Could a stratified sample be used? If so, how?

Yes, a stratified random sampling technique could also be used as the specific companies that comprise the population could be stratified or layered by their specific criteria. The most common one used is revenues, yet spending on capital equipment, or the number of production plants could also be used to stratify the sample. Of these, financial measures are the most basis of stratification in research populations.

A d) Could a cluster sample be used? If so, how?

Yes, clustering based on the specific types of production processes shared across each manufacturing segment could be used, as could the use of specific types of steel. This second attribute specifically could be used for defining the various grades of the new power press being manufactured. Clustering could also be done by the percentage of customers each company has in specific sub-segments of the markets they serve, or by the use of distribution channels. Clustering as a sampling technique is often used when a population has a unique set of attributes that need to be captured as part of the research methodology.

A e) Which sampling technique would you recommend? Why?

Of all techniques, a stratified random sample based on revenues is the best approach as there will be major differences in the approaches these companies take in funding a new power press. There will be differences in the approaches each of these stratified samples have in terms of how they also track and report Return on Invested Capital (ROIC), and their approaches to managing the financial evaluation process for new equipment.

A f) Identify two possible nonsampling research errors that might occur in the...

Explain how you would deal with these nonsampling errors.
Using a stratified random sampling technique based on revenue could provide two nonsampling research errors. The first would be that the assumption of purchasing processes varying significantly by size of company based on revenue could prove to be erroneous, yet sample means testing would bring this out quickly during a preliminary statistical analysis of the results. A second error could potentially be the lack of ROI activity across both groups, a fact that would not emerge until after the analysis was completed. The first problem could be alleviated through the use of a set of screening criteria applied to the entire research population prior to the survey being completed to ensure the hypothesis of processes being different actually exist. The second error could be alleviated through speaking with industry research firms who possibly have secondary data in the form of ROI analyses in the industries of interest. Specifically, industry associations for automobile, construction equipment and major appliance or durable goods industries could be contacted to see if they have any visibility into the ROI processes in place by size of company.

Question Three

Refer to the Pella Satisfaction Questionnaire. Identify what meaningful and useful analyses you would like to conduct and why: What cross-tabulations for which variables?

Cross-tabs need to include Question 1 on customer satisfaction by questions 3, 4, 5, 6, and both 7a and 7b. These cross-tabs will explain the specific attributes that lead to greater customer satisfaction from a sales cycle execution standpoint, product features, performance and perception of them by owners as