people in each of the strata that are going to be included as opposed to getting, sort of, some random representation of people in each strata like you might do in stratified sampling.
What are common sampling errors?
Errors that learners will make or even experienced professionals may make in sampling: The errors tend to be focused on people not thinking through their sampling framework completely on the front end and maybe not having their groups be as exclusive as they need to be. This is especially important if you are doing a type of design were people need to be distinction from one another, then people have to set up a sampling method instead of procedures that will really put people in unique groups.
Treatment and Intervention Models
So sometimes researchers may find that their protocols for sampling, needs revision. In quantitative studies this is important too especially as people are doing what is called a probability-based sampling method.
In a probability-based sample, population measures have a defined probability of being selected and this, you know, varies by the type of method used. For instance, in random sampling one example is where all members have equal probability of being included and because that is one of the hallmarks of random sampling if somehow the researchers introduce some type of error into the sample then that can cause problems.
So, for instance, maybe their protocol requires that everybody who is sampled goes through a set of procedures where they then are brought into the study using some type of written letter, but somehow somebody wiggles through and they get in touch with the researcher through an email or through word of mouth and they do not go through the standard requirement process, but now we have somebody who could have gotten in, in a slightly different way and then now our sample may not be as random. Who knows maybe they knew someone or they had some special insight or they had some special motive to participate and that can contaminate things, so, again, following protocol is important. Treatment and Intervention Models
Another issue around sampling that can come up is that we have people where they come into the study and they do not quite meet the inclusion or exclusion criteria, but may be somebody who is working for research project really wants to try to get that in, they are really focused on getting done with the study and so they start monkeying around with how they ask the questions that might, you know, put somebody in the study or out of it and again they are not following protocol and now we start getting people who really are different from other people in the study.
So, again, just following protocol is one way that sampling methods can be kept in good shape or really seriously knocked out of whack if people start getting in a hurry to get their end or they are not paying attention to the details in terms of how they do their recruitment and bring people into the study.
Another issue with sampling can also be that we are not clear in terms of the criteria we use to find any kind of strata or selection criteria that we have. So if there is some kind of ambiguity in terms of how we are asking people questions during that process of getting people into our work, then we can also have errors and bias creep in to a study. Treatment and Intervention Models
In fact, how you actually define your sampling methods on the front end of your study is very important. There is a bias called pre-inclusion attrition bias, which is where people are identified for a study appropriate, but then somehow they disappear and never quite make it in.
There is always a portion of people who will fit that bill and what you want to try do is keep that number as small as possible so that your sample remains representative to whoever you wanted it to represent, but pre-inclusion attrition bias can definitely affect the study and, again, I think it is important to remember that no study is bias-free, but we always want to try to control any kind of bias that we are either accidentally or through our producers putting into a study.
Read the following in Royse, Thyer, and Padgett’s Program Evaluation: An Introduction to an Evidence-Based Approach:
- Chapter 8, “Sampling,” pages 211–225.
Consult the Library to locate three peer-reviewed journal articles that are relevant to your final project. Use these articles in this unit’s “Recruitment and Sampling Choices” discussion. Treatment and Intervention Models
For this assignment, you will conduct a literature review.
The Library is a good place to begin your research. As you conduct a literature review, keep in mind the major project requirements, and your selected population, sampling procedures, methodology, and possible interpretations.
APA Style Central provides guidance for citing sources and formatting your paper in proper APA style. See the APA Style and Format section of the Writing Center for instructions on accessing and using APA Style Central. Treatment and Intervention Models
For this project component, you are expected to locate 15 references, with 10 coming from peer-reviewed journals, to support your final project. The articles you choose should cover best practices in addiction counseling, intervention and prevention theories and models that relate to addiction counseling with your chosen population, culturally and ethnically sensitive counseling or therapy services, and appropriate assessments to evaluate, diagnose, and develop treatment plans for your chosen population.
Write two to three sentences about each peer-reviewed journal article. Your descriptions should synthesize results into a summary of what’s known and not known, identify areas of controversy in the literature, and formulate questions that need further research.
Remember, peer-reviewed journal articles must be published in scholarly journals, such as the Addiction Journal or American Journal on Addictions. Additionally, these journal articles must be relevant to your project and published during the last five years. If you use an article that has been published more than five years ago, you may use it but it does not count toward the requirement of the 20 peer-reviewed journal articles.
Recruitment and Sampling Choices
By this time, you should have selected your target population for your research project. An important question to ask and answer is, “How will I recruit participants?”
Depending on the research design, program evaluation, or therapeutic program, a completely inclusive process is not always appropriate. From the course readings and your selection of the journal articles in this unit’s first studies, what are the sampling choices and concerns for your specific proposed study? Treatment and Intervention Models
Discussion 2: 5 days order with 1 page needed with minimum of 250 words and 2 references.
Treatment and Intervention Models
Examine various addiction counseling models for treatment and intervention. Demonstrate how you would apply addiction counseling intervention and prevention theories, models, and approaches for individuals and groups.