Non-probability Samples in Management Research - Essay Example

Critical Evaluation of the use of Non-probability Samples in Management Research I. Introduction The primary purpose of sampling is to help a researcher obtain a fairly accurate representation of the entire population (Rubin & Babbie 2011, pg08). The researcher can use the sample to make accurate and precise inferences, as well as deductions about the population. Non-probability samples are increasingly gaining prominence in business and management research. A non-probability sample refers to a sample or sampling technique that is not based on the methods of random selection. On the other hand, probability samples are based on random selection methods. Hence, all individuals have equal chances of being considered for selection in probability sampling. The outcomes of probability sampling are more likely to give a fairly accurate representation or reflection of the entire population. It is imperative that researchers consider the availability, time, cost, and the subject to research about when choosing a sampling technique (Rubin & Babbie 2011, pg12). Probability sampling has gained vast popularity among scholars in both fields of quantitative and qualitative study. Overall, researchers or project managers would employ non-probability in the survey while holding on to a basic assumption that the entire population has evenly distributed characteristics. In this case, non-probability samples would be relevant in generating accurate results, as well as inferences about the population under study. Even though non-probability samples are ineffective for generalizations of results about entire population, they are highly beneficial when the researcher faces workforce constraints, inadequate funding, accessibility problems, and limited time. Definition of non-probability sample and probability samples A non-probability sample refers to a sample or sampling technique that is not based on the methods of random selection. On the contrary, probability samples are based on random selection techniques. All individuals or subjects in a probability sample have equal chances of being considered for selection during probability sampling. Main benefits and limitations of using non-probability samples Benefits First, non-probability samples are valuable in circumstances where only the sample units that are conveniently and easily accessed. Secondly, non-probability samples enable researcher to generate ideas and get constructive feedback. A typical case is when a project manager uses quota samples (females and males) to generate important ideas and obtain pertinent feedback. Thirdly, non-probability sampling is less costly and more convenient (Rubin & Babbie 2011, pg31).The sample is widely applicable in situations where the researcher wants to generate ideas through sampling but lacks adequate funding to undertake a more comprehensive study of the entire population. As a result, the researcher merely generates valuable ideas about the population under study but does not need to draw inferences and deductions that emulate the entire population. It also helps many researchers to conduct a pilot, exploratory or qualitative study. Lastly, non-probability samples are highly beneficial when the researcher faces workforce constraints, inadequate funding, and limited time (Oakshott 2009, pg25). Limitations Non-probability samples are not founded on random selection methods. Thus, they contribute to unequal, imbalanced representation of the population under study. In addition, it is difficult to generalize results or outcomes based on non-probability samples. Moreover, the researcher cannot use non-probability sampling technique (or non-probability samples) to generalize his results about the entire population under survey. Incidentally, the confidence level is relatively low when non-probability samples are used to represent the characteristics/traits of an entire population (CHAPMAN, HOPWOOD, & SHIELDS 2006, pg115) I. Literature Review Most researchers and project managers are bounded by various drawbacks and restrictions, including limited workforce, time, and money. Therefore, it is normally difficult to use random selection methods to sample the entire population. In light of these constraints, most of them end up using non-probability sampling techniques (Oakshott 2009, pg48). Unlike probability sampling, non-probability sampling is not characterized by a randomized selection process. Accessibility of the subjects/participants is the basis of selection in non-probability sampling. Since a non-probability sample may or may not accurately reflect, or represent the entire population, the results cannot be applied in generalizations. According to Emily and Roger (2010, pg54), non-probability sampling is crucial when the researcher has limited resources or simply unable to identify members of the population. For instance, a researcher cannot accumulate enough resources to conduct a probability research about underworld criminals. It may also apply when a researcher is carrying out an exploratory study. Description of various types of non-probability samples The major types of non-probability samples include convenience samples, consecutive samples, quota samples, judgmental samples, and snowball samples (Oakshott 2009, pg63). A researcher would consider convenience samples for selection primarily because they are easily accessible to him. For instance, a researcher would choose subjects merely because they are relatively easy to recruit. Convenience sampling technique is arguably the cheapest, least-time consuming, and the easiest. On the other hand, consecutive samples are similar to convenience samples except that all subjects the researcher can access constitute the sample. As a selection technique, consecutive sampling is arguably the finest of all the sampling techniques that do not involve randomized selection. Apparently, consecutive sampling includes all accessible units/subjects thus making the sample more representative of the entire population (Bryman and Emma 2011, pg45). Meanwhile, quota samples ensure proportionate or equal representation of the subjects, dependent on the characteristic being considered the basis of a particular quota. For instance, a researcher may consider age, education, religion, race, or gender as the basis of the quota when selecting 100 college students for non-probability sampling. For purposes of quota sampling, he must select 25 students from each college level. Judgmental sampling (also known as purposive sampling) is where a researcher chooses subjects to be part or portion of the sample, but with a specific idea in mind. In this regard, the researcher normally considers certain subjects/individuals more fit for the survey than others. Accordingly, the researcher carefully and purposively chooses the subjects to be part of the sample population under study. Finally, snowball samples represent incredibly small population size. The researcher will obtain a snowball sample after asking the initial subjects or participants to identify any potential subject who equally meet the criteria of the survey. Snowball samples are also barely representative of the population. II. A critical discussion of the circumstances under which Non-probability Samples are most appropriate Non-profitability samples are widely pertinent in situations where there is no exhaustive list of population accessible. For instance, the management researcher may find it difficult to obtain or select certain subjects (TAYLOR, SINHA & GHOSHAL 2006, pg89). As a result, he has limited options or means of knowing the size as well as effects of the sampling error. Since non-probability samples are not based on random selection methods, they contribute to unequal, imbalanced representation and reflection of the population under study. Similarly, it is normally difficult to generalize results or outcomes based on non-probability samples (Bryman and Emma 2011, pg59). Non-probability samples may be valuable in circumstances where the researcher intends to generate ideas and get constructive feedback. For instance, the management researcher may use quota samples (females and males) to generate important ideas and obtain relevant feedback. However, the researcher cannot use non-probability sampling technique (or non-probability samples) to generalize his results to the entire population (FOX, & BAYAT 2007, pg78). Apparently, the confidence level is comparatively low when non-probability samples are used to represent or reflect the characteristics/traits of an entire population. Scholars should take into consideration the basic information and concepts about survey sampling techniques or designs. In particular, management researchers should note the difference between probability and non-probability samples so that they get to understand the merits and demits of each approach. Researchers should particularly consider whether or not the sample to be used for management research is reflective of the entire population (Bryman and Emma 2011, pg62). Non-probability sample is widely applicable in situations where the researcher intends to generate ideas through sampling but lacks enough funds to undertake a more comprehensive study of the entire population. In this regard, the researcher merely generates valuable ideas about the population under study but does not need to draw inferences and deductions that reflect the entire population (Buckingham and Peter 2008, pg28). In addition, non-probability samples are more expedient and convenient compared to probability samples. Nevertheless, non-probability samples often generate an incomplete sampling frame. They do not have contact information and data about the entire population. Similarly, all units or persons have unequal chances for selection during the survey. In most instances, the results of a non-probability sampling comprise information or data in which the researcher has utterly and deliberately missed several portions of the audience/participants. Moreover, non-probability sampling is pertinent in situations where the project manager or management researcher simply desire to make descriptive comments about the selected portion of the population. Non-probability samples are also commonly applicable in applied social research, particularly if it is impractical to carry out probability sampling. A typical case in point is Statistics Canada. Apparently, Statistics Canada often employs non-probability sampling for certain preliminary studies during a survey’s development stage, as well as for questionnaire testing (Buckingham and Peter 2008, pg41). III. Implications of using non-probability samples for research design and presentation of findings Non-probability samples are vital for research design and presentation of findings. However, these samples are only useful if accurate representation and generalizations about an entire population are less emphasized in the design and presentation. Since most of the non-probability samples are normally accidental or haphazard, they are valuable in circumstances where only the sample units that are conveniently and easily accessed would be considered for selection during a survey. For example, a food critic may try different entrees or appetizers to judge or evaluate the quality, as well as a variety of the menu. Although the sampling approach is largely characterized by bias, it can deliver a fairly accurate result in scenarios where the population under study is homogenous (Buckingham and Peter 2008, pg55). For instance, a scientist could apply non-probability samples to determine the rate or level of pollution in the lake. The primary assumption would be that any sample used in the survey would yield similar results, and the water is well-mixed. He could draw water from any point without bothering about whether or not the sample (water drawn from the lake) is representative. In the same way, a business or management researcher could consider the first 200 customers entering a department store for a survey. The researcher would obviously make major assumptions about the gender, physical attributes/qualities, marital status, age, and occupation when conducting non-probability survey. IV. Conclusion In summary, non-probability samples do not represent or reflect the entire population because individuals/participants in the sample population are not accorded equal chances or prospects of being selected. However, most researchers end up using non-probability samples due to workforce constraints, inadequate funding, and limited time. Therefore, non-probability samples are a valuable resource in situations where probability sampling or random selection of the entire population is impossible. It is imperative to reiterate that a non-probability sample is not an outcome (or product) of a random selection process (Rubin & Babbie 2011, pg15). The individuals or units in non-probability samples are normally selected based on the accessibility and convenience rather than their potential to represent the entire population. Similarly, the researcher’s purposive personal view or judgment of the units/subjects/individuals may influence his or her selection of a non-probability sample. Despite the researcher’s assumptions about the representative nature of the subjects/individuals selected, the sample may or may not represent the entire population perfectly and accurately (Bryman and Emma 2011, pg71). Hence, most of the non-probability samples cannot be considered when generalizing the entire population under study. In the same way, the researcher’s inferences from non-probability samples cannot accurately describe the actual characteristics of the population. However, non-probability samples are fairly accurate and representative in management research when the population is homogeneous. The different types of non-probability samples and sampling techniques pertinent to management research include convenience samples/sampling, consecutive samples, quota samples, judgmental samples, and snowball samples. Non-probability samples are useful when a researcher or project manager primarily wants to demonstrate that a particular characteristic or trait exists in the population under survey. Non-probability samples may also be used when the management researcher intends to conduct a pilot, exploratory or qualitative study. Similarly, the samples may be used when the population under study is almost boundless (ZIKMUND & BABIN 2007, pg91). It is impractical to use randomization or probability samples when the population is limitless. Since most surveys require massive funding to accomplish, non-probability samples cut down the typical survey expenditures. Apart from the convenience and cost-effectiveness, a non-probability sample is useful in situations where the researcher does not necessarily need to generate results or outcomes for purposes of generalization. A limited workforce, time, and budget may compel a researcher to use non-probability samples/sampling techniques. Lastly, non-probability technique can be used at the initial stages of the survey before finally conducting the study using a randomized, probability sampling techniques. References Adler, E. S., & Clark, R. (2008). How it's done: an invitation to social research. Belmont, CA, Thomson/Wadsworth. 12 Bryman, Alan and Emma Bell (2011). Business Research Methods, 3rd Edition, Oxford University Press. www.oxfordtextbooks.co.uk/orc/brymanbrm3e Buckingham, Alan and Peter Saunders (2008). The Survey Methods Workbook, Cambridge, Polity Press. (Chapters 1-5) CHAPMAN, C. S., HOPWOOD, A. G., & SHIELDS, M. D. (2006). Handbooks of Management Accounting Research, Volume 1. Burlington, Elsevier. http://public.eblib.com/choice/publicfullrecord.aspx?p=285762. FOX, W., & BAYAT, M. S. (2007). A guide to managing research. Cape Town, Juta. Lee, N. & Lings, I. (2008). Doing business research. London: Sage. Oakshott,L. (2009). Essential Quantitative Methods for Business, Management and Finance. 4th Edition. Palgrave Macmillan. RAO, A. B. (n.d.). Research methodology for management and social sciences. [S.l.], Excel Books. Rubin, A., & Babbie, E. R. (2011). Research methods for social work. Belmont, CA, Brooks/Cole Cengage. 355 TAYLOR, B., SINHA, G., & GHOSHAL, T. (2006). Research methodology A guide for researchers in management and social science. New Delhi, Prentice-Hall of India. ZIKMUND, W. G., & BABIN, B. J. (2007). Exploring marketing research. Mason, Ohio, Thomson South-Western. Read More