Advantages and Disadvantages of large Sample Size Free Samples

1.In our case being that the population size was sixty nine thousand, a sample size of fifteen thousand represented by over twenty percent covers over one fifth of the population. This sample therefore has a bigger sample size than expected. i.e. 383 bank workers were to make the sample size for this population size (69,000) with confidence level of 95 percent and margin of error of 5 percent. Large sample sizes are therefore associated with advantages and disadvantages. Large sample sizes ensure for the reliability of the sample mean as the estimator of the population parameter. For a sample to fully reflect the correct population mean, larger sample size is always contemplated of. The quantity need to be pinned down. Standard error (Se) of the mean is therefore used to quantify the reflection of population mean. This standard error is essential for all calculated sample means. This is taken as an advantage of the large sample sizes due their wide coverage of the population. Working with large samples is important since it helps in sweeping out the outliers in the sample. Small samples are perceived prone to outliers which may misrepresent the data in the sample. Bigger samples capture greater odds of outliers in the sample. However, in most of the cases, outliers tend to complicate analysis of statistical data but accounting for them help in giving realistic picture and the characteristics of the population. Another advantage of large sample sizes is that they help in obtaining a quality and precise mean. This is so because the mean will have covered many elements of the population. Determination of the mean is important for it help the researchers to do away with the outliers from their data. Outliers in the dataset are important to be dealt with because they totally differ from the mean greatly and may give a deceitful image about the sample or population. Since large sample size is suitable due to its large and wider coverage of the population of study, it is in the same way time consuming and expensive to work with. For instance, sampling 15,000 workers who work in the Belgian bank will require a lot of time and also the expense that will be involved will be high. A lot of time is required since the larger sample size is spread in the manner that the population is spread and thus collecting data from the entire sample will involve much time compared to smaller sample sizes. Due to its wider coverage, the expense that is involved in data collection process is also higher compared to expense that could be incurred in a small sample size. Overrepresentation of population data in a population involves large sample size. Collection of data from this sample size in a well distributive way will require high financial involvement for the success of the process as planned Decision on what sample size to use will depend on the population size i.e. 69,000 bank workers and cost that will be involved in data collection. If the researcher wants to incur low cost in the process, smaller sample size will be preferred. In that case, it will also help in determining how precise we should be with our data. Sampling whole lot of 15,000 Belgian bank workers will mean high cost incurred in the data collection process. Prior information concerning the subject of study will help in determining the sample size for use in the study. This prior information can be considered in deciding whether to reduce the sample size or not. The key elements that will be considered from prior information is the prior mean and variance estimates, this is according to (Moher et al, 1994). Practicality is another factor to be considered when choosing for the sample size. The sample size chosen for use must make sense and practical in real life situation. Margin of error also forms another key factor for it will be relied on in determining how reliable and perfect a sample is. It will be showing the width or interval at which the calculated mean will lie and also help in construction of the confidence interval level. 2.The bank workers who were to be involved in the study were given equal chances of being selected by employing probability sampling methods. The chances will be made in such a way that they are greater than zero; this helps in reducing human biasness that may arise through their judgments thus making the process free and fair for inclusion of all banks and the bank workers in the process (Bacchetti, 2002). Probability sampling method used by the research institutions was stratified sampling method. The research institutions first randomly identified the banks which formed the strata then in the identified banks; they randomly selected the workers for fairness in their selection. Compared to other probability sampling methods like the simple random sampling, stratified sampling method gives more precision of the same sample size. Precision is important in the estimation process of the population parameter, each stratum’s statistic will be calculated and their closeness compared to one another. The process is found to be cost effective as it only involves random selection of different baking institutions and workers over the entire population which makes it half completed because of its precision. It is also flexible in that any number of participants can be selected with ease and efficiency. Also, this process tends to be more effective as it results to accuracy in selection of data since it involves lesser degree of judgment of the researcher. It as well forms easier way of sampling as compared to other sampling methods since it does not involve long and complicated processes. Moreover, probability sampling method does not require any technicality the refore any person can carry it out even non-technical persons. Since it only require random assignment of numbers over the specified strata. This method (stratified probability sampling method) of selecting the sample results to the selection of only specific class of samples. This sampling method is as well time consuming as the researcher is required to follow all due procedure such as first identifying strata and also going down to the strata to do the selection of individuals that will now participate in the process. The process result to monotony as the researcher or the surveyor will be repetitively assigning numbers in order to obtain the required information through this method; this may have further effects such as reducing the efficiency of the surveyor. 3.The chosen sampling method will have influence on the outcome data for use in the analysis. For instance, if the method that was used in sampling the banks was found to be biased, this will affect the results and the conclusions that will be drawn from this sample study (Mann, 2003). So to eradicate such short comings, the researcher is supposed to ensure that they reduce biasness as much as possible to save on the results and their dependability. This can be done through randomization. This ensures that all the possible samples are given equal chances of being selected for the sample of study. This so far is the effective technique that can be applied by the researcher in ensuring for equality of all possible samples when using simple random sampling. To reduce and improve stratified sampling technique, the groups are divided into groups referred to as strata that must be showing relationship that is meaningful in the study. In some cases, responses from the strata may be different from one another in a survey. Stratification is done in response to help in reflecting the population and ensuring that each stratum’s opinion is represented and reflected in the sample. In most of the cases, stratification is done by gender in order to take care of the divergent opinions and have all of them represented. Because each sampling method is concerned with precision in the analysis thereafter, testes methods are supposed to be conducted. This is done with the aim of ensuring that each sampling method chosen for use to satisfy research goals. The level of precision and the cost associated will be important to determine for each potential method. In this case, since standard error will be used, it will help in measuring the level of pre cision whereas the smaller the standard error, the greater the precision of our sample. 4.More often, questionnaires have been widely used in the collection of data from the respondents. In as much it has been preferred method for data collection, it is always associated with some problems (disadvantages). Dishonesty has been a big problem rocking the use of questionnaires in data collection. This arises as a result of the respondents abscond the truth from the researcher when answering the questions. In our case since the questionnaires were sent to the respondents, this may result to lack of clarity of questions for easy understanding by the respondents (Zaza et al, 2000). The matter of dishonesty may be as a result of hiding what they consider private for the fear of disclosure and desirability bias. Though this kind of problem can be dealt with by ensuring them (the respondents) about their privacy and also that their identifications will be hidden. Also, conscientiousness of the responses provided by the respondents can be missed since some of the respondents do not carefully think when responding to the questions. In some cases, they preselect the answers before they go through the whole question to know the requirement of the question. Validity of the data is affected when the respondents try to split the questions and even go further ahead to skip some of the questions thus missing out potential answers. The research institutions involved in this study can collect the most accurate data through structuring simple questions that are easy to read and understand by the respondents. If the questions are not presented to the respondent face-to-face like in this scenario, the respondents may have difficulty with understanding the questions and interpreting them since the researcher is not around to give clarity of what the questions need and offer guidelines. This will lead to a variation in interpretation of the questions thus resulting to different responses which some may not even be meaningful and related in any way with the subject of discussion (Zaza et al, 2000). Skewed results from this can be combated by well structuring the questions and making them easy to read, understand and interpret. Questionnaires should always be made accessible. The choice of which data collection tool to be used should be made by considering the respondents. For instance, people with other forms of physical disability such as visual impairment or hearing impairment, survey should not be used with them to collect data. Problems of this sort are eliminated or dealt with by making appropriate choice of which data collection tool to use. At sometimes, some respondents do have their own hidden agenda and this may lead them to provide biased information. Interest of the participants may steer them towards either the product or services. Questionnaires that only make use of open-ended questions are difficult to analyze by the respondents. Answers obtained through these types of questions are individualized opinions hence they cannot be quantified by the analysts since they vary across all the individual groups. Structuring a questionnaire with many open ended questions will result to more data to be analyzed. So it can be dealt with by reducing the number of open-ended questions and using the closed ended questions instead. Some of the questions remaining unanswered are other problems that are being encountered by the researchers when using questionnaires especially when the questions are optional. This risk can be avoided by making the questionnaires online and terming all the fields required. In the same way, the questions are supposed to be precise and easy to respond to. 5.The dataset that will be used to check for the representativeness of the sample will be obtained from the National Bank of Belgium in conjunction with Employment industry in Belgium. They will be used as the checking point for collected data for study. They will also be used to obtain data that are termed relevant from other sources like from the previous study. Additionally, secondary data provide descriptive information that is used to support the study that is currently being carried out thus helping in the development of the study with facts. Variables used in the study are in most of the cases tested if at all there is a relationship that exist between variables thus helping in building up the model. Secondary data are as well used in data mining where computer technology is used in studying the trend for the previous research by visiting large volumes of data. Among other uses of the secondary data, they are as well used in the identification of relevant sources in order to d o away with plagiarism. Moher, D., Dulberg, C. S., & Wells, G. A. (1994). Statistical power, sample size, and their reporting in randomized controlled trials.  Jama,  272(2), 122-124. Bacchetti, P. (2002). Peer review of statistics in medical research: the other problem.  British Medical Journal,  324(7348), 1271. Mann, C. J. (2003). Observational research methods. Research design II: cohort, cross   sectional, and case-control studies.  Emergency medicine journal,  20(1), 54-60. Zaza, S., Wright-De Agà ¼ero, L. K., Briss, P. A., Truman, B. I., Hopkins, D. P., Hennessy, M. H., ... & Pappaioanou, M. (2000). Data collection instrument and procedure for systematic reviews in the Guide to Community Preventive Services.  American journal of preventive   medicine,  18(1), 44-74.