Mff715 w1 1_introto_research_fall11
Download as PPTX, PDF1 like355 views
1. The document discusses the key concepts of scientific research, including deductive and inductive reasoning. It provides examples to illustrate the differences between the two types of reasoning. 2. Scientific research is defined as a systematic process of determining, acquiring, analyzing, and disseminating empirical data and insights to help decision makers maximize business performance. 3. While deductive reasoning allows determining if conclusions follow logically from premises, inductive reasoning involves using specific observations to derive general conclusions, though these cannot be proven with absolute certainty. Popper argued theories should be falsifiable and have predictive power.
Mff715 w1 1_introto_research_fall11
- 1. Week 1: What is Research? MFF 715: Forensic Research & Analysis1
- 2. What’s due on 11/20/2011 at 11:55pmPractice 1 (individual – PASS/FAIL)Quiz 1 (individual)Graded Discussion 1: Initial postings are due by the end of 11/17/2011. The individual replies are due by the end of 11/20/2011.Assignment 1 (group)2
- 3. Learning ObjectivesUnderstand what scientific research is and is notAppreciate the need for sound reasoning to interpret research results3
- 4. Why should I care about research?Research makes us better decision makersExample: How should we allocate resources for fraud detection?Example: Why are accounting fraud problems more prevalent in some divisions than others?Research improves practiceExample: What are the best ways to perform police lineups?Example: Are certain types of customers more likely to commit fraud than others?4
- 5. Truth or Myth?Eyewitnesses are better at identifying criminals when all suspects are presented together in a lineup as opposed to one at a timePeople are more likely to cheat alone when they can pocket all the profits, than when they have to cheat with colleagues and split the profitsPeople cheat more as the payoff gets biggerPeople cheat more as the risk of getting caught decreases5
- 6. “We have many intuitions in our lives. The point is, many of those intuitions are wrong. The question is, are we going to test those intuitions?” – Dan Ariely6
- 7. What is Research?From the following examples, identify those cases that you would consider scientific research:An investigator gathering information about several potential cases of financial fraudAn investigator asking a customer about her identity fraud experienceAnswer: Both or none of them could be research !The key lies on the methodology used, it should be scientific reasoning ! Scientific methodology is a systematic approach toward understanding of the world 7
- 8. What’s Scientific ResearchA process of determining, acquiring, analyzing, synthesizing, and disseminating relevant empirical data, information, and insights to decision makers in ways that mobilize the organization to take appropriate business actions that, in turn, maximize business performance8
- 9. 9
- 10. Scientific Reasoning: DEDUCTIVEReasoning (or inference) in which there is a relation between the premises and the conclusion so that the following property exists: if the premises are true, the conclusion must also be true.Example 1:All people born in the United States are U.S. citizensJohn was born in the U.S.Therefore, John is a U.S. citizenCan you provide your own examples of deductive reasoning?10
- 11. NON-DEDUCTIVE INFERENCESDeductive inference is fundamental toward the development of generalizations (theories) about the world.However, it is constantly misused (e.g. in the media).Example:The first five eggs in the box are rottenAll eggs have the same expiration dateTherefore, all eggs are rotten11
- 12. Scientific Reasoning: INDUCTIVEWe move from premises about objects we have examined to conclusions (or generalizations) about objects we have not examinedExample: Everyday the sun rises; therefore, the sun will also rise tomorrow.We use this reasoning everydayCan you provide your own inductive example?12
- 13. Inductive Inference in ScienceScience also uses inductive reasoning (e.g. how do we know that copper conducts electricity?)Still, inductive reasoning is a risky business (e.g. all swans are white)David Hume argued that inductive reasoning can not be rationally justified ……but that we should use a practical skepticism based on common sense 13
- 14. Problem with Inductive Inference: Popper’s ApproachTo solve the problem with inductive reasoning, Karl Popper argued that we can never know if a theory is true but only if it is false.Karl Popper (1934) proposed that all theories should be falsifiable (that is, they could be proved wrong ).In general, it is accepted that a theory should be falsifiable, explanatory, and have predictive power!14
- 15. Research Approach: DeductiveDeductive ApproachExampleDeduce a hypothesis from existing theoryExpress the hypothesis in operational termsTest the hypothesisExamine the outcome (does it prove the theory false?)If necessary, modify the theory in lieu of the resultsBase Theory: Cognitive Dissonance (Festinger 1957)Deduced Hypothesis: Consumers of snack foods experience cognitive dissonance due to adverse health-related publicitySurvey snack food consumers15
- 16. Research Approach: InductiveInductive ApproachExampleBerg et al (2003) studied use of cell phones by teensResearch question: How do teenagers use their cell phones?Exploratory study in a small college using ethnographic methodologyThey found that phones are used to establish and maintain the status of social networks via text exchangesResults were used for a new design for 3G cell phonesIn this approach theory follows dataDefine the context of your study (e.g. a small group of students) and your data collection and analysis method (e.g. qualitative analysis)Examine the data and develop your theory and hypotheses16
Editor's Notes
- #2: Welcome to class!
- #6: All are myths – Check out these research findingsPolice lineuphttp://www.npr.org/2011/07/06/137652142/to-prevent-false-ids-police-lineups-get-revampedCheating is more likely when benefits are sharedhttp://www.sciencedirect.com/science/article/pii/S0749597810000841Dan Ariely’s TED presentation (that you saw during the Community Experience course)http://blog.ted.com/2009/03/17/why_we_think_it/
- #7: This quote is from the TED video by Ariely that you watched as part of the Community Experience course. We think that this quote sums up very well why research is needed. Statements on the previous slide would seem reasonable, believable, or even convincing based on our intuitions or personal experiences. However research has proven all of them to be mere myths. Hopefully with your capstone work, we can help the field of the fraud and forensics to practice based on scientific evidence rather than myths or intuitions.You will watch a video by Dr. Friedrichs discussing the value of research for the field of fraud and forensics.
- #8: For cases 1 and 2 to constitute research, it is necessary to have both a research question (RQ) and a scientific methodology toward approaching the question. For example, in case 1, the RQ could be “What is the demographic of people most likely to commit financial fraud?” or, in case 2, “What are the factors that facilitate the occurrence of identify fraud?” Based on the research domain, we tend to name different types of research such as descriptive (e.g. case 1) or predictive (e.g. case 2) and so on. Also, these 2 cases constitute examples of applied research (i.e. for practical purposes), called so, to differentiate it from basic research (for theoretical purposes).
- #9: A systematic (scientific) process to find out things (purposefully); thereby increasing knowledge in the business field.Although business research could be pursued for the sake of knowledge (basic research), it is done, quite often, for practical purposes (applied research).Although modern business research tends to use both quantitative and qualitative analysis, the quantitative approach (rooted in a positivist tradition) is still the dominant approach; in particular, in applied research.Data collection and analysis are key to the scientific approach to business researchAnother key is the research question is about business/organizational issues. For example, Amy said that she’s interested in the effectiveness of weight-loss programs.She can ask several different research questions:What is the nutritional value of Diet A vs. Diet B?Does Diet A or Diet B produce higher customer satisfaction?Does Diet A or Diet B lead to higher profit margin?While 1 is a valid question, it really is more a hard science question (which is important for people like nutrition scientists) as opposed to a business research question. 2 and 3, on the other hand, are more relevant for business managers.Can you think of potential fraud and forensics research questions that may be of interest to you?
- #10: This diagram illustrates the process of doing research. This is probably the most important diagram in the course.Over the three research course, you will learn how to complete the entire research process by proposing and conducting your own research project.[More discussion about this process is available in your Ch. 4 Cooper textbook ]
- #11: The foundation of scientific research is scientific reasoning. Discussions of the two primary approaches: deductive and inductive – are not just a thought exercise or a philosophical debate. They are very practical because many myths, misconceptions, and poor decisions can be avoided with more scientific reasoning!Some examples of deductive reasoning are:General Tso’s chicken is sweet. All sweet food has sugar in it. Therefore, General Tso’s chicken must be made with sugar.All cookies are baked with an oven. Fortune cookies are a kind of cookies. Therefore, to make fortune cookies, we much bake with an oven.In the Fraud and Forensic field, we have the following deductive reasoning:3. All large corporations are likely to experience fraud. Worldcom is a large corporation. Therefore, Worldcom is likely to experience fraud4. All humans are prone to make biased decisions. Accountants are humans. Therefore, accountants are likely to make biased decisions When deductive reasoning is employed, it may be possible to develop theories that can be tested empirically (e.g. observing a large sample of accountants to confirm that they make biased decisions)
- #12: Many newspapers and politicians fall into this trap of providing true premises but that they are not related to the conclusion in a deductive way! See if you can come up with some examples!Deductive reasoning is also called deductive logic. Notice that the key condition for a valid inference is that the premises are true. However, in the following inference:All students work hard on the courseWorking hard on the course is enough to pass the courseTherefore, all students will pass the courseAs you can easily notice, this inference is not valid because the first premise “All students…” is not true (not all students work hard on the course). Similarly, the second premise is also false (working hard is not enough to pass the course, you need to master the key concepts). Therefore, the conclusion cannot be held true in light of the stated premises.An example of a failed inference in the field of fraud and forensics could be:All employees are thievesMy company has many employeesMy company will experience many theftsAs you can easily see here, the first premise (“All employees are thieves”) is false therefore cannot be used to argue for the validity of the conclusion (“My company will experience many thefts”)Can you provide your own example of a failed (or non-deductive) inference?
- #13: The word “objects” is used in a very general sense, it could also mean events or situations Some students may quickly realize the potential problems with this type of reasoning. Nicolas Taleb provides an excellent example in his book “The Black Swan” when he shows the inductive reasoning of a turkey the day before thanksgiving! ( “I have been lavishly fed during the past 364 days; therefore, I will be fed up lavishly tomorrow!” says the turkey). However, when you think about it, this is what we do all the time when we try to forecast business trends (e.g. using regression) !Another example would be “I have never known of any fraud in this company in my 20 years working here, therefore, I will not see any fraud anytime soon”
- #14: Whenever science performs an experiment and generalizes the result to the whole population, it is using inductive approach. For example, it has been determined (by examining ill people) that people suffering from Down syndrome have 47 chromosomes (instead of 46). To say that ALL Down syndrome patients have 47 chromosomes is inductive reasoning because not ALL Down syndrome patients in the world have been examined.Until the eighteenth century, Europeans thought that all swans were white. Black swans were not discovered until Europeans settled in Australia and New Zealand. This is the opening line of Nicolas Taleb’s popular book “Black Swan” that argues about the limitations of our predictive methods (based on induction reasoning). On the other hand, we could not live without inductive reasoning. For example, when I turn the steering wheel to the right, my car turns to the right; I am going to turn the steering wheel to the right now; therefore, my car will turn in to the right. My reasoning is based on my observation of the car doing so countless times in the past. As can be seen, it would not be possible to drive if we were questioning what would happen every time we turn the steering wheel! [The point of view that inductive logic was necessary and perhaps more natural for people was made by David Hume who argued for a practical skepticism; that is, rather than stating that inductive reasoning cannot provide true conclusions, we should use common sense (like in the case of arguing if the sun will rise tomorrow) in its use. However, others such as Karl Popper have denied even the possibility of such as thing as inductive reasoning. Now you may start to suspect why I said earlier that we can never prove a hypothesis to be true, only that it is false. The answer is given by the problem of inductive reasoning as illustrated by the black swan example. If this is not clear yet, do not despair, we will get back to this problem in the next slide ]Most applied research uses inductive inference. For example, you study a company or a group of companies and from this you generalize to the rest of the companies. Similarly, you can interview an accountant, or a group of accountants about their opinion on the likelihood of fraud in their companies and then generalize it as the opinion of most accountants.
- #15: Popper argued that you can never prove anything to be right (even if you show that an apple falls to the ground each time you try the experiment, you can never be certain that this will always be true. Certainly, the apple will not fall if you are in the Space). However, Popper said that you can prove if something is false (if the apple doesn’t fall to the floor just once it means that the statement “apples left to themselves fall to the ground”). This approach was proposed by Popper in The Logic of Scientific Discovery in 1934 and is the one currently accepted. Notice that when we say that a theory is falsifiable means that we can prove it to be wrong. For example; in physics, the 17th century “phlogiston theory” that posited the existence of a fire-like substance, phlogiston, present in combustible material has been proved false through multiple experiments. Why we cannot prove a theory to be true? Because, if the result of an experiment shows that conclusion A is “true,” we could replicate the experiment to confirm it but the question is how many experiments we would need to perform to make sure our conclusion is true. The answer would be the whole universe of possibilities because we need only one situation to prove the whole theory wrong. If this doesn’t seem to make sense, think about an experiment, raising a swan to see what color it turns out to be. You could have thousands of experiments that would turn out white swans; however, if you move to New Zealand, the swans there could also be black1
- #16: The previous philosophical discussion (yeah! We were doing philosophy of science) is important for practical purposes because it indicates there are two approaches to fraud and forensics research: Deductive (deducing hypotheses from existing theory) o Inductive (deducing hypotheses from existing data or experimentation). Both are equally useful in fraud and forensics research and can be used separately or even combined.
- #17: One interesting observation is that manufacturers were not particularly interested in Berg et al (2003)’s findings but the operators were. Any idea why?