22 (United States Census Bureau). Finally, Best Market Research surveyed a disproportionate amount of those with incomes under $20,000 per year (United States Census Bureau), sample issues that most likely arose from the media of the survey, a survey. In addition to these severe sample discrepancies, the fact that the data was collected significantly earlier than the implementation date of the project is a serious time fallacy. Additionally, the statistical data only collected survey data from customers who logged onto the U.S.A. World Bank's web site. This limits the types of persons questioned to either customers of USA World Bank, those who are interested in USA World Bank, and those who accidentally stumbled onto the bank's web site. Thus, the population will not be random because it includes mostly those who already have a preference toward banking with USA World Bank, making the question regarding changing banks nearly meaningless. Best Market Research's Question Statement Problems

In addition to the sample data errors, another issue with the sample survey was the wording of is questions. For instance, the questions make assumptions, such as assuming that spouses would know about separate bank accounts, that those in non-marital living arrangements do not have shared accounts, and that ranking information that may be arbitrary. Also, asking demographic information such as race and financial information is likely to produce false results, as people are likely to lie about such features of their lives.

Jim's Data Collection Methodology

Although Jim's attempt to find an accurate representation of the population through focus groups is admirable, his data collection strategies fell short on a few crucial points. Like Best Market Research's survey, Jim's focus groups only dealt with those customers who were already members of USA World Bank. Once again, this makes the question about switching banks nearly irrelevant if the small businesses represented in the survey were already members of the bank in question. Furthermore, this group is obviously already biased toward the bank because they have their accounts, as well as potentially other services and products through this bank. Whether it is because they enjoy banking with the institution, or because the convenience of keeping all of their accounts in the same place is great, current customers of USA World Bank are more likely than others to say that they would appreciate the new products offered by the institution. Though the fact that the surveys were conducted both by e-mail and by postal mail, as Best Market Research's were not, they still subjected themselves to bias through certain aspects of the focus group. For instance, money was offered as an incentive for the CEOs to complete the survey. While this can be construed as simply remuneration for time and honesty, people who needed the money were probably more likely to respond, suggesting that their business' income was lower. Furthermore, the instrument for gathering Jim's data requested that the CEOs and founders make themselves available for a telephone conversation in which to answer the bank's questions. While this method is better than the Internet survey conducted by Best Market Research because it allows for the elimination of customers who are not sure about their answers, it is still subject to sample error because certain CEOs, those whose businesses are doing extraordinarily well, are usually unable to find the time to sit down for a chat with a bank for only $250.

Solution

Although Jim's focus group methodology was significantly less prone to error than Best Market Research's data collection methods, both faced errors. In order to have the best chances of determining which product should be implemented, both Mary and Jim must obtain new statistical data using better methodology. Although focus groups and surveys are two accepted methods of conducting...

This time, both Mary's and Jim's sample must be truly random. The best way to do this is to study the U.S. Census Bureau and the Internal Revenue Service to collect data about the demographics of both the United States' population and the characteristics of small businesses throughout the United States. After determining the types of demographics that should be represented, such as more females than males, less low-income earners, etc., the company must conduct a truly random sample that reflects this information. According to the United States Department of Education, sometimes the best ways of random sampling are just that, truly random -- such as names pulled out of a hat or randomly selected on a list ("What's the Best Way" 1998). Once the sample has been drawn, researchers should check that sample against the U.S. Census Bureau and Internal Revenue Service in order to determine any possible sample errors. While a sample cannot be exactly representative of the population in each and every way, knowledge of sample errors will facilitate a presentation to the board that includes possible fallacies, explanations, and why they do or do not impact the data.
Once USA World Bank executives know what type of sample they're looking for, it is time to begin polling the sample population. Although focus groups are a perfectly legitimate way to conduct research, in this case, surveys will be much more efficient. In order to eliminate the biases associated with computerized surveys, these surveys should be printed and mailed at random to both individuals and small business owners across the United States and the world. Paper surveys should come with an attached, postage-paid pre-stamped envelope. To avoid a bias toward those who are older and more likely to answer postal mail, the surveys will be supplemented with a series of telephone surveys, as well as personal surveys of those who are "tagged" while shopping in malls across the United States.

Although this way of conducting surveys suggests an improvement over both Jim and Best Market Research's biased survey information, the questions asked during the survey must be similarly clear and well worded. Along with demographic information, survey questions will address the likelihood of customers and potential customers becoming involved in the Instant Rewards program, along with questions that will ask consumers if they would switch banks in order to be able to take advantage of the program. Instead of asking customers to rank their top benefits, a question that would be useless given that it would simply give ordinal data, customers will simply be asked which of the incentives is their favorite, which they would use and which they would not, and if they would still use the program if only three of the services they would use are offered as rewards. By using this type of questioning, the researchers will be able to gain more insight into what card incentives the potential customers find rewarding.

After giving these new, more accurate survey questions to the truly random sample population, USA World Bank will have more representative data regarding the implementation of the two programs. Data should be studied carefully by more than one expert and analyzed for error. Before the implementation of either program, researchers and executives at USA World Bank must determine whether or not the probability of errors is great enough to result in the retraction of one or both of the programs. Furthermore, this new data may suggest that both Mary and Jim were right: there is a demand for both products in two very different business sectors.

Conclusion

From psychology to business, the use of statistics has greatly influenced institutions in a number of fields to make accurate or semi-accurate predictions about new cures, methods, and products. While statistics have been an invaluable contributor to many fields, the situation at USA World Bank suggests that firms must be cautious when using statistics to make decisions, as statistics can be easily manipulated to represent what certain firms want to hear. Although the CEOs at USA World Bank were confident that their methodology used when collecting the research data for statistical analysis was sound, it took only an experienced statistician to tell them otherwise. When a board meeting could have quickly flown by and a new proposal approved, both time and money had to be spent on collecting new data. Although this is a frustrating situation, Brian commented that he did not care about the money. Implementing the wrong product based on faulty statistics would be more costly in the long run. What the situation at USA World Bank can teach all of those who use statistics in their daily work is that statistics should be used to supplement other data, like the customer comments that Jim and Mary received, and that they should be used with caution, employing careful methodology that includes no biases. With the solution provided to USA World Bank, they will be able to conduct more research, obtain more data,…