Data Collection and Sampling in Qualitative Research †Free Samples

Question: Discuss about the Data Collection and Sampling in Qualitative Research. Answer: Introduction: Interviewing or collecting data from the entire population would result to accurate information since each element will have given their opinion unlike when they would have been represented by others. The spread, expenses, time and all difficulties involved in surveying the entire population makes the researchers to prefer a sample. The quantity representing the proportion of the population is referred to as sample size. Coming up with suitable and correct size of the sample is vital in the collection of accurate information. For the collected information to be reliable, the sample size has to be correct and accurate. The sample size is calculated using the formula as follows; The percentage guess can be a pick such as 40%, 50%, 60% etc. now, depending on which percentage pick has been made the size of the sample will vary. In most of the cases, 50% is normally preferred due to its bisection of the population and conservativism when determining the sample size. Considering that we had a population of 69,000 bank workers, taking the percentage pick at 50%, margin error of 0.05 and the confidence level at 95%, the accurate and correct number of bank workers that were supposed to be surveyed is 384. Choosing to survey 15,000 bank workers lead to working with large sample size since the number was far much above the recommended sample size of 384. Working with large sample sizes has advantages and disadvantages as they will be discussed. Advantages and disadvantages of large sample size. It is advantageous working with large sample sizes since they help in minimizing marginal error. As a result, the outcomes accuracy of the ample are therefore improved. The statistic and point estimator confidence interval will tend to be in such a way that the population parameter is covered, this is according to (Clearly et al. 2014). The two research institutions that were incorporated to conduct the research being that they worked with the sample size of 15,000 bankers, it is therefore in our speculation that they must have obtained more accurate information concerning the subject in question as compared to when they would have worked with smaller sample size. Large sample size is also preferred because of their representativeness since most of the characteristics or elements in the population are covered including the outliers unlike small sample sizes (Belli et al. 2014). One of the disadvantages associated with large sample size is that it is more expensive. The expenses involved in collecting data from large sample size might involve covering a wider geographical area which will involve more cost unlike small sample size (Goodman et al. 2013). For instance, Union of Belgian Banks will incur much cost through the research institution in surveying 15,000 banker that were spread around the country. Additionally, working with large sample sizes is time consuming since a lot of time is involved to reach the individuals from various banks in various part of the country. The prior information that is known by the researchers about the topic that is under study is one of the factors that should be considered when choosing sample size. This prior information might help to make a decision on whether to increase or reduce the sample size since the estimators such as mean and variance can be used to carb the variation in the sample (Button et al. 2013). Another factor is the risk of values involved, that is, if the risk involved is to be high then small sample can be used but when the risk involved is to be low, then the sample size is to be made large to reduce the marginal error. Out of the population of 69,000 bank workers, only 15,000 bankers were ought to be surveyed by the research institutions. The process of selecting members that will represent the groups from the population under study is referred to as sampling method. In this case therefore, the research institutions used stratified sampling method. This sampling method was preferred for used due to some of the advantages it offers. One such advantages is that stratified sampling method reduces the sampling errors, this is according to (Ye et al. 2013). The population is divided into subgroups called strata where they are spread to ensure for representativeness of the population. Characteristics in the strata are selected by simple random sampling method in order to reduce or eliminate selection bias. The spread of strata and wide coverage by stratified sampling method ensures that the population of interest is highly and well represented in the selected sample. One of the disadvantages of stratified sampling method is that it takes a lot of time to identify and select the sample from the strata through simple random sampling method, this is according to (Acharya et al. 2013). Devising what to base on in categorizing the population into strata tend to be difficult and as a result researchers do tend to shun this method hence making it not widely used. Research institutions first selected the bank institutions in Belgium then further select bankers from their various working banks which in this case acted as strata where they were to be picked to form a sample by simple random sampling method. Effectiveness of the sampling methods is what drive the researchers to choose them for use in collecting certain types of data. I hereby recommend that the number of strata should be increased to ensure for high effectiveness of stratified sampling method. By so doing, representativeness of the population will increase in the same manner. A tool used by the researchers to collect a particular point time information from the already collected data is referred to as cross-sectional design. Use of this tool is always associated with some advantages and disadvantages. Worthiness of assumptions can be established in the study through cross-sectional study by cross-sectional research design, this is according to (Shen and Bjrk, 2015). Little time is spent when cross-sectional research design is used when it is compared to other research designs. It is associated with taking less time since it is concerned with extracting information from already information that had been collected and only taking specific point time information. Additionally, cross-sections research design incurs less cost as compared to other research designs such as longitudinal research design. On the other hand, longitudinal research design is seen to bear the potential of showing the design of variables or variables for a certain time coverage as one of its advantages. Some of the disadvantages that are encountered when using cross-sectional design is its lack of reliability to give the prediction of the existence of a relationship between variables and results due to unavailability of time element being that only point time information is measured. Cases for events that last for relatively longer time, cross-sectional research design tend to show the prevalent of the results from such events even if they could be of less importance. Due to the time factor covered by longitudinal research design, being that the time involved is long, the research design is termed more expensive and also time consuming. Time consumption of longitudinal research design tend to be higher than that of cross-sectional research design due to its ability to forecast the pattern over a period of time. (Shen and Bjrk, 2015) further stated that longitudinal research design becomes less efficient wen the results that are expected are less. Data from the respondents can be collected using different data collection methods such as using interviews or questionnaires. The research institutions used questionnaires to collect data from the banker in various banks in Belgium. A questionnaire is a set of questions structured in accordance to the subject under study (stress in this case) with the aim of collecting responses from the respondents. The questions in the questionnaire can either be closed ended or open ended or the scaled questions like those provided in the Likert scale. On the spaces provided, the participants are to provide their responses. This method of data collection is in most cases faced with some of the problems as discussed below; Dishonesty by the respondents is one major problem faced when the questionnaires are used to collect data from the participants (Chernik et al. 2011). The respondents can willfully or intentionally be untruthful in the answers they proved to the presented questions in the questionnaires. This can be experienced when the participants feel that their identities will not be kept private. When this is let to happen and continue, the questionnaire will risk collecting deceitful information that will later affect the results of the study in the results and discussion part of the report. This problem can be combated by assuring the respondents that their privacy will be held and highly valued, ensuring that unauthorized persons are not given access to the data and also assuring the participants that confidential information will be maintained confidential. This will boost the confidence of the respondents and the chances of the problem reoccurring will be reduced. Lack of common understanding of the questions as provided in the questionnaire is another problem. This problem mostly occur where the questionnaires are sent to the respondents without any physical contact between the researcher and the respondents hence no clarity of the questions. Varied understanding that people have will lead to varied responses to the same questions as indicated in the questionnaires. Complicated questions can also lead to such problems due to its complexity. Dealing with this problem require the researcher to compose and create the questions whose structure are simple and are easy to understand and answer. Problem with analyzing responses provided for questions in the questionnaires. Construction of the questions in the questionnaires are supposed to be well thought of since so many open ended questions will always result to respondents opinions that are often varying from one individual to another. Coding and analysis of such data becomes too difficult as the data also becomes too much than can be handled. This problem can be eradicated by making the correct choice of question types i.e. using the closed ended question or Likert scale questions other than the open ended questions (Chernik et al. 2011). The closed ended and Likert scale questions are easier to code and therefore as well simpler to analyze. Skipping the questions and leaving them unanswered is another crisis with questionnaires. In some of the cases, the respondents can decide to leave some questions with the idea that they will answer them later only to end up collecting the without answering them. Failure to answer the questions can be as a result that either the questions were complicated and not well understood by the respondents or the question required the information they do not have knowhow about. This problem can be dealt with by ensuring that the constructed questions on the questionnaires are uncomplicated, simple to understand and making the survey as short as possible to help in raising the completion rate of the questions. For the online surveys, they normally tend to make all the fields required property such they all have to be filled before proceeding to the next step. Second hand information that are collected from archives or databases are referred to as secondary information. Secondary data to be used must first be checked and ensured that they are relevant to the subject being studied. Much such as accuracy and competency of data in regards to the subject of study is to be confirmed before data is used to check for the representativeness of the sample (Piwowar and Vision, 2013). Secondary data provides the researcher with clear picture of what is expected from the sample, as a result, this saves time. Secondary data is cheaper and easier to retrieve as compared to primary data. Secondary data that were collected from large sample do have high statistical precision since most of the population elements are represented. References Acharya, A.S., Prakash, A., Saxena, P. and Nigam, A., 2013. Sampling: Why and how of it.Indian Journal of Medical Specialties,4(2), pp.330-333. Belli, S., Newman, A.B. and Ellis, R.S., 2014. Velocity dispersions and dynamical masses for a large sample of quiescent galaxies at z 1: Improved measures of the growth in mass and size.The Astrophysical Journal,783(2), p.117. Button, K.S., Ioannidis, J.P., Mokrysz, C., Nosek, B.A., Flint, J., Robinson, E.S. and Munaf, M.R., 2013. Power failure: why small sample size undermines the reliability of neuroscience.Nature Reviews Neuroscience,14(5), pp.365-376. Chernick, M.R., Gonzlez-Manteiga, W., Crujeiras, R.M. and Barrios, E.B., 2011. Bootstrap methods. InInternational Encyclopedia of Statistical Science(pp. 169-174). Springer Berlin Heidelberg. Cleary, M., Horsfall, J. and Hayter, M., 2014. Data collection and sampling in qualitative research: does size matter?.Journal of advanced nursing,70(3), pp.473-475. Goodman, J.K., Cryder, C.E. and Cheema, A., 2013. Data collection in a flat world: The strengths and weaknesses of Mechanical Turk samples.Journal of Behavioral Decision Making,26(3), pp.213-224. Khberger, A., Fritz, A. and Scherndl, T., 2014. Publication bias in psychology: a diagnosis based on the correlation between effect size and sample size.PloS one,9(9), p.e105825. Piwowar, H.A. and Vision, T.J., 2013. Data reuse and the open data citation advantage.PeerJ,1, p.e175. Shen, C. and Bjrk, B.C., 2015. Predatoryopen access: a longitudinal study of article volumes and market characteristics.BMC medicine,13(1), p.230. Ye, Y., Wu, Q., Huang, J.Z., Ng, M.K. and Li, X., 2013. Stratified sampling for feature subspace selection in random forests for high dimensional data. Pattern Recognition,46(3), pp.769-787.