Reference no: EM13301069

Week 05 Discussion - The Ethics of Classification: "Student Posts"

Provide me a half page response to each discussion post in detail.  Make sure that you are responding to the post describing what you agree with and disagree with if anything.  Also, provide any alternatives if any.  Tell me in detail what you like and or dislike and explain your responses to each discussion.

Discussion #1:

"The Good"

The Data mining tool regarding classifications is essential for showing students under 23 years of age who have the ability to fund their educational pursuits themselves.  Through this determination, admission committee members see a group of candidates who are serious enough about learning that they are willing to obtain their own loans or work to pay for tuition and books.  Typically, these students are classified accurately as most people not serious about their educational endeavors, wouldn't risk their credit being lowered or dispose of their hard earned money for no reason.

Another example of positive data mining is the determination that students over the age of 23 with over 5 years of work experience should be consider low risk and accepted into a program.  This is ideal thinking as applicants within this category have shown discipline within their respective industries.  The dedication they show is typically displayed on their resumes with measurable achievements at their prior and current organizations.

"The Bad"

The Data mining tool can be a detriment when you assume that those individuals over the age of 23 with less than 5 years of work experience and no children are high risk and should be rejected.  For example, an applicant could have 4.5 years of work experience, but that work experience could have been at a high level, working with senior level executives on major initiatives that led to serious changes within an organization.  Due to the high level of dedication to career pursuits, the opportunity to have children has not presented itself.  Does this mean that the candidate should be eliminated from consideration all together?  Certainly not.

Another example of bad classification is students under 23 years of age who have their parents as a source of funding, but have less than a 3.0 GPA.  Some students aren't as dedicated to academic pursuits at a high school level.  This shouldn't disqualify them from a potentially successful future.  Sometimes, it takes the realization of a college experience for students to find themselves and mature.  Through this matriculation, the seriousness of academics tends to kick in and students who seemed to slack in one arena, thrive in another and begin to excel with new experiences.

Discussion #2:

Data mining allows large values of data to be analyzed from many different dimensions through an automated process.  As data gets increasingly large, the automation simplifies the process to find correlations and patterns within the data necessary for decision making.  Data mining can segment and group data to aid businesses in marketing campaigns, identifying high risk customers, make predictions for future customer behavior and draw conclusions about website traffic patterns which result in business decisions that can improve customer service or increase profits.

Data mining is for the most part an automated process, statistically sound process.  That being said data mining techniques should not be a threatening or harmful. However, it seems that when the results of data mining techniques identify patters across behaviors, race, gender and nationality the results can be interpreted as biased or stereotypical.  It is often because in the decision making around these classifications can become subjected to less favorable outcomes.  For instance, higher insurance rates for females.  This is also frequently mentioned in the cases of healthcare insurance and loan approval processes. 

Discussion #3:

There are definitely positive aspects to data mining. Data mining allows companies and institutions to refine information. This makes the organization more efficient, saves time, and can make them more effective.

The example provided is a good example of how data mining can save significant time and resources. College admissions departments get an enormous number of applications each year. Each application is an extensive document. If every document is reviewed in full, a lengthy complicated process would in sue. Refining this process down to a meaningful selection of applicants is only logical.

This example also demonstrates a very bad aspect of data mining. The criteria used needs to be valid, unbiased and appropriate. The software should simply eliminate candidates or refine candidates, but not exclude anyone based on unfair criteria. For example, according to the chart provided, I should have been automatically rejected from grad school.  When I applied, I had less than 5 years experience in the workforce and no children. Yet, I graduated my undergrad with Honor's and have performed well at Scranton. The data mining should be limited to actually criteria that is used to evaluate a student. For example, if a school has a policy that they do not admit anyone with less than a 3.3 in undergrad, than that would be criteria. 

On the other hand, data mining can also be a great way to put some candidates into a pile to examine closer. For example, anyone with less than 5 years experience and no children may go into a pile to be looked at closer, rather than rejected. This allows data mining to still serve a person without being bias against any groups. 

Discussion #4:

Data mining software provides statistical random results of information, so a user can utilize this information for decision making. The good of classification with using data mining is that it provides numerous results or data that can be drill down to multiple common results. Another good is the results are obtained by using statistically techniques or tools, and acquiring information from statistical tools can eliminate some information from being bias.

The bad of classification using software like data mining is having too much information, especially if the information does not help in the decision making process. Having too much information can cause the classified information to be irrelevant.  For example, in our textbook the case of classifying applicants for college; the data obtained in order to create a decision tree for how to select applicants; it included information about an applicant having children. The fact an applicant has children is irrelevant to the decision of selecting and admitting students into the university. The information provided in these data mining software can narrow down information to be irrelevant that can be stereotyping or unethical to the population.

Discussion #5:

Data mining can provide useful and helpful information to a number of industries. For example, data mining in marketing can help predict who will respond to new marketing campaign such as direct mail or online marketing campaigns. Retail stores can use market basket analysis to appropriately arrange products in a way that customers bundle products together, or give insight into particular products that will attract customers. Financial institutions can use data mining to obtain information about loans and credit reporting. The bank or financial institution can estimate the risk levels are involved in giving loans and repayments. However, as you can imagine this is where data mining can have certain ethical implications and disadvantages.

I would think that the two biggest ethical issues regarding data mining would be those surrounding privacy, and when the results are used in decision making processes that effect people, like in the loan example I gave above. The data mining itself does not represent an ethical issue to me; however how the data is utilized is where the ethical dilemma lies. For example, if it is determined that single females of a certain race were a risk for loan payments, and were denied a loan as a result of this data, I think it would be an ethics issue. A decision making process is effecting a person based on results that are subject to error due to issues such as dirty data, missing values, or inconsistent data. Stereotyping based on inaccurate information could certain lead to legal issues as well. 

Regarding privacy, a simple online search returns any number of lawsuits that have been brought against companies for data mining and invasion of privacy. One lawsuit filed against Microsoft, McDonald's, Mazda, and CBS claims data-mining was used "to identify the websites people visit, invading people's privacy, misappropriating their personal information and interfering with the operations of their computers." (Smith, 2010) Of course I now know that in a data warehouse somewhere is information that I visited this article online...

https://www.networkworld.com/community/blog/lawsuit-claims-microsoft-mcdonald%E2%80%99s-mazda-cbs