When collecting data for business research, there are three major factors to be considered. These include, how much data is to be collected, how it is collected and how to handle it after collection. The sampling method used in business research activities is determined by the first two factors. The data collection method and the amount of data required influence the choice of sampling procedure.
The effectiveness of a sampling procedure is determined by how well sampling biases are avoided. These biases are mostly manifested through underrepresentation (Denzin & Lincoln, 1994).
Even sampling produces better results when conducting business research. In this case, two types of sampling methods will be explored. These are stratified sampling and cluster sampling. Management examples where these methods can be employed and their accompanying characteristics will be explored.
Stratified sampling involves picking random samples from each strata of the population. Members of a population are categorized into mutually exclusive groups. In addition, these groups have to be collectively exhaustive.
Cluster sampling also known as block sampling involves dividing the population to be sampled in clusters. These groups unlike in stratified sampling are heterogeneous as opposed to the homogeneous groups used in stratified sampling.
Stratified sampling is best employed in business research situations where the population naturally divides itself into classifiable strata. These homogeneous groups of the population have to show substantial interstratum disparities. For example, when conducting a customer satisfaction survey, the customers are randomly selected then subdivided.
Out of all these customers, ten percent could be managers, seventy percent could be end users, and the other twenty percent could be operators. This represents a homogeneous population with disproportionate strata. The sampling method has to take this into account to avoid overrepresentation or underrepresentation of a particular stratum.
Cluster sampling is applied when conducting business research on a population that divides itself into geographical clusters. Each cluster has to be as heterogeneous as the rest of the population. For example, a company may wish to research on twenty similar projects. The company may opt to pick five of these projects to represent the whole group. These are the clusters used in this sampling method.
The data collected on those five projects, is representative of the data expected from the rest of the projects. There is need to pick these clusters randomly to avoid generalization. In addition, picking very few clusters may result in the same problem.
Shifting Oil Demand: A Content Analysis of Changing Oil Demands in Asia
Europe and North America have always served as the indicators of world’s oil demand. In this article authored by Nelder a Smart Planet’s columnist, the shifting trends of world’s oil demands are addressed. America has been the leading swing consumer in oil products for close to a century (Nelder, 2012).
This means that due to America’s high volume demands, the country could have afforded to outbid any developing economy. However, the situation has changed since 2005 when instances of developing countries out-bidding America were first witnessed. This puts America in a tight spot because it no longer has the ability to influence global oil prices.
This article covers various aspects that concern oil demand. The demand being addressed in this case is the one primarily driven by fuel demand in American and Asian economies. The author of the article is mainly focussed on export driven demand. The countries that have control over oil prices are cited as either belonging to the Asian faction or the American faction.
The author hypothesises on the need to use fuel-efficient cars and light trucks in order to cut fuel consumption. According to the article, Asian markets actualize this and it has a positive impact on their economies. The article variously references OECD as a source of oil consumption statistics. The overall aim of the article is to explore what factors are responsible for shifting oil demand in the global markets.
This article’s content analysis reveals a list of frequently mentioned names. The most mentioned name is demand occurring a total of thirty three times. Oil is second with twenty-seven mentions. The word year is mentioned thirteen times and fuel is mentioned twelve times.
The author mentions organization for Economic Development and Cooperation- OECD ten times. Export and Economy are mentioned eleven and eight times respectively. The article also mentions World, Asia, and people eight times each. Price is mentioned seven times in the text.
This word count indicates that demand is the central focus of this article. The article addresses how demand for oil products is shifting from the developed countries in the West to the developing ones in the East. Soaring population is the main factor influencing oil demand.
According to the article, China will have one hundred and twenty five million more vehicles on the road within a period of five years. This equals to half the number of vehicles in the United States today.
Oil is the next most mentioned word. This is to be expected given that the demand being addressed concerns oil. Mention of oil is in tandem with the reference of the word demand. It is also used to indicate the factors surrounding the shift in demand for oil. For example, the article covers oil consumption, costs, export, and import trends.
The number of times oil is mentioned in relation to pricing is also notably high. Price itself is mentioned seven times in the article. Naturally, a rise in demand results in high prices. The addressed shift in demand is in line with fluctuating oil prices. This ties all these three factors together. The shifting oil demand and its fluctuating prices form the theme of the article.
The main theme is also supported by various other factors and references. These factors are exemplified by frequently occurring word mentions throughout the article. For instance, year is mentioned thirteen times in the story. This is because it is used to show the timeframes of the shifting demand.
Fuel is the major by product for oil. This is why it is severally mentioned because the shifting demand directly influences oil demands. Others are supporting factors of shifting demand like people, economies, and production.
Linear regression models and Statistical Testing
Linear regression refers to the measure in the ability of an independent variable to give a prediction of a dependent variable’ depending on their interrelationship. There are two main models of linear regression. These include simple or multiple linear regression and logistic regression. In simple regression, the dependent variable’s outcome is predicted by one independent variable (Fox, 1997).
Multiple regression is a model predicted by more than a single independent variable. On the other hand, logistic model is used where dummy independent variables are used. In this case, one or more of the independent variables are continuous variables.
Simple linear regression models can be effectively employed in SPSS programs. This form of linear models is very common and quite easy to employ. This is because it is possible to approximate the non-linear relationships using the straight lines (Trexler & Travis, 1993).
Using this model, there is a possibility of transforming data previously in non-linear models into linear models formats. Due to its simplicity, this model is often considered a general methodology.
An example of how to statistically test this model is detailed below. You could use this model of measure the mobility of old people with the expected score being TED. In this case, there is a response variable and an explanation variable.
Linear regression is used together with least squares model to give an estimation of the parameters involved in calculating the mobility of an old person. Each increase in TED has to be directly related to these parameters.
Logistic regression linear model is mostly used in the exploratory phase of a research. This makes it possible to eliminate the redundant variables early on in the research (Pedhazur,1982). After eliminating any variable, it is important to test the model. The analysis can be considered complete in case when there are no more variables to be excluded.
One statistical test used with this model is the likelihood-ratio test. This test makes use of the “ratio of the likelihood of a maximized value” in order to function for a full mode (L1) over the likelihood function when a simpler method is used (L0) (Trexler & Travis, 1993).
The outcome of this equation is the likelihood-ratio test determinant. The value of the outcome is known as chi-squared statistic. It is often used when a model is being built through backward step elimination.
References
Denzin, N.K. & Lincoln, Y.S 1994, Handbook of Qualitative Research, Sage Publications, Thousand Oaks, CA.
Fox, J 1997, Applied Regression Analysis, Linear Models and Related Methods, Sage, California.
Nelder, C 2012, Oil demand shift: Asia takes over, <https://www.zdnet.com/article/oil-demand-shift-asia-takes-over/>
Pedhazur, J 1982, Multiple regression in behavioral research: Explanation and prediction, Rinehart and Winston, New York.
Trexler, J.C. & Travis, J 1993, “Nontraditional Regression Analyses”, Ecology, vol 74, pp. 1629-1637.