Non-Probability Sampling in Management Research - Coursework Example

Non-probability Sampling in Management Research Student’s Name Code+ course name Instructor’s Name University Name City, State Date Table Contents Table Contents 2 Introduction 3 Benefits and limitations of non-probability samples 4 Convenience sampling 4 Purposive sampling 5 Snowball sampling technique 5 Quota Sampling 6 Application of non-probability samples 7 Implications of non-probability samples 8 Misrepresentation 8 Cost implications 9 Degree of sampling error 9 Challenges of conducting a test on statistical significance 9 Bias 10 Conclusion 10 Reference List 11 Non-probability Sampling in Management Research Introduction Sampling techniques are essential tools whenever students or researchers intend to collect data for research. In research methodology, a researcher has to identify the type of data needed, the type-desired outcome of the results and the technique to be used to collect the data. Data collection may require categorization that is often referred to as sampling. By definition, a sample is a sub-collection of a population or a representative of an entire target population. Two sampling techniques exist that are probability sampling and the non-probability sampling. The probability sampling is at times referred to as random as a selection of the samples and analysis of outcomes relies on chance. Non-probability sampling technique does not consider or rely on chance during the selection of samples. The term probability is used to refer to the allocation of opportunity during the selection of samples. Therefore, where samples from the study population are not selected on a chance basis, the method is referred to as non-probability sampling. Where the samples from the study population are selected on a chance basis, the technique is referred to as probability sampling (Latham, 2007). There are principles of non-probability samples that in some way represent the advantages of this sampling technique. Other general advantages of non-probability samples include their applicability in exploratory studies, studies where a sampling frame is missing, studies where there is wide spread population that limits use of cluster sampling (Fan, 2013). The general challenges include high risks of bias and misrepresentation of the study population. There is also a risk of committing statistical errors like under coverage and non-sampling This paper discuses the application of non-probability sampling technique, its benefits and limitations in research methodology. Benefits and limitations of non-probability samples Despite the fact that many study use probability sampling, not all type of researches can fully benefit from it. Some researches like social surveys are effectively conducted with non-probability design (Babbie, 2013). According to Battaglia (2011), there are various types of non-probability sampling techniques that applicable to various research methodologies. To understand the benefits and limitations of non-probability samples, it is important to establish the different types of non-probability sampling techniques that exist. Four types of non-probability sampling that exist include quota sampling, accidental sampling, purposive sampling, and expert sampling and snowball sampling. The overall advantage of non-probability sampling are; cost convenience for studies where representative samples are not mandatory; It is a useful technique for pilot studies and more specifically when sensitive information is needed for a study. The general disadvantages include failure to select representative for the study population. This section will discuss the benefits and limitations of each type of non-probability sampling techniques. Convenience sampling Convenience sampling is a non-probability sampling technique where samples for study are selected basing on the sample that is readily available. A researcher may find it necessary to make use of easily accessible participants for a study due to time constraints and financial constraints (Battaglia, 2011). The samples, in this case, meet the basic requirements for the study. The primary benefit of the technique is that information samples have high chances to deliver good response for the questionnaire. The method, however, has shortcomings and cannot be efficient in the study of large population. Convenience sampling is vulnerable to bias during the selection of samples thus; results may not portray the actual state on the ground. The research technique may fail to offer adequate representation for the entire study population. Besides, a researcher is likely to miss on useful samples that can provide prime information for the study. Purposive sampling Purposive sampling is a non-probability sampling technique where samples are strictly selected to meet the demands of the research. Some researchers like Battaglia (2011) also refers to this criteria as judgmental. Like the convenience sampling technique, purposive sampling is also very selective to only samples that meet the study. Equally, the method proves to save time and cost spent during data collection and analysis. Purposive samples present a researcher with a high chance for valid and consistent data that ease the process of analysis and interpretation. On the other hand, purposive sampling design requires one to understand the trends within the study population thus limit individuals who lack such useful information. According to Babbie (2013), purposive sampling is also vulnerable to sources of bias during the process of data collection. Snowball sampling technique The snowball sampling design is a method that technically develops a network during the selection of samples for a study. In this method, one sample that meets the study requirements leads the researcher to the next sample. The method is most useful when a research intends to collect sensitive information only known to a few individuals within a study population. With such kind of practice in a methodology, a researcher is sure to obtain relevant and valid information for a study. The study design is also a useful technique for establishing a quick cause-effect relationship. Snowball design has several limitations that can affect the results of the study. For instance if the researcher fails to accurately identify the first sample, he or she may establish a wrong sampling network thus end up with incorrect data. A researcher is likely to rely too much on the network provided, and that can impair the researchers judgment in the identification of samples (Babbie, 2013). The method has high risks of bias during data collection, and one sample can intentionally mislead the researcher. Also, the snowball has no prior planning and identification of samples before the data collection phase. For this reason, a researcher is likely to spend more time in the field to identify samples. There is also uncertainty over the distribution of samples of the population; thus a researcher may find it challenging to develop a budget and a timetable for the research. Snowball technique does not provide proper representation for a study population. Quota Sampling In quota sampling, a researcher selects samples by identification of subgroups within the study population. Specific units are then identified to serve as a subject for the interview. According to Battaglia (2011), there are several challenges that a researcher are likely to phase when using quota sampling. There is a problem of establishing or maintaining accuracy during identification of sub-groups. The quota design has high chances of bias especially in the selection of samples within the identified subgroups. These challenges in quota sampling are inherent and can ultimately affect the results of research especially in cases that require statistical description (Babbie, 2013). Application of non-probability samples Sampling techniques are selected basing on the type of data needed, the characteristics of the study population, time and available budget for the study. Non-probability samples are significant in research and can be used in official statistical agencies (Doherty, 1994). For instance, many business surveys officially conduct researches that adopt the non-probability sampling technique. Such businesses use purposive sampling since it may be hard for the researchers to get respondents willing to take part in the survey. According to Doherty (1994), the principal technique used to determine the consumer price index is the non-probability sampling technique. The method is preferred to probability sampling since it is hard for researchers to a researcher to identify markets where they price. If a researcher may decide to go ahead and determine the markets where they price, the cost of the survey will rise thus accumulating cost inconveniences. Also, more time may also be spent on the identification process thus resulting to time inconveniences. In New Zealand, the nonprobability sampling technique has been approved for use by the Consumer Price Index Advisory Committee, consumer as well as the statistical experts who advise the government (Doherty, 1994). The non-probability technique of quota sampling is applicable when a researcher is likely to lack adequate information of the study population (Battaglia, 2011). For instance, if a country wishes to conduct a survey on the consumer behavior of foreign tourists, quota sampling stand out as the best technique for the methodology. According to Farrokhi and Hamidabad (2012), convenience sampling techniques are mostly used in opinion polls where the researcher may only target a group that will respond to the questionnaire or interview. In cases where there is high chance that individuals may shy away from participating like in political poll survey, convenient sampling will be useful. Convenience may be regarding response, accessibility, cost and time that a researcher considers. Snowball samples useful especially when a researcher is pursuing a sensitive topic that is hidden from the public. Such sensitive cases may include cases of drug abuse or peddling, cases of rape or HIV/AIDS infection, cases of corruption or security intelligence. In most cases, it is difficult to identify these samples due to issues of secrecy, social security or personal preference. In such cases, the only option is to determine one sample for the study that can eventually provide a link for the next sample. Implications of non-probability samples The non-probability sampling techniques can widely be used in social surveys such as opinion polls and product analysis but may have certain implications. The implications depend on the method of non-probability design used to select samples in a social survey. The implications are discussed in this section. Misrepresentation Misrepresentation of the study population is one of the significant limitations linked with non-probability samples (Monette, et al., 2013). The absence of probability in the selection of samples eliminates the possibility of good representation of the study population. There is no way of establishing if the study population is adequately represented in the study population. A researcher cannot pose a claim of a study population representation and defend the claim. For this reason, research that adopts non-probability design cannot generalize the findings of a study. Cost implications Some probability sampling may reduce cost while some may considerably increase the cost of a survey. For instance, the snowball sampling technique may pose a researcher with a serious challenge of establishing the cost of a field research. In such cases, the research can end up spending too much cost during the field research (Monette, et al., 2013). More cost can be incurred in the processes of validation and samples. Degree of sampling error According to Monette, et al., (2013) there is a considerable uncertainty to the extent of sampling error that can occur when using non-probability sampling technique. When using non-probability sampling, there is no clarity of study population representation. Researchers have not so far established an efficient arithmetic formula for determining sampling error in non-probability samples. The methods used to estimate errors, in this case, may not be accurate and reliable (Monette, et al., 2013). For this reason, there is no chance that formulas for establishing study population are not applicable to non-probability sampling. The concept of homogeneity and are often not considered hence the fraction that such data represents might not be easily identified. Challenges of conducting a test on statistical significance After doing research, a researcher has to establish the importance of such findings to the study conducted. Often statistical tools are used to conduct these measures. In probability sampling, the arithmetic approach can be adopted to determine the statistical significance of results. On the other hand, a researcher may face a serious challenge to establish statistical significance in non-probability samples. Some methods, however, exist for statistical significance that requires a researcher to exercise caution when using them. The statistical options that exist rely upon an assumption, and there is a high risk of violating some of the basic assumptions. Bias According to Boslaugh & Watters (2008) non-probability sampling is prone to measurement bias at various stages of a study. Measurement bias can result during identification and retention of samples for a study. With this kind of bias, a researcher is likely to collect unreliable data and eventually bias information and conclusion (Wesberg, 2011). The two major types of bias in this case are; the sample selection and retention bias; and bias in the collected information (Boslaugh & Watters, 2008). Selection bias and volunteer bias are common in the non-probability sampling specifically in the convenience and snowball sampling. There is also a probability that a researcher can lose important follow-up links when adopting methods such as convenience sampling (Wesberg, 2011). Conclusion There are many reasons as to why researches are conducted that may include business purpose, political purposes, medical purposes or education. The purpose of research, the target population, time and budget are important in identifying the process of data collection. Data is collected from samples that are identified using two main criteria, the probability criteria and the non-probability criteria. The probability criteria give a chance of selection to the study population while the non-probability criteria deny an equal chance for sample selection in the study population. Four types of non-probability samples exist that include quota sampling, accidental sampling, purposive sampling, and expert sampling and snowball sampling. The use of samples selected through such method has both benefits and limitation. Some benefits include reduction of cost, time spent usability where there is inadequate information of the study population, and the simplicity involved in identifying samples. Some of the challenges include the high possibility of misrepresentation, uncertainty in the distribution of samples and variety sources of bias in the entire process. The non-probability samples can be applied instances of a social survey such as consumer characteristics, political opinion poll and public health service delivery opinion Reference List Babbie, E., 2013. The Basics of Social Research. 6 ed. New York City:Cengage Learning. Battaglia, M., 2011. Non Probability Sampling: Encyclopedia of Survey Research Methods. Retrieved on November 30th , from Boslaugh, S. & Watters, A. P., 2008. Statistics in a Nutshell: A Desktop Quick Reference. SanFransisco: OReilly Media Inc. Doherty, M., 1994. Probability versus Non-Probability Sampling in Sample Survey. The New Zealand Statistics Review. Retrieved on November 30th , from http://www.nss.gov.au/nss/home.NSF/75427d7291fa0145ca2571340022a2ad/768dd0fbbf616c71ca2571ab002470cd/$FILE/Probability%20versus%20Non%20Probability%20Sa mpling.pdf E, D., 2015. Sampling Techniques in Educational Research. 1st edition ed. s.l.:Lulu.com. Fan, P. D., 2013. Reconceptualizing Survey Representativeness for Evaluating and Using Nonprobability Samples. The premier e-journal resource for the public opinion and survey research community. Retrieved on November 30th , from < http://www.surveypractice.org/index.php/SurveyPractice/article/download/43/pdf> Farrokhi, F. & Hamidabad, M. A., 2012. Rethinking Convenience Sampling: Defining Quality Criteria Retrieved on November 30th , from http://www.academypublication.com/issues/past/tpls/vol02/04/20.pdf Latham, B., 2007. Sampling: What is it? Quantitative research method. Retrieved on November 30th , from http://webpages.acs.ttu.edu/rlatham/Coursework/5377(Quant))/Sampling_Methodology_Paper.pdf Monette, D., Sillivan, T. & Dejong, C., 2013. Applied Social Research: A Tool for the Human Services. 9th Edition ed. New York City:Cengage Learning. Wesberg, H., 2011. Bias and Causation: Models and Judgment for Valid Comparisons. New Jersey: John Wiley & Sons. Read More