One point I found interesting in the chapters is the use of scales to present information numerically. Scales, as illustrated in the chapter are a form or a criterion through which numerical information is presented to facilitate its interpretation either in analysis, or in the presentation of the research findings. It is important to use scales in the presentation of this data owing to a variety of reasons. The choice of scales used in the presentation of this information also differs primarily due to the amount and nature of information the researchers use. One scale may be more representative of a certain population of data in comparison to the other. The understanding, therefore, of all the scales, their strengths, weaknesses and limitations enable the researcher to present the information required in the most suitable manner possible. For instance, there are the basic types of scales, the nominal, ordinal, interval, and the ratio. The use of these scales implies that in each scale, ranking from nominal to ratio, the information presented on the variable increases with each subsequent scale.
The nominal scale, for instance, gives some basic gross information as to the variables, such as mean calculation between two mutually exclusive and collectively exhaustive categories, such as male and female variables in the study. The information that would be derived from the assignment of code numbers one and 2 is the percentage of each group of participants to the study. The ordinal scales go a step further, not only differentiating the variables, but also by rank ordering them in a meaningful way. For instance, in a job research drive, the researcher not only differentiates the percentages of men and women queried in the drive, but also their preference to the jobs offered. The interval scaling is an improvement to the ordinal scale in that it helps in showing the magnitude in the difference in ranking of the ranks. The interval scale thus allows the measurement of the distance between any points in the ranks. The final scale under this consideration is the ratio scale which allows the researchers to measure the proportion of the differences between points in a scale. It is the most powerful of the four scales discussed.
The use of different scales for different types of surveys is an important attribute of good research methods. The choice of the scale type and complexity determines the level of information, and depth of the information available to the researchers for analysis. However, the choice of scale types is in many cases left at the discretion of the research team. Choice of one research tool might hinder the representation of certain aspects of the investigation that would have been possible by the use of a different scale. How do the researchers determine the best scales to use under certain circumstances, avoiding choice of scales presenting the least effort in their use, or obscuring important aspects of the investigation deliberately?
Another interesting point learnt in the chapters is the need for precision in data analysis and the maximization of the confidence levels of the data obtained from the samples. In maximizing both, the sample size has to be increased as much as is possible. In the event that this is not possible, then, on is sacrificed in the elucidation of the other. For instance, if we wish to maximize the confidence levels in the sample size, then the level of precision is sacrificed. It is therefore a matter of paramount interest to the researchers in determining the approach to take in each case in maximizing the two values. In the ideal situation, the use of the largest population samples is recommended. If this approach proves untenable, then it is important that the determination on four factors be made in ensuring that the results presented are representative of the findings. These variables under consideration are;
1. The precision required of study, i.e. the margin of allowable error
2. Confidence needed, how much chance of error is tenable
3. To what extent variability exists in the population traits
4. The cost/benefit of increasing the sample size if possible to do so
The use of precision and confidence in results analysis is a crucial part at ensuring that the research findings indicate the most representative and useful portrayal of the research intent. Sacrifice of either leaves the research under the risk of misrepresentation and consequent misinterpretation. However, and in cases where the sample size cannot be expanded beyond the current size, how do the researchers determine that the sacrifice of one aspect, such as confidence is prudent/ desirable in comparison to precision? Both being complementary variables, how does one win over the other in representing a trustworthy result of the research findings?
Sampling is another important point learnt in these chapters. The sample size, sampling methods, and suitability of each sampling method are critical factors to consider in the choice and implementation of a sampling method to a given research situation. Sampling for qualitative analysis is an interesting proposition in particular. Fidelity to procedure and proper research methods is a key attribute in choosing qualitative sampling methods. The target population, for instance has to be clearly defined. The sampling utilized in qualitative research is known as purposeful sampling as the inference required from the population is of a non-statistical nature. The wealth in the information required should thus reflect in the population chosen. Such wealth is achieved through the application of a sampling method introduced by Glaser and Strauss known as the grounded theory. This theory holds that the theory in the data emerges through an iterative process through which repeated analysis of the data produces no new information to the on the population chosen.
The choice of sampling methods is of importance to the accuracy and usefulness of research findings. The nature of the population and the inclusivity of the required results are some of the variables that influence the choice of the sampling methods to apply. The researchers, in the instance of qualitative data that is non-represent able statistically, have to utilize a rigorous and trustworthy sampling regime to obtain reliable results. However, in the choice of the sampling methods, and in the interpretation of the research findings, the proficiency and trustworthiness of the researcher are paramount. The probability that they will choose the appropriate sampling method, and execute it to a desirable level of thoroughness depends on their commitment to doing so. The drawback in that qualitative findings are not statistically verifiable ensures that the findings presented by the researchers are taken as the actual position of the study population investigated. In the light of this drawback, how do researchers and managers ensure that there is no abuse of sampling methods, or failure to procedure adherence with the aim at obtaining biased results?
Sekaran, U., &Bougie, R. (2014). Research Methods for Business: A Skill-Building Approach, 6th edition, 9781119942252 | CourseSmart. Retrieved March 10, 2014, from http://www.coursesmart.com/9781119942252/firstsection