Words: 2197
To this end, the assignment will Investigate the nature of cluster sampling y finding a strategic fit between theoretical principles and practical applications Witt regard to cluster sampling. I en mall adjective AT ten selectmen Is to exeunt understanding towards cluster sampling by defining what a cluster sample is while further highlighting the advantages and disadvantages of cluster sampling. In addition, the assignment will select a research study that utilized the cluster sampling method.
This section of the assignment will critically and practically analyses why a cluster sampling structure was used for the study while comparing this method to other possible sampling methods. This section of the assignment will further suggest possible flaws and key success factors involving the research stud’s sampling technique. The assignment was referenced and based primarily on secondary data sources such as academic textbooks, the internet and Journal articles.
All the sources used are not older than ten years which means that no resources exceeding the year 2003 were used. The data collection method was based on the fact that the information could be collected in a cost effective manner, the information is easy to access and swift to access. This decision was supported by the mime barriers experienced as a result of the third year marketing students’ academic schedule. However, the assignment fashioned some challenges.
One of the challenges experienced was the lack of relevant and current academic textbooks. This was overcome by making use of the electronic sources such as the internet in the form of Google Books, Google Scholar and credible websites. Furthermore, finding cluster sample studies proved challenging, however, this challenge was overcome by selecting a suggested research study from the third year Marketing Research study guide. 2 CLUSTER SAMPLING
McDaniel and Gates (2008:343) as well as Molar (2004:328) agree that cluster sampling is when one selects a sub-group of a targeted population that is divided into mutually exclusive and collectively exhaustive groups called clusters. After the clusters have been identified the clusters are selected in two ways. The samples can either be selected probabilistically which means that not all of the clusters will be included in the survey process or all of the clusters will feature in the survey procedure (McDaniel & Gates 2008:343).
According to Molar (2004:329) cluster sampling can be categorized into three main groups. Molar (2004:329) claims that cluster samples can be categorized into one stage sampling, two stage sampling and multi-stage sampling. Single stage sampling occurs when one lists all the clusters in the population and from that list clusters are selected (Aimed 2009:3). The clusters are then selected by using the simple random selection method where all the sample elements are included in the sample (Aimed 2009:3).
This means that every sample element and unit has an equal chance of being selected for the survey process. For example, all possible Commensurate homes and home owners have an equal chance f being selected in the Commensurate cluster. In contrast, two stage cluster sampling occurs when one lists all the clusters in a population from which one selects the specific clusters in the first stage (Molar 2004:329). From these clusters the interviewers will survey the sample elements that fall within the first stage’s chosen clusters (Molar 2004:329).
Moreover, the selection of the clusters can be conducted by using a simple random selection approach or by using a systematic selection approach (A ) For example, a researcher may randomly select Port Elizabeth suburbs and then use systematic sampling to select which households ND residential blocks will be surveyed. Moreover, multi – stage cluster sampling is when the sample is selected in different stages (Wide & Diggings 2009:208). For example, one can divide the sample into different clusters consisting of different elements and then one can randomly select a representative sample (Wild & Diggings 2009:208).
Moreover, cluster sampling is said to either overrepresented or underrepresented the sample and thus the data collected may be skewed (Castillo 2009:49). To narrow down the negative effects, one should select many small clusters instead of selecting a few large clusters. In the author’s opinion, this may have the effect of producing a more representative sample and narrow the effect of obtaining similar results because of similar living conditions from similar clusters.
STUDY This section of the assignment will attempt to explain and suggest logical reasons as to why the research study entitled “Student recruitment marketing in South Africa – An exploratory study into the adoption of a relationship orientation” by Bennie and Human (2010:435-477) utilized the cluster sampling method. This section of the assignment will further propose suggestions and evaluations by cross referencing similar studies as well as comparing the advantages of cluster sampling with the disadvantages of the remaining seven other sampling approaches. Moreover, the Bennie and Human study can be referred to in Annexed
A. Firstly, the relevant study investigated by Bennie and Human (2010:435-447) o en ten qualitative data In ten study Day enumerating eleven experience interviews in which several student recruitment marketers and administrators located at higher education institutes were interviewed throughout South Africa to avoid region bias (Bennie & Human 2010:439). Furthermore, 895 first and second year students from higher education institutes were interviewed from the Free State, the Western Cape and Sautéing which formed part of the quantitative stage of the study (Bennie & Human 2010:439).
The reason for selecting higher education first and second year students was based on the fact that high schools were not willing to participate in the study and higher education students have experienced higher education decision making (Bennie & Human 2010:439). Moreover, these respondents were chosen because the respondents were familiar with the higher education application process as well as higher education recruitment efforts (Bennie & Human 2010:439).
This ultimately means that higher education students have a more realistic and holistic view towards the relationships between high school and higher education recruitment. Furthermore, the researchers surmised that respondents who have a high involvement with the topic would be more likely to participate in the study (Bennie & Human 2010:439). The study then utilized cluster sampling by targeting the survey at classes within various faculties at the different higher education institutions in the Free State, Western Cape and Sautéing (Bennie & Human 201 (“439). The main advantage of using cluster sampling for the previously mentioned study is based on the premise that the researchers can have a larger ample with greater reductions in financial means and resources. This is because interviewers can interview several respondents at once because entire classes are surveyed at one specific higher educational institute at a time. To this end, a fairly large sample can be used while visiting less higher education institutes.
This will ultimately substantially shorten the time it takes to complete the survey provided the survey is administered in a time efficient manner and the survey provides the benefit of lower costs per respondent. These benefits would be most effective if classes completed the questionnaires all at once. Furthermore, because fewer institutes will be surveyed fewer interviewers will be required which once again reduces the cost of the study.
Likewise, because higher education institutes generally have more than one campus in close proximity of each other, interviewers will be relatively close to the sample for each higher education institute they survey in the respective provinces. Supportively, Rust et al (2004:24) argue that cluster samples result in significant cost reductions with respect to training data collectors, the standardization of data collection methods and allow provision for quality control ND analysis. All of which would prove to be advantageous for the Bennie and Human study. .Whilst the previous section of the assignment investigated why it would be beneficial to use cluster samples for the Bennie and Human study this section of the assignment will provide suggestions as to why the other seven sampling methods were not utilized. Moreover, the knowledge learned and observed from the disadvantages of the seven other sampling methods will assist in Justifying the use of cluster sampling in the Bennie and Human study.
The study did not tells ten snowball sampling approach primarily Decease ten respondents were easy to locate and define because of the nature of the study whereas if the study had to find respondents who purchase second hand home appliances a snowball sample would be appropriate. Secondly, the snowball method was not executed because the approach is time consuming and difficult to control. This then would lead to significantly higher costs when conducting the study. In support of this, Molar (2004:331) and McDaniel and Gates (2008:347) both agree that snowball sampling is mime consuming and respondents are difficult to locate.
Moreover, in the author’s opinion the study should not make use of convenience sampling or Judgment sampling. This is cemented on the fact that convenience sampling is not fully representative of the population; this technique is notorious for large sample errors and it is difficult to generalize the results to the wider population (Shall 2008:62). Likewise, this approach is not beneficial for descriptive or causal research and there is a high degree of respondent selection bias (Molar 2004:331). In addition, it is in he best interests of the researchers to not have selected Judgment sampling.
This is based on the notion that the method has the tendency to be subjective to the end of expert opinions and as a result does not allow for the results to be generalized to the wider population (Molar 2004:331). Furthermore, Wild and Diggings (2009:200) suggest that Judgment samples are most appropriate when large samples are not selected. As it can be seen in the Bennie and Human study the sample size is significantly large (895 respondents). This then has the effect of ruling out the judgment sample approach for the study.
Once again the use of cluster sampling for the study is more beneficial because the use of quota sampling or stratified sampling, in the author’s opinion, provides too many barriers. In support of this statement, Wild and Diggings (2009:201) and Shall (2008:62) agree that it is difficult to determine which population characteristic or characteristics to use and it is difficult to access this type of information. This then may negatively affect the cost and time taken to complete the study. This is significant based on the fact that the sample consists of 895 respondents.
Similarly, stratified sampling would not be social because of the strata selection barriers. This is based on the argument that relevant stratification variables are difficult to select because of the often limited amount of information on the demographic variables which undesirably leads to negative cost and time implications (Molar 2004:331; McDaniel and Gates 2008:341; Shall 2008:61). In the author’s opinion, research studies should be conducted accurately without adding unnecessary costs and unnecessary difficulties.
Although Stratification and quota sampling have their own set of advantages, stratification and quota sampling are in the author’s opinion considered necessarily complicated for the Bennie and Human study so these methods should not be used. Lastly, the author believes that one of the reasons why Bennie and Human did not choose simple random sampling and systematic sampling is because the target population sample frames were difficult to obtain and as with any marketing research study it is essential that the sample is representative of the population.
In the same breath, Molar (2004:331) and Shall (2008:60) seemingly agree with the notion that systematic sampling has the tendency to reduce the representatives of the sample while McDaniel and Gates (2008:340) and Shall PUB:U) are adamant Tanat enlarge patterns In ten data may present themselves Ana contribute to a less representative sample. Furthermore, simple random sampling should not be considered because simple random sampling is generally more expensive than other methods because sampling frames have to be sourced (Molar 2004:331).
In the case of the Bennie and Human study it would be easier, less time consuming and appropriate to use a cluster sample because purchasing student numbers and information may be extremely time consuming because of the large population of students nationally. Furthermore, some higher education institutes may be reluctant to provide such information to the researchers or the researchers would have to have ethical clearance which ultimately means that the study would have more barriers and restrictions. 4 CONCLUSION In conclusion, it seems apparent that cluster sampling would be the ideal sampling method to be used.
This is based on the fact that other sampling methods encourage unnecessary costs, do not accurately represent the population, encourage selection bias and barriers or elicited subjective opinions that cause lower levels of precision. Furthermore, because higher education institutes may restrict sample frames and significantly prolong the research process the author of the assignment highly supports and recommends the use of cluster sampling in the Bennie and Human study entitled, “Student recruitment marketing in South Africa – An exploratory study into the adoption of a relationship orientation. This recommendation is further supported by positive cluster sampling variables such as cost efficient trends, the lack of demanded resources and the close geographic proximity of the clusters.