"NONPROBABILITY SAMPLING 1NONPROBABILITY SAMPLINGBy Student’s NameCode+ course nameInstructor’s NameUniversity NameCity, StateDate NONPROBABILITY SAMPLING 2ContentsIntroduction ................................................................................................................................................... 3Benefits and limitations of non-probability samples .................................................................................... 4Convenience sampling .............................................................................................................................. 4Purposive sampling ................................................................................................................................... 5Snowball sampling technique ................................................................................................................... 5Quota Sampling ........................................................................................................................................ 6Application of non-probability samples ........................................................................................................ 6Implications of non-probability samples ....................................................................................................... 8Misrepresentation ...................................................................................................................................... 8Cost implications ...................................................................................................................................... 8Degree of sampling error .......................................................................................................................... 9Challenges of conducting a test on statistical significance ....................................................................... 9Bias ........................................................................................................................................................... 9Conclusion .................................................................................................................................................. 10References ................................................................................................................................................... 12NONPROBABILITY SAMPLING 3IntroductionSampling techniques are essential tools whenever students or researchers intend to collectdata for a 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. Datacollection may require categorization that is often referred to as sampling. By definition, asample is a sub-collection of a population or a representative of an entire target population. Twosampling techniques exist that are probability sampling and the non-probability sampling. Theprobability sampling is at times referred to as random as a selection of the samples and analysisof outcomes relies on chance. Non-probability sampling technique does not consider or rely onchance during the selection of samples. The term probability is used to refer to the allocation ofopportunity during the selection of samples. Therefore, where samples from the study populationare not selected on a chance basis, the method is referred to as non-probability sampling. Wherethe samples from the study population are selected on a chance basis, the technique is referred toas probability sampling (Latham, 2007).There are principles of non-probability samples that insome way represent the advantages of this sampling technique. Other general advantages of non- probability samples include their applicability in exploratory studies, studies where a samplingframe is missing, studies where there is wide spread population that limits use of clustersampling (Fan, 2013). The general challenges include high risks of bias and misrepresentation ofthe study population. There is also a risk of committing statistical errors like under coverage andnon-sampling This paper discuses the application of non-probability sampling technique, itsbenefits and limitations in research methodology. NONPROBABILITY SAMPLING 4Benefits and limitations of non-probability samplesDespite the fact that many study use probability sampling, not all type of researches canfully 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. Tounderstand the benefits and limitations of non-probability samples, it is important to establish thedifferent types of non-probability sampling techniques that exist. Four types of non-probabilitysampling that exist include quota sampling, accidental sampling, purposive sampling, and expertsampling and snowball sampling. The overall advantage of non-probability sampling are; costconvenience for studies where representative samples are not mandatory; It is a useful techniquefor pilot studies and more specifically when sensitive information is needed for a study. Thegeneral disadvantages include failure to select representative for the study population. Thissection will discuss the benefits and limitations of each type of non-probability samplingtechniques.Convenience samplingConvenience sampling is a non-probability sampling technique where samples for studyare selected basing on the sample that is readily available. A researcher may find it necessary tomake use of easily accessible participants for a study due to time constraints and financialconstraints (Battaglia, 2011). The samples, in this case, meet the basic requirements for thestudy. The primary benefit of the technique is that information samples have high chances todeliver good response for the questionnaire. The method, however, has shortcomings and cannotbe efficient in the study of large population. Convenience sampling is vulnerable to bias duringthe selection of samples thus; results may not portray the actual state on the ground. The research NONPROBABILITY SAMPLING 5technique may fail to offer adequate representation for the entire study population. Besides, aresearcher is likely to miss on useful samples that can provide prime information for the study.Purposive samplingPurposive sampling is a non-probability sampling technique where samples are strictlyselected to meet the demands of the research. Some researchers like Battaglia (2011) also refersto this criteria as judgmental.Like the convenience sampling technique, purposive sampling isalso very selective to only samples that meet the study. Equally, the method proves to save timeand cost spent during data collection and analysis. Purposive samples present a researcher with ahigh chance for valid and consistent data that ease the process of analysis and interpretation. Onthe other hand, purposive sampling design requires one to understand the trends within the studypopulation 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 techniqueThe snowball sampling design is a method that technically develops a network during theselection of samples for a study. In this method, one sample that meets the study requirementsleads the researcher to the next sample. The method is most useful when a research intends tocollect sensitive information only known to a few individuals within a study population. Withsuch kind of practice in a methodology, a researcher is sure to obtain relevant and validinformation for a study. The study design is also a useful technique for establishing a quickcause-effect relationship. Snowball design has several limitations that can affect the results of thestudy. For instance if the researcher fails to accurately identify the first sample, he or she mayestablish a wrong sampling network thus end up with incorrect data. A researcher is likely to rely NONPROBABILITY SAMPLING 6too much on the network provided, and that can impair the researchers judgment in theidentification of samples (Babbie, 2013). The method has high risks of bias during datacollection, and one sample can intentionally mislead the researcher. Also, the snowball has noprior planning and identification of samples before the data collection phase. For this reason, aresearcher is likely to spend more time in the field to identify samples. There is also uncertaintyover the distribution of samples of the population; thus a researcher may find it challenging todevelop a budget and a timetable for the research. Snowball technique does not provide properrepresentation for a study population.Quota SamplingIn quota sampling, a researcher selects samples by identification of subgroups within thestudy 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 phasewhen using quota sampling. There is a problem of establishing or maintaining accuracy duringidentification of sub-groups. The quota design has high chances of bias especially in theselection of samples within the identified subgroups. These challenges in quota sampling areinherent and can ultimately affect the results of research especially in cases that require statisticaldescription (Babbie, 2013).Application of non-probability samplesSampling techniques are selected basing on the type of data needed, the characteristics ofthe study population, time and available budget for the study. Non-probability samples aresignificant in research and can be used in official statistical agencies (Doherty, 1994). Forinstance, many business surveys officially conduct researches that adopt the non-probability NONPROBABILITY SAMPLING 7sampling technique. Such businesses use purposive sampling since it may be hard for theresearchers 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-probabilitysampling technique. The method is preferred to probability sampling since it is hard forresearchers to a researcher to identify markets where they price. If a researcher may decide to goahead and determine the markets where they price, the cost of the survey will rise thusaccumulating cost inconveniences. Also, more time may also be spent on the identificationprocess thus resulting to time inconveniences. In New Zealand, the nonprobability samplingtechnique 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 likelyto lack adequate information of the study population (Battaglia, 2011). For instance, if a countrywishes to conduct a survey on the consumer behavior of foreign tourists, quota sampling standout 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 mayonly target a group that will respond to the questionnaire or interview. In cases where there ishigh chance that individuals may shy away from participating like in political poll survey,convenient sampling will be useful. Convenience may be regarding response, accessibility, costand time that a researcher considers.Snowball samples useful especially when a researcher is pursuing a sensitive topic that ishidden from the public. Such sensitive cases may include cases of drug abuse or peddling, casesof rape or HIV/AIDSinfection, cases of corruption or security intelligence. In most cases, it is NONPROBABILITY SAMPLING 8difficult 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 eventuallyprovide a link for the next sample.Implications of non-probability samplesThe non-probability sampling techniques can widely be used in social surveys such asopinion polls and product analysis but may have certain implications. The implications dependon the method of non-probability design used to select samples in a social survey. Theimplications are discussed in this section.MisrepresentationMisrepresentation of the study population is one of the significant limitations linked withnon-probability samples (Monette, et al., 2013). The absence of probability in the selection ofsamples eliminates the possibility of good representation of the study population. There is noway of establishing if the study population is adequately represented in the study population. Aresearcher cannot pose a claim of a study population representation and defend the claim. Forthis reason, research that adopts non-probability design cannot generalize the findings of a study.Cost implicationsSome probability sampling may reduce cost while some may considerably increase thecost of a survey. For instance, the snowball sampling technique may pose a researcher with aserious challenge of establishing the cost of a field research. In such cases, the research can endup spending too much cost during the field research (Monette, et al., 2013). More cost can beincurred in the processes of validation and samples. NONPROBABILITY SAMPLING 9Degree of sampling errorAccording to Monette, et al., (2013) there is a considerable uncertainty to the extent ofsampling 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 notso 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 andreliable (Monette, et al., 2013). For this reason, there is no chance that formulas for establishingstudy population are not applicable to non-probability sampling. The concept of homogeneityand are often not considered hence the fraction that such data represents might not be easilyidentified.Challenges of conducting a test on statistical significanceAfter doing research, a researcher has to establish the importance of such findings to thestudy conducted. Often statistical tools are used to conduct these measures. In probabilitysampling, the arithmetic approach can be adopted to determine the statistical significance ofresults. On the other hand, a researcher may face a serious challenge to establish statisticalsignificance in non-probability samples. Some methods, however, exist for statisticalsignificance that requires a researcher to exercise caution when using them. The statisticaloptions that exist rely upon an assumption, and there is a high risk of violating some of the basicassumptions.BiasAccording to Boslaugh & Watters (2008) non-probability sampling is prone to measurement biasat various stages of a study. Measurement bias can result during identification and retention of NONPROBABILITY SAMPLING10samples for a study. With this kind of bias, a researcher is likely to collect unreliable data andeventually bias information and conclusion (Wesberg, 2011). The two major types of bias in thiscase 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 samplingspecifically in the convenience and snowball sampling. There is also a probability that aresearcher can loose important follow-up links when adopting methods such as conveniencesampling (Wesberg, 2011).ConclusionThere are many reasons as to why researches are conducted that may include businesspurpose, political purposes, medical purposes or education. The purpose of research, the targetpopulation, time and budget are important in identifying the process of data collection. Data iscollected from samples that are identified using two main criteria, the probability criteria and thenon-probability criteria. The probability criteria give a chance of selection to the studypopulation while the non-probability criteria deny an equal chance for sample selection in thestudy population. Four types of non-probability samples exist that include quota sampling,accidental sampling, purposive sampling, and expert sampling and snowball sampling. The useof samples selected through such method has both benefits and limitation. Some benefits includereduction of cost, time spent usability where there is inadequate information of the studypopulation, and the simplicity involved in identifying samples. Some of the challenges includethe high possibility of misrepresentation, uncertainty in the distribution of samples and varietysources of bias in the entire process. The non-probability samples can be applied instances of a "