According to Business Dictionary, probability sampling is a method of sampling that uses individual units as a sample, and each sample has a fixed and known probability of selection so that each sample has an equal chance of being randomly selected in the research process (Probability Sampling). Examples of randomization in probability sampling are often used by people without them even realizing it, such as when picking names out of hat or choosing straws and somebody chooses the short straw. This is done manually; however, in today’s advanced technological world, computers are used to assist with random sampling by generating random numbers for research sampling purposes. A table of random numbers can also be used.
In research, probability sampling involves analyzing observations of a population of data which is made by studying a representative sample of the whole population. These samples are used to make assumptions about the population in its entirety. The data from the sample is then used to report what is probable about the entire population, as it relates to trends and future events. Probability sampling allows for any one individual unit of a population to be selected and this can be mathematically calculated as a probability (Herek, 2012). This is indicative of quantatitive research.
Additionally, probability sampling is a form of quantitative research because the sample findings are numerically analyzed and mathematically calculated and can be used for future estimates (Quantitative Research). For example, quantitative research could involve analyzing a list of people in the local phone book whose last name begins with a certain letter.
Obtaining probability sampling results requires the development of a sampling plan. This, in turn, requires including and evaluating various factors of the sampling plan such as the sampling plan frame, sampling unit, sample size, target population, and stratification. The sampling plan is the guideline for how the research sampling project will be carried out in the research process. It is important that the sampling plan is clearly defined to reduce the chance of sampling process error. If the sampling process is inaccurate, then this leaves room for inaccurate analytical results; therefore, precision is a must when developing a sampling plan (Leedy & Ormrod, 2005).
The sampling frame is basically made up of the sampling units of the population being used in the sample. In the research process, the population units or frame, need to be defined to show what part of a population will be analyzed in the sample such as telephone numbers of local residents, a certain age group in a city, or a roster of members enrolled in an organization (Herek, 2012). A sampling frame can consist of various stages with different sampling units. The first stage of the sampling frame is made up of primary sampling units and the subsequent stages are made up of enumeration units, which are listing units (Levy & Lemeshow, 2008).
To explain, an example would be a primary sampling unit in a multi-level probability sample may be a school, and the second level may be the science department within the school, and the third level may be all teachers within the science department within the school, and finally the fourth level may be all female teachers within the unit of teachers within the science department within the school. The second through the fourth levels are the enumeration units, and along with the primary unit they make up the sample frame which will be the basis for the research.
The sampling frame is developed from the sampling unit which is made up of groups of individual variables that are potentially part of the research study, or the sampling unit could be respondents to a survey or something similar.
According to Business Dictionary online, the sampling unit is a single section selected from a population for gathering statistics for research purposes (Sampling Units). An example would be, when studying fitness center memberships, a single member at a specific fitness center is a sampling unit. Another example would be a group of people who have the same characteristics within a larger group, such as all students in a school who have a grade point average of 3.0. These students represent a subgroup within a primary group, which is all students. Sampling units can be one individual item or many items in a group.
According to Kotler & Keller (2008), a sample size is literally a count of how many units will be included in the research process. Quantitative research (which is probability sampling) simply takes a representative sample (made up of individual units) of the whole population in a target. The most reliable findings are obtained with large samples rather than small ones. This means it is better to take big pieces of a population to analyze so that there is more representation of the entire population when analyzing results of the sampling.
Using larger samples, as opposed to using smaller ones, takes into consideration the size of the sample, confidence intervals and confidence levels. Confidence intervals relate to margins of error in research through the estimation of specific range in a population. With this, calculating margin of error allows for assessing how reliable the estimation is, which allows for making estimations about the population, based on the sample. Confidence levels refer to the reliability of data gathered from a research sample.
In research, the entire population of units for a survey of data is called the target population. It is the definitive area of the population from which results are to be generalized. For accurate research findings, the target population needs to be specific and defined to determine if sampled cases will be included in the survey findings (Cox & Lavrakas, 2012). Target populations are useful in the research process because they are used to focus research initiatives on answering specific questions and solving specific problems. To do this, a population needs to be grouped into subgroups or targets. This is referred to as stratification.
According to Investopedia, stratification involves dividing a population into groups, also referred to as strata. The groups or strata are segmented based on similar attributes or shared characteristics of the members within the groups. For example, a group or strata could be all black females between the ages of 25 and 35 with type 2 diabetes living in a population of a specific city. The process of stratification would separate this group into a subset for research purposes. With stratification, a random sample from each group is put into subsets and the random sample is formed from this number, proportional to the size of the entire population. Data from the random samples are accurate indicators for making predictions and inferences about the whole population, based on the samples (Stratified Random Sampling).
Sampling Plan Components
All of the above components of a sampling plan, which are the sampling frame, sampling unit, sampling size, target population and stratification are essential to the development of the sampling plan with includes the type of sampling methods considered such as probability sampling. Two types of probability sampling that are commonly used are simple random sampling and systematic sampling.
Simple Random Sampling
According to Babble (2011), the method of probability that is simple random sampling draws subsets from a population and gives estimates of characteristics of the subsets that exist within the total population. Consequently, each random sample will give an estimate of a percentage of the total population that conforms to whatever the research criteria is.
Social research uses simple random sampling, which is a method of statistical computation, after a sampling frame has been established. Simple random sampling research requires assigning random numbers to each unit in a sampling frame so that they can be randomly selected. These numbers can be manually assigned by using a table of random numbers or the assigned numbers can be computer generated.
Randomization of sample selection is necessary to eliminate the possibility of preferential treatment or selectivity bias in the selection process. This can occur with human inquiry and observation when dealing with research sampling, but randomization is used to ensure sample selection integrity. In addition, simple random sampling is used to scale and compare sample designs. The best way to estimate population characteristics is to randomly select data from sample units (Ahmed, 2009).
When dealing with a simple random sample, all units within a sampling frame can be selected and they also have an equal chance of being selected. In addition to this, each unit is independently selected within each stage, instead of being selected in groups (Herek, 2012). This means that although each unit has an equal probability of being selected, their selection does not affect the other units that might be selected from the population. Additionally, random sampling is one of the most used methods of sampling techniques (Teddlie & Yu, 2007).
With simple random sampling, units can either be replaced or not be replaced after they are selected. Any units selected by simple random sampling with replacement can be selected more than once in the study, but simple random sampling without replacement means units are selected one time for the research. Simple random sampling without replacement is used more often in research studies than not (Basic Survey Design).
In addition, it is important to note that simple random sampling depends on the luck or bad luck of the draw, and this means that it is possible that the representative sample may not be a very good interpretation of the population’s subgroups. This is where other sampling techniques are better choices such as with systematic sampling.
According to Babble (2011), systematic sampling uses a list of units from a sampling frame and every so many units in the total list is selected in a systematic fashion to be a part of the sample for research process. For example, every 10th person in a list of 10,000 people may be chosen to participate in a sample, if the research criteria called for a sample of 1,000. With this, the first element is to be selected at random to eliminate any possible selectivity bias. To begin the selection process with this method, a random number between one and ten is selected and assigned as the first unit in the sample. Then, every 10th unit thereafter is chosen systematically.
Advantages of Systematic Sampling
Using the systematic sampling method is easier than using the simple random sampling method mainly because it allows for a little bit more control in the process by using a system to randomly select units, instead of manually selecting them. Another advantage of using the systematic sampling method is having a population that is sampled evenly. Additionally, clustered selection is systematically eliminated when using the systematic method of sampling (Castillo, 2009).
Disadvantages of Systematic Sampling
There is a disadvantage of using the systematic sampling method if the population sampled has specific aspects that occur on a periodic basis or in a definite pattern. This is called periodicity and it can skew sample results because periodic occurrences would compromise the randomness of the sample and would not be an accurate representative sample of the whole population.
Usefulness of Probability Sampling in Quantitative Research
Quantitative sampling is a solution to studying whole populations because it is more efficient than trying to actually study each unit of a population one-by-one. Therefore, a representative sample is obtained and is as accurate as studying the entire population, if the sampling process is done correctly. The representative sample is appropriate for making generalizations and assumptions about a population as a whole. This is the most common method for researching and studying populations.
Quantitative methods are commonly used in social sciences research such as in the areas of sociology and psychology. In relation to this and specifically within these areas, quantitative research methods are useful in social disciplines such as social work, social welfare, criminality or criminal justice, etc. In addition, quantitative research methods use a systematic approach to investigating criteria based on mathematical computations and measurements such as with statistical information. The results that come from quantitative research are meant to be unbiased and accurate so that the researcher can draw conclusions about a population, based on the sample data from the quantitative study (Given, 2008).
Probability sampling is a preferred method because it eliminates any element selection bias that may arise from human selection processes. This ensures that all elements in the population are equally weighed in the selection process and they also have an equal chance of being selected in the process. Consequently, there is a higher probability that the representative sample of the population accurately represents the entire population. In addition, margin of error rates can be estimated with probability sampling.
Quantitative Research in Criminal Justice
The purpose of quantitative research is to collect relevant statistical data for various purposes, depending on the field being researched such as criminal justice. Criminal justice research is used to obtain information on data for exploration purposes, descriptive purposes, explanation purposes, or application purposes according to Maxfield & Babbie (2008).
Exploratory Research in Criminal Justice
There are various problems and issues involved with criminal justice that researchers explore in order to find solutions to the problems. Exploration is a common method of seeking answers to questions. For example, a researcher may want to ascertain how a new corrections approach implemented in one state is fairing in another state. If the new approach is successful in the state in which it was implemented, the assumption is perhaps the approach would be just as successful in other states. This type of exploratory research would collect data based on results of the newly implemented procedures and use that data to predict future trends and also to initiate changes if necessary.
Another example is estimating the prevalence of drug use in the United States. Quantitative research methods, in this regard, might involve conducting a random sampling of various groups of people in the United States such as students, working adults, non-working adults, etc. In addition to this, statistical reports such as public records and census reports can be used to gather information for the research as well. Exploratory questions asked in the research may pertain to drugs sales, drug arrests, drug use, addiction reports from rehab centers, and reports from the various subset groups via anonymous surveying methods (Maxfield & Babbie, 2008).
Descriptive Research in Criminal Justice
According to Maxfield & Babbie (2008), criminal justice research studies are usually based on detailed descriptions of a criminal issue or policy guidelines in relation to criminal issues. Observation is used in the research and then descriptive responses are made from that. This means that people will observe events and report about those events by describing what they perceived from their observations. Descriptive studies in criminal justice count observations reported from what people have witnessed or perceived, as opposed to exploratory research that seeks to understand the nature of criminality problems.
According to the FBI’s Uniform Crime Reports (UCR), data about criminal activity is reported to the media on a regular basis. The data in these reports include information such as number of murders in a given city in a given year, or the number of auto theft cases in a city. For example, uses for this type of data could include research and gathering information about crime in a specific neighborhood to assist in organizing crime watch groups for the neighborhood, and based on observations from this effort, other neighborhoods could follow suit and organize crime watch groups as well (Maxfield & Babbie, 2008). This would be facilitated by observations made in the neighborhood.
Explanatory Research in Criminal Justice
Criminal justice research is used to explain phenomena in the area of criminal justice such as why people in urban areas are not fond of the police. This question posed for research would be the basis for the sampling. This type of research helps researchers draw conclusions on various aspects of specific criteria involved in the criminal justice system. This type of research is also instrumental in assessing certain situations related to crime and relating them to sampling criteria (Maxfield & Babbie, 2008).
Applied Research in Criminal Justice
Applications in criminal justice include research that evaluates and analyzes problems. Applied criminal justice research can evaluate specific justice programs such as analyzing a new police policy neighborhood program to reduce break-ins, to see if the program actually worked. This is an applied research technique. Applied criminal justice research compares the intent of the program with the actual outcome of the program results.
The second type of applied research is problem analysis and it pertains to using data to predict future trends and events, in an attempt to anticipate what may happen in the future relative to the past. In criminal justice, problem analysis is helpful for studying cases and their patterns of outcomes so that similar cases in the future may have a guideline for law enforcement to go by to initiate responses to recurring problems (Maxfield & Babbie, 2008).
The following review of literature highlights key points from gathered data that are relevant to developing a sampling plan and also to the usefulness of sampling as applied in quantitative research study in the field of criminal justice.
Ahmed, S. (2009). Simple Random Sampling, Systematic Sampling: Lecture 2. Johns Hopkins University Bloomberg School of Public Health, Biostatistics Department: School of Hygiene and Public Health.
This literature is a lecture on methods in sample surveys by professor Saifuddin Ahmed, MBBS, Ph.D. at Johns Hopkins University. This is his second lecture and it covers simple random sampling and systematic sampling. Dr. Ahmed begins with an explanation of the sampling procedure process in which he defines sampling and stresses that randomization in the sampling process is important to eliminating possible selectivity bias in the selection data. This is significant because any bias would render the entire population data unusable for accuracy. With randomization, researchers are able to get a better view of the a population by having a more accurate representative sample to start with.
The literature further states that each element in a simple random sample is unique and has an equal probability of being selected for the research. In addition, there are two types of single random samples and they are with or without replacement of units in the sampling process. This is significant, as an example, not interviewig the same individual twice would facilitate the need for the simple random sampling with replacement.
As it relates to systematic sampling, Dr. Ahmed reviews advantages and disadvantages. Advantages of using systematic sampling methods include better random distribution than that with simple random sampling and easier implementation. A disadvantage of systematic sampling is noted to be cyclic variations which would render the sample inaccurate. For example, as it pertains to criminal justice, some crimes happen in populations many times over again and this would mean that the systematic approach to sampling would not be the best choice.
Babble, E. R. (2011). The Basics of Social Research (5th ed.). Belmont, CA: Wadsworth Publishing.
As criminal justice relates to social research, this literature is significant in exploring the meaning of social research and how it relates to human inquiry. Human inquiry is the basis for research but this literature points out the error potential in human inquiry. For example, it is stated that inquiry is inaccurate when people depend on authority figures to tell them everything and then they believe what they hear only because someone in an authority position stated it. This leaves room for inaccurate observations from human viewpoints without scientific input. This is why computer-generated random sampling is important. It reduces the possibility of error in sample selection because it has no selectivity bias potential.
The literature further states that devices of measurement assist in the use of inaccurate observations in analyses. This is important with sampling, particularly in the criminal justice research realm, because probability sampling is used and with this quantitative analysis is also used. The literature states that quantitative data makes observations more plausible, and it helps to make gathering and summarzing data easier. It also creates statistical data for analyses.
Kotler, P., & Keller, K. L. (2008). Marketing Management (13th ed.). Upper Saddle
River, NJ: Prentice Hall.
As it relates to the subject of this report, this literature defines how a researcher creates a research plan based on a selected research approach, and points out that designing the research plan involves selecting data sources, the research approach, and instruments to be used for the research. It also involves creating the sampling plan which can include both primary and secondary data sources. Kotler also outlines the use of the sampling units and sample size to determine the target population and what parts of the population is to be sampled.
In criminal justice, Kotler touches on how research is often specific in its findings. In contrast to this, he also explains that some research is exploratory, descriptive or casual. He states that exploratory research highlights the nature of a problem and allows researchers to suggest possible solutions to the problem, or to come up with new ideas to eliminate the problem.
Descriptive research has the purpose of determining certain scales such as how many people would pay a quarter to use a pay-toilet in an airport. Casual research gets it results from testing cause and effect relationships. For example, a question may be posed of how many people would pay the quarter to use the pay-toilet if it were closer than the free toilets in the airport.
Leedy, P. D., & Ormrod, J. E. (2005). Practical Research: Planning and Design.
Upper Saddle River, NJ: Prentice Hall.
This literature goes in-depth into the fundamentals of research and the meaning of research and how it is used. As it relates to the subject matter of this report, research is essential to criminal justice inquiry and probability sampling.
This literature also explains theory as a means to explain specific phenomenon through an organized body of conceptualized principles. Theories are tentative explanations that a researcher can modify to account for data findings or offer alternative explanations for the theory. Additionally, when researchers have theories that help explain specific phenomenon, it serves as a springboard for further research to find answers to new questions and suggesting alternative hypotheses about potential outcomes of specific investigations, based on assumptions. Regarding specific phenomenon, it is necessary for the phenomenon to be predictable and not completely randomized events, and also cause and effect relationships can predict patterns observed in the phenomenon. This is the basis for assumptions.
As it pertains to research, data must be collected and interpreted so that findings of any analyses can help researchers resolve problems for which the research was proposed. For example, studying patterns of crime waves in a community requires that statistical data and past criminal history patterns be analyzed in order to find possible solutions to change the outcomes related to the crime waves. In addition, data collected from events, observations, calculations, measurements, etc. are self-contained units.
The data is only significant when it is giving meaning by a researcher who analyzes the data. This is how data helps answer questions and solve problems, when researchers interpret the data and give it meaning. However, this means that different interpretations that occur must be taken into account.
Finally, the literature states that the activity of research follows a cyclical process that includes 1) a question or questions posed, 2) a problem stated, 3) the problem divded into subproblems, 4) gathering preliminary data, 5) forming a hypothesis, 6) collecting more systematic data, 7) data is processed, 8) data is interpreted, 9) a discovery is made, 10) a conclusion is reach, and 11) the question is answered or not answered. This completes the cycle.
As it pertains to probability samping, this process is critical to quantitative research for finding solutions to problems in the criminal justice system, in particular.
Maxfield, M. G., & Babbie, E. R. (2008). Basics of Research Methods for Criminal
Justice and Criminology (2nd ed.). Wadsworth Publishing.
This literature gives in-depth information on how research methods are used in the fields of criminal justice and criminology. As it relates to the subject of this report, the literature highlights the importance of scientific inquiry, surveying and sampling technology in criminal justice research.
Additionally, as stated previously, sometimes making blind assumptions and having blind beliefs in something than people may do, just because someone in authority says it is so can cause inaccurate observations and can skew research findings. This literature points this out as well. Authority can err and criminal justice research can sometimes have errors in the research findings and this can lead to problems in results of any research. The errors associated with this are inaccurate observation, overgeneralization, selectice observation and illogical reasoning.
Criminal justice research serves specific purposes and keeping errors to a minimum is essential to explaining critical associations that have to do with exploration, description and application, as necessary in quantitative research, explained earlier in this report. In addition, thorough knowledge of research methods and principles is important for successful criminal justice professionals. This literature makes remarks about various new technological methods of doing survey research and this is something that can enhance research initiatives, particularly in the field of criminal justice.
Criminal justice research is best carried out through the use of probability sampling, both simple random and systematic. Moreover, probability sampling is based on quantitative analysis methods and techniques that are best used in data selection processes to analyze observations within a population. This is important for researchers so that they can draw pertinent conclusions and help answer questions to solve problems. This is done by gathering representative sample data to assist in making assumptions about a population so that probable outcomes can be inferred about the population to predict future trends and events.
To facilitate the research initiatives, a sampling plan is required which includes a sampling frame, sampling unit, the sample size, and information about the target population. For example, a specific research population may be a school and the sampling units are the students. The research question may look for percentage of students of the school who smoke marijuana. Sampling is a useful tool in this regard because the data can be measured quantitatively.
In conclusion, as it relates to the purpose of this report, it evaluates the usefulness of sampling as applied in a quantitative research study in the field of criminal justice, as well as gives an in-depth look into probability sampling methods and techniques in the area of research.
Ahmed, S. (2009). Simple Random Sampling, Systematic Sampling: Lecture 2. Johns Hopkins University Bloomberg School of Public Health, Biostatistics Department: School of Hygiene and Public Health.
Babble, E. R. (2011). The Basics of Social Research (5th ed.). Belmont, CA: Wadsworth Publishing.
Basic Survey Design. (n.d.). (Australian Bureau of Statistics) Retrieved from National Statistial Service: http://www.nss.gov.au/nss/home.nsf/SurveyDesignDoc/B0D9A40C6B27487BCA2571AB002479FE?OpenDocument
Castillo, J. J. (2009). Systematic Sampling. Retrieved from Explorable: http://explorable.com/systematic-sampling.html
Cox, B. G., & Lavrakas, P. J. (2012). Encyclopedia of Survey Research Methods: Target Population. SAGE Journals.
Given, L. M. (2008). The Sage encyclopedia of qualitative research methods. Los Angeles, CA: Sage Publications.
Herek, G. M. (2012). A Brief Introduction to Sampling . (The Regents of the University of California, Davis Campus) Retrieved from UCDavis: http://psychology.ucdavis.edu/rainbow/html/fact_sample.html
Kotler, P., & Keller, K. L. (2008). Marketing Management (13th ed.). Upper Saddle River, NJ: Prentice Hall.
Leedy, P. D., & Ormrod, J. E. (2005). Practical Research: Planning and Design. Upper Saddle River, NJ: Prentice Hall.
Maxfield, M. G., & Babbie, E. R. (2008). Basics of Research Methods for Criminal Justice and Criminology (2nd ed.). Wadsworth Publishing.
Quantitative Research. (n.d.). Retrieved from Business Dictionary online: http://www.businessdictionary.com/definition/quantitative-research.html
Sampling Units. (n.d.). Retrieved from Business Dictionary online: http://www.businessdictionary.com/definition/sampling-units.html
Stratified Random Sampling. (n.d.). Retrieved from Investopedia: http://www.investopedia.com/terms/stratified_random_sampling.asp#axzz2JC2MdEKK
Teddlie, C., & Yu, F. (2007). Mixed Methods Sampling: A Typology With Examples. Journal of Mixed Methods Research, 1(1), 77.