Research tools

Business is often faced with the need to conduct studies towards certain issues. Research often seeks to increase production, reduce losses, determine new markets and improve old products. The methods obtained should give as clear a picture as possible to determine the most reasonable course of action. The most reliable method should always be applied to research work if possible.

Data collection methods

In statistics, data collection is a vital process. It helps analysts to analyze the data collected and determine the best possible action. Various data collection methods may be used. The method used should be based on the data accuracy of data required, the size of the population and available resources (Good, 2013).

Census

This method involves the collection of all the data available for collection. It involves the collection of data from the entire population. This method of data collection is very expensive and time consuming for large populations. It is however very accurate as it is capable of reproducing the actual field data and is unlikely to be biased.

Sample data collection method

This is the collection of data from only a part of the entire population. The data is then assumed to be representative of the population data. To improve on the level of accuracy, the data may be stratified so as to ensure that it is representative. A big sample is also more accurate since it is more inclusive. This method is cheap and less time consuming.

Administrative method

Sometimes data is obtained as part of the usual day to day activities of the business. Retail outlets may use the gender, age, marital status and shopping preferences to their advantage. They may therefore opt to collect this data during the day to day running of their bsiness for their loyal customers. Hospitals on the other hand collect these details as a part of their business.

Comparison of sampling methods

The most cost effective method of data collection is the administrative method. In this case, the institution does not have to provide for research to be conducted and the data is always available. The data collected this way is also reliable due to the trust that is built between the business and the consumer.  Sample method is more effective for large populations. It is the less time consuming and cheaper than census method. They are however both more expensive and less reliable than the administrative method.

Sampling methods

Sampling is the determination of the sample size. It involves both the selection of the sample size and the selection of the members of the samples. A good sample should be such that it is representative of the actual population. It should also provide as accurate information as possible. Different methods are used towards this selection.

Random sampling

This is the simplest method of sampling. Each member of the population stands an equal chance of being selected in the sample. This chance of being selected is usually known and is mainly dependent on the sample size and the population size. For very large populations, the possibility of identifying every member of the population becomes minimal. This results in a biased sample.

Systematic sampling

This is also referred to as the k-th name selection technique. After the sample size has been determined, the sample is selected by picking every k-th member of the population. This method is as workable as random sampling provided that the population does not have any hidden order. It is easier than random sampling to execute. This method is often used to select samples from lists or computerized data.

Stratified sampling

This method is favorable to random sampling since it reduces the sample error. It involves grouping the members of the population into groups according to their similarities. These groups are known as stratums. Random sampling is then used to select small samples within these samples. To determine the sample to be selected from each sample, fractional representation of each subset is considered.

A comparison of sampling methods

Random sampling is the most commonly used method of sampling. It is cheaper than other methods and less time consuming. On average, it is usually representative of the population. Systematic sampling is as good as random sampling. It however suffers the risk of having a hidden trend in the sample in which case it often biases the sample. Stratified sampling is he most representative sampling method. Every member of the population is adequately represented (Tuggle, 2012).

The article China: Commercial Banking Reportwas obtained by the use of administrative data collection method. Every banking company is required to give their financial records to the rightful authorities. The authorities on the other have a very easy time getting access that would provide for the composition of this article. The data that would be required in this case would be Banks’ Bond Portfolios, 2014, Loan/Deposit & Loan/Asset & Loan/GDP ratios, 2014, Total Assets & Client Loans & Client Deposits (US$bn), and US$ Per Capita Deposits, 2014. With this information, the rest of the article can be written (China: Commercial Banking Report, 2014).

The case of the Chinese banks is a unique one. Due to the ease of access to information, and the small population, it was possible to involve all the banking institutions. This way, the accuracy of the information was maximised.

The inferential statistical method is used in this case. The data obtained about specific banks is analyzed to determine the economic situation of China. The trends are also observed and used to obtain forecasts for the industry. Since the data obtained is further analysed to obtain inferences, the method can only be said to be inferential.

 The article The Toyota Group and the Aisin Fire by Toshihiro Nishiguchi and Alexandre Beaudet is another unique article. In this case, the information was obtained from different workers and witnesses of the two companies randomly. In this case, the willingness to contribute to the research and the possession of information were the only things that mattered. In this case, the timeline of the events that occurred after the plant was on fire were the most important. Other data that was collected included the dates of the events, the losses, and the contributions towards recovery. To substitute the information obtained this way, the records of the events was also obtained from other sources. Good sources in this case included other publications, bank records and the information that had been granted to the authorities (Nishiguchi,  & Beaudet, n.d.).

The sample was selected at random. In this case, all the witnesses that were willing to divert information were allowed to do so. As mentioned above, to complicate the method further, the data was compared with other sources so as to obtain the most reliable data. The statistical method used in this case is descriptive. The information obtained was only rearranged to presentable form and then granted to the readers. In this case, no inferences are made. The reader is left to his own devices to make any conclusions where necessary.

The most appropriate sampling method that may be used in the communications industry, say, to determine the consumption of airtime by mobile phone users, would be systematic sampling. All customer information is available within databases. It would be therefore very easy for a statistician to select every k-th consumer and obtain their consumption information. The information should be used to infer to determine the future of the industry or to understand consumer calling habits. Once the habits are obtained, it is possible to inference towards making use of consumer in formation in terms of encouraging calls during hours when consumers hardly call and to create new products. The statistical method that would be most appropriate would therefore be the inferential method (Beaulieu, 2012).

Conclusion

In statistics, it is the role of the statistician to present reliable data. For data to be termed as reliable, it has to be representative and accurate. The statistician then selects the methods that are most reliable for his case. He is also challenged to keep the costs and time used as low as possible since they are often the reason why research is often carried in the first place. A census is conducted whenever a small population is being analyzed. This maximizes the accuracy of the information presented. If a census is impossible, other methods may be used but a high sample size is highly encouraged.

References

Beaulieu, D. (2012). An Introduction to Sampling in Statistics. New Delhi: World Technologies.

China: Commercial Banking Report. (2014). China Commercial Banking Report, (2), 1-63.

Good, P. I. (2013). Introduction to Statistics Through Resampling Methods and R. Hoboken, New Jersey: John Wiley & Sons, Inc.

Nishiguchi, T., & Beaudet, A. (n.d.). The Toyota Group and the Aisin Fire | MIT Sloan Management Review. Retrieved July 19, 2014, from http://sloanreview.mit.edu/article/the-toyota-group-and-the-aisin-fire/

Tuggle, L. (2012). An Introduction to Statistics (concepts and Applications). Delhi: Library Press.