Data Collection - Lecture notes 1-2 PDF

Preview

Summary

Analysing the different types of sampling techniques which will be important to remember for the upcoming exam this year. ...

Description

Data Collection: Data collection involves gathering and measuring data on a particular subject in a structured way. The data collected can be used to answer pre-specified questions, test for differences between different groups of respondents and evaluate the outcome of a change. Examples of data collection in business:    

Manufacturers might collect data to see how well the processes in their factories are running and to identify where improvements could be made. Marketing professionals might collect data to identify customer trends and allow them to target customers more effectively. Human resource professionals might collect data to help them to understand and evaluate the effect of their staff on the business. Finance professionals might collect data to understand and analyse trends in markets.

Data collection process: Looking at the process by answering the following questions:     

Why to collect data? What research questions are there to be answered and how will the data help to achieve that? Who is the data collected from? Is there a specific group of people or could any random person provide valuable data? What data is interesting? Is the data quantitative, qualitative or both? How to collect the data? Will there be questionnaires, observations, interviews or measurements? When will the data be collected? Is the timing important and repeated sets of data needed to be collected or will one set of data be enough?

Why to collect data? There are three broad categories of research study: exploratory, descriptive and explanatory. An exploratory study forms the initial research into a hypothetical or theoretical idea. The researcher may have observed something and conducts research to find out more. It is often the first step in more extensive research and can be followed by descriptive and explanatory studies. Exploratory studies. Exploratory studies frequently use secondary data sources, such as existing literature, and qualitative data obtained through informal discussions or formal interviews. A descriptive allows to identify and describe the differences or similarities in variables related to the research question. An explanatory study looks at and attempts to explain the relationships between variables related to the research question. Questionnaires can be used to collect both qualitative and quantitative data for descriptive and explanatory studies. A data requirements table can be useful when planning the data collection. Who are the data is collected from? A sample is a selection of observations taken from an entire population. For example, the people that live in a street are a sample of the entire population of a given city. If a survey about public transport was to be conducted, then it would be out of the question for the interviewers to approach every

single person living in a given city. It would be far too costly and it would take too long. Instead, the interviewers would approach a sample of the city’s population thereby ensuring quicker and cheaper results. The problem that arises in selecting a sample concerns who to select to form the sample. The sample should be so chosen such that it is a small scale replica of the entire population. It is said to represent or be representative of the population. To achieve an unbiased sample, we must select from the population in such a way that every member of the population has an equal chance of being selected. For example, if we select names from the telephone directory, then we introduce a bias. All the people that we select will have telephones. Those people who do not have a phone cannot be selected. This produces a biased sample-it is biased towards those people who possess a telephone. Consequently, any results obtained from a study of the sample will only apply to people who have telephones. The second source of error is the sampling error. It is the difference between the value of a statistic obtained from the sample and the value of the same statistic obtained from the entire population. However, measuring the entire population is not feasible so in reality it is not possible to actually measure the sampling error. In order to reduce the error, we may however increase the sample size. There are four main sampling strategies: Random Sampling: The only sample that is free of bias is the random sample. If a sample is chosen from a population under the condition that every member of the population has an equal chance of being selected for the sample, then that sample is called a random sample. Clustered sampling: This is a sampling strategy with the sampling done according to specific characteristics. For example, we may choose ten particular streets to conduct our survey. We could then choose a sample of only households with two children being the specific characteristic. Stratified sampling: In a stratified sample the population is divided into blocks where all the members of a block have at least one feature in common. For example, the population can be divided into three blocks, namely those under 21 years old, those who are aged between 21 and 40, and finally over 40 years old. In this way, no one person is in more than one block. The random samples are then chosen from each block. Systematic Sampling: Samples are selected such that, for example, names are selected from a list by choosing a name at regular intervals such as every 100th name. These samples are biased against those people who have no chance of being chosen once the first name has been selected. Sample size: The larger the sample size, the closer it is to the population and the more likely it is to reflect the population. However, larger samples cost more to collect. A useful rule of thumb used commonly in statistics is to collect a sample of at least 30 for each group that you wish to analyse. So if you are

using stratified sampling and you have three different groups that you need to represent, you will need a minimum of 30 from each group. Conclusion: Always remember to use all of these sampling techniques as this will be useful for the exam upcoming....