What is required for a final statistics project?

1) Describe how you obtained the data. Is the data obtained from a reliable source? Did you collect the data yourself ?

2) State the statistical software used for the project. (For example, Microsoft Excel.)

3) Present the reader of your paper with summary statistics including histograms and other charts as deemed appropriate.

4) Use statistical methodology such as confidence intervals or hypothesis testing.
5) Relate how your knowledge of statistics allowed you to find one or more outcomes that may not have been anticipated at first.

6) Give conclusions and recommendations for further study.

This is a chance to look over the context and relevance of the field of statistics. Since I call it "common sense" mathematics one should use only those statistical methods that are suitable to the topic or to the data collection.

For a final project students are typically asked to work on an in-house collectionor search the internet for an applied data set. If all fails, the best place to go to is a Web-site called StatLib. This site is located at several addresses including: lib.stat.cmu.edu The StatLib Web-site was created many years ago at Carneige Melon University and includes a Data Story Library (DASL) which has data collections related to many topics.

Any survey or poll is frequently not the job of one person. The team of survey specialists should not only include people working in the field to be investigated but at least one of the team is to be responsible for statistical accuracy. In this manner, we can consider breaking down a Survey(Poll) Project as follows:

Stage One: Planning and defining of duties and responsibilities

A project coordinator is necessary. Other project team members will volunteer or be assigned duties and responsibilities appropriate to their areas of expertise. Since time is the most precious commodity it will be clear to everyone how much time they have to invest in survey project. Several meetings must be held. The overall foundations and timelines are drawn up. If extra statistical or sampling expertise is required it is important to obtain this help at an early stage.

Stage Two: Writing and Editing the Survey Questions

All team members are typically contributing to the content of the Survey Questionnaire.
The language and phrasing of questions are critical. However, deadlines must be met. If there have been successful company surveys in the past then it is useful to use these as guides for the present survey.

Stage Three: Transmission of Survey

In Stage One planning the team has to agree on the mode of conducting the survey.
The survey can be driven by way of in-person interviews, the mail, company e-mail, company voice mail or combinations of these. In a company survey a hard copy poll is usually undertaken. Then to complete the task, employees or managers who have not handed back their questionnaires can be phoned to complete their surveys.

Stage Four: Analysis of Data

This last stage can be easiest and most straight-forward. Electronic data entry will allow statistical software to display the survey results. The sooner the timely questionnaire summaries get out to people who can use them the better.

Stage Five: Conclusions and Summary Paper

The investigators have had this stage in mind all along. This is the deliverable. On the date promised the survey documents and conclusions are presented. Recommendations are made if there is survey analysis support.

Simple Random Sample(SRS) One of the most common methods is that of the simple random sample. It allows the analyst to collect a subset of a population in a very approved manner. The SRS should be a collection of n data elements where and subset of size n was possibly chosen from the population. Investigators often use random number generators to help choose the sample.

Stratified Sampling This technique of sampling can be used on its own or connected to other sampling methodology. A population is broken down into categories and then sufficiently large sub-samples are collected from each category of the population.

Systematic Sampling In industrial workplaces one of the simplest ways of sampling is to check every nth product. Or to check product samples every n minutes. In either case, the sampling is can be a quality assurance feature that is a vital part of manufacturing.

Convenience Sampling When a large truckload of potatoes are delivered to a potato chips manufacturer a conveniece sample is necessary. One person from the manufacturer takes as best a sample as he can find and sends these sample potatoes to the company laboratory for testing. If the tests indicate approval then the whole truckload is delivered. Otherwise the delivery is sent back.

Matrix Sampling A population collection is sometimes arranged as a matrix or an array. The sampling can be accomplished in a mathematical fashion so that the investigator is retrieving a subset of the population in a very structured way. The population of measurements determines whether this technique can be used.

Cluster Sampling Knowledge and familiarity of the population measurements is essential for many sampling techniques. Clusters of data points may be retrieved from the population. Analysts may look at some individual clusters as well as the overall total sample.

Statistics is still a developing science and not as rigid as Calculus and Algebra. To use statistics and its methodology intelligently and correctly you have to see the "big picture". In starting a project based on statistical techniques get all the help you need and look at successful projects that were accomplished that are similar to yours. All statistical models are based on an array of assumptions. Should any one assumption fail then the outcome of the project may be in jeopardy. The famous American statistician George Box has been quoted as saying: "All statistical models are wrong". Don't be discouraged. There is a great validity in statistics and modern agriculture and manufacturing could not be where it is today if not for good statistical reasoning.