120-150words each QN. Total 450words maximum.
angella123
planninganddecisionmakingch12class17.pptxJudgment in Managerial Decision Making 8e Chapter 12 Improving Decision Making
Copyright 2013 John Wiley & Sons
1
Use Decision-Analysis Tools
Linear models
What are they?
A formula that weights and adds up the relevant predictor variables in order to make a quantitative prediction
An example,
A parent asked his boy’s pediatrician to predict how tall his son would grow. The pediatrician offered a simple linear model in response.
She said that a child’s adult height is best predicted with the following computation.
First, average the parents’ heights.
Second, if the child is a boy, add two inches to the parents’ average. If the child is a girl, subtract two inches from the parents’ average
One way to improve decision-making is to use decision-analysis tools.
These tools allow people to guide their decision-making process by converting qualitative preferences into quantitative data that can be converted into expected values and then compared. They typically consider the following:
Weights placed on the importance of different attributes.
The value of each attribute.
The probability of uncertain events occurring.
The objective costs associated with each option.
Linear models facilitate the process of performing a quantitative decisions analysis when there is uncertainty.
These models account for relevant variables that predict future outcomes and combine them in a linear fashion to provide estimates of future outcomes.
They can be used to predict a wide range of things such as:
A child’s adult height given the height of his or her parents and his or her gender.
A baseball player’s future performance given his past performance, age, height, and weight.
A movie studio predicting the potential revenue that will be generated by a movie.
Linear models are effective at improving decisions.
Research in a variety of domains has demonstrated that they make better predictions than “experts”.
More complex models provide limited improvement to simple linear models.
We have inconsistent preferences that are prone to bias while models are consistent and unbiased.
They also are effective at allowing people to gauge the effectiveness of various practices, which helps them learn and fine-tune their managerial strategies.
Despite the demonstrated effectiveness of linear models, people frequently object to their use.
People still believe that brief interviews are more effective at predicting performance than a long history of prior performance and aptitude tests to draw from. Neither of these factors truly give people a sense of an individual’s unique characteristics and the latter requires little effort and provides a lot of unbiased data.
People also believe that it is impossible to model the qualitative preferences and tastes of individuals, but research has demonstrated that linear models are reasonably effective at capturing such qualitative preferences.
One domain in which linear models could improve the decision-making process is admissions decisions.
In a manifestation of what is known as the correspondence bias, graduate school admissions tend to be biased in favor of individuals with high GPAs from lenient institutions because they fail to sufficiently discount the role of lenient grading.
One researcher developed a linear model to make graduate admissions decisions and compared it to the decisions of an admissions committee.
The model was capable of ruling out 55% of applicants who the committee also rejected.
The model also proved more effective at predicting the future ratings of applicants than the committee.
Overall, the effectiveness of this simple model suggests that the implementation of simple models using decision rules to make quick admissions decision can save universities hundreds of millions of dollars in time and resources saved from not having committees review applications.
Decision-analysis tools can also be of assistance in making hiring decisions.
Job interviews predict only 14% of the variability in employee performance. However, people continue to believe in them as an effective diagnostic tool.
Interviewers are biased by the availability of subjective qualities that they associate with effective performance when interviewing potential applicants.
Interviewer affect is influenced by superficial cues associated with a candidate and this influences their perceptions of the candidate.
Individuals who are extroverted, sociable, tall, attractive, and ingratiating are often considered as ones that represent the qualities of effective leaders even though they are far less predictive of performance than less observable traits such as conscientiousness and intelligence.
Managers tend to seek information that confirms the quality of their decision after hiring an applicant, which prevents them from learning. In addition, they never have a sample of performance from those who were interviewed but not hired to use as a comparison group.
Interviews may be difficult to justify based on the time that goes into them and the little predictive validity they have, but if the ratings of interviewers are combined into a linear model with measures of aptitude and past performance, hiring decisions could be improved.
2
Linear Models
Why they lead to superior decisions?
Lead to superior predictions than experts
More complex models produce only marginal improvement above a simple linear framework
Linear model are superior because:
People are much better at selecting and coding information (what variables to put) than they are at integrating the information (using data to make a prediction)
We are inconsistent. Given the same data, we will not always make the same decision
A linear model can be programmed to avoid biases that are known to impair human judgment
Allow organizations to identify the factors that are important in the decisions of its experts, making them valuable tools
One way to improve decision-making is to use decision-analysis tools.
These tools allow people to guide their decision-making process by converting qualitative preferences into quantitative data that can be converted into expected values and then compared. They typically consider the following:
Weights placed on the importance of different attributes.
The value of each attribute.
The probability of uncertain events occurring.
The objective costs associated with each option.
Linear models facilitate the process of performing a quantitative decisions analysis when there is uncertainty.
These models account for relevant variables that predict future outcomes and combine them in a linear fashion to provide estimates of future outcomes.
They can be used to predict a wide range of things such as:
A child’s adult height given the height of his or her parents and his or her gender.
A baseball player’s future performance given his past performance, age, height, and weight.
A movie studio predicting the potential revenue that will be generated by a movie.
Linear models are effective at improving decisions.
Research in a variety of domains has demonstrated that they make better predictions than “experts”.
More complex models provide limited improvement to simple linear models.
We have inconsistent preferences that are prone to bias while models are consistent and unbiased.
They also are effective at allowing people to gauge the effectiveness of various practices, which helps them learn and fine-tune their managerial strategies.
Despite the demonstrated effectiveness of linear models, people frequently object to their use.
People still believe that brief interviews are more effective at predicting performance than a long history of prior performance and aptitude tests to draw from. Neither of these factors truly give people a sense of an individual’s unique characteristics and the latter requires little effort and provides a lot of unbiased data.
People also believe that it is impossible to model the qualitative preferences and tastes of individuals, but research has demonstrated that linear models are reasonably effective at capturing such qualitative preferences.
One domain in which linear models could improve the decision-making process is admissions decisions.
In a manifestation of what is known as the correspondence bias, graduate school admissions tend to be biased in favor of individuals with high GPAs from lenient institutions because they fail to sufficiently discount the role of lenient grading.
One researcher developed a linear model to make graduate admissions decisions and compared it to the decisions of an admissions committee.
The model was capable of ruling out 55% of applicants who the committee also rejected.
The model also proved more effective at predicting the future ratings of applicants than the committee.
Overall, the effectiveness of this simple model suggests that the implementation of simple models using decision rules to make quick admissions decision can save universities hundreds of millions of dollars in time and resources saved from not having committees review applications.
Decision-analysis tools can also be of assistance in making hiring decisions.
Job interviews predict only 14% of the variability in employee performance. However, people continue to believe in them as an effective diagnostic tool.
Interviewers are biased by the availability of subjective qualities that they associate with effective performance when interviewing potential applicants.
Interviewer affect is influenced by superficial cues associated with a candidate and this influences their perceptions of the candidate.
Individuals who are extroverted, sociable, tall, attractive, and ingratiating are often considered as ones that represent the qualities of effective leaders even though they are far less predictive of performance than less observable traits such as conscientiousness and intelligence.
Managers tend to seek information that confirms the quality of their decision after hiring an applicant, which prevents them from learning. In addition, they never have a sample of performance from those who were interviewed but not hired to use as a comparison group.
Interviews may be difficult to justify based on the time that goes into them and the little predictive validity they have, but if the ratings of interviewers are combined into a linear model with measures of aptitude and past performance, hiring decisions could be improved.
3
Linear Models
Why do we resist them?
Some have raised ethical concerns
People still believe that brief interviews are more effective at predicting performance than a long history of prior performance and aptitude tests to draw from.
How could they tell what I’m like?
Neither of these factors (interviews or linear models) truly give people a sense of an individual’s unique characteristics and the latter requires little effort and provides a lot of unbiased data
Despite the demonstrated effectiveness of linear models, people frequently object to their use.
People still believe that brief interviews are more effective at predicting performance than a long history of prior performance and aptitude tests to draw from. Neither of these factors truly give people a sense of an individual’s unique characteristics and the latter requires little effort and provides a lot of unbiased data.
People also believe that it is impossible to model the qualitative preferences and tastes of individuals, but research has demonstrated that linear models are reasonably effective at capturing such qualitative preferences.
One domain in which linear models could improve the decision-making process is admissions decisions.
In a manifestation of what is known as the correspondence bias, graduate school admissions tend to be biased in favor of individuals with high GPAs from lenient institutions because they fail to sufficiently discount the role of lenient grading.
One researcher developed a linear model to make graduate admissions decisions and compared it to the decisions of an admissions committee.
The model was capable of ruling out 55% of applicants who the committee also rejected.
The model also proved more effective at predicting the future ratings of applicants than the committee.
Overall, the effectiveness of this simple model suggests that the implementation of simple models using decision rules to make quick admissions decision can save universities hundreds of millions of dollars in time and resources saved from not having committees review applications.
Decision-analysis tools can also be of assistance in making hiring decisions.
Job interviews predict only 14% of the variability in employee performance. However, people continue to believe in them as an effective diagnostic tool.
Interviewers are biased by the availability of subjective qualities that they associate with effective performance when interviewing potential applicants.
Interviewer affect is influenced by superficial cues associated with a candidate and this influences their perceptions of the candidate.
Individuals who are extroverted, sociable, tall, attractive, and ingratiating are often considered as ones that represent the qualities of effective leaders even though they are far less predictive of performance than less observable traits such as conscientiousness and intelligence.
Managers tend to seek information that confirms the quality of their decision after hiring an applicant, which prevents them from learning. In addition, they never have a sample of performance from those who were interviewed but not hired to use as a comparison group.
Interviews may be difficult to justify based on the time that goes into them and the little predictive validity they have, but if the ratings of interviewers are combined into a linear model with measures of aptitude and past performance, hiring decisions could be improved.
4
Linear Models
Why do we resist them?
Choosing mushmelons for his grandmother’s cake requires judgment, taste, experience, and a lot more that are not possible in the linear model
Research, however, has shown that they can
They rule out intuitions or gut feelings
Requires difficult changes within organizations
What will bank loan officers do when computers make the decisions?
These models still need us to specify the inputs, monitor the outcome, update variables, and so on
Despite the demonstrated effectiveness of linear models, people frequently object to their use.
People still believe that brief interviews are more effective at predicting performance than a long history of prior performance and aptitude tests to draw from. Neither of these factors truly give people a sense of an individual’s unique characteristics and the latter requires little effort and provides a lot of unbiased data.
People also believe that it is impossible to model the qualitative preferences and tastes of individuals, but research has demonstrated that linear models are reasonably effective at capturing such qualitative preferences.
One domain in which linear models could improve the decision-making process is admissions decisions.
In a manifestation of what is known as the correspondence bias, graduate school admissions tend to be biased in favor of individuals with high GPAs from lenient institutions because they fail to sufficiently discount the role of lenient grading.
One researcher developed a linear model to make graduate admissions decisions and compared it to the decisions of an admissions committee.
The model was capable of ruling out 55% of applicants who the committee also rejected.
The model also proved more effective at predicting the future ratings of applicants than the committee.
Overall, the effectiveness of this simple model suggests that the implementation of simple models using decision rules to make quick admissions decision can save universities hundreds of millions of dollars in time and resources saved from not having committees review applications.
Decision-analysis tools can also be of assistance in making hiring decisions.
Job interviews predict only 14% of the variability in employee performance. However, people continue to believe in them as an effective diagnostic tool.
Interviewers are biased by the availability of subjective qualities that they associate with effective performance when interviewing potential applicants.
Interviewer affect is influenced by superficial cues associated with a candidate and this influences their perceptions of the candidate.
Individuals who are extroverted, sociable, tall, attractive, and ingratiating are often considered as ones that represent the qualities of effective leaders even though they are far less predictive of performance than less observable traits such as conscientiousness and intelligence.
Managers tend to seek information that confirms the quality of their decision after hiring an applicant, which prevents them from learning. In addition, they never have a sample of performance from those who were interviewed but not hired to use as a comparison group.
Interviews may be difficult to justify based on the time that goes into them and the little predictive validity they have, but if the ratings of interviewers are combined into a linear model with measures of aptitude and past performance, hiring decisions could be improved.
5
Linear Models
Evidence to better outcomes:
Graduate school admission
Hiring
Graduate school admission
In a manifestation of what is known as the correspondence bias (assuming students’ behavior or GPA corresponds to their innate traits), graduate school admissions tend to be biased in favor of individuals with high GPAs from lenient grading institutions because they fail to sufficiently discount the role of lenient grading and institutions’ quality.
Its easier to make a linear model avoiding this error
One researcher developed a linear model to make graduate admissions decisions relying on three factors:
Graduate school examination scores
Undergraduate GPA
Quality of undergraduate school
Compared to the decisions of an admissions committee, the model was better at predicting future student performance
One domain in which linear models could improve the decision-making process is admissions decisions.
In a manifestation of what is known as the correspondence bias, graduate school admissions tend to be biased in favor of individuals with high GPAs from lenient institutions because they fail to sufficiently discount the role of lenient grading.
One researcher developed a linear model to make graduate admissions decisions and compared it to the decisions of an admissions committee.
The model was capable of ruling out 55% of applicants who the committee also rejected.
The model also proved more effective at predicting the future ratings of applicants than the committee.
Overall, the effectiveness of this simple model suggests that the implementation of simple models using decision rules to make quick admissions decision can save universities hundreds of millions of dollars in time and resources saved from not having committees review applications.
Decision-analysis tools can also be of assistance in making hiring decisions.
Job interviews predict only 14% of the variability in employee performance. However, people continue to believe in them as an effective diagnostic tool.
Interviewers are biased by the availability of subjective qualities that they associate with effective performance when interviewing potential applicants.
Interviewer affect is influenced by superficial cues associated with a candidate and this influences their perceptions of the candidate.
Individuals who are extroverted, sociable, tall, attractive, and ingratiating are often considered as ones that represent the qualities of effective leaders even though they are far less predictive of performance than less observable traits such as conscientiousness and intelligence.
Managers tend to seek information that confirms the quality of their decision after hiring an applicant, which prevents them from learning. In addition, they never have a sample of performance from those who were interviewed but not hired to use as a comparison group.
Interviews may be difficult to justify based on the time that goes into them and the little predictive validity they have, but if the ratings of interviewers are combined into a linear model with measures of aptitude and past performance, hiring decisions could be improved.
6
Linear Models
Decision-analysis tools can also be of assistance in making hiring decisions.
Job interviews predict only 14% of the variability in employee performance. However, people continue to believe in them as an effective diagnostic tool.
Interviewers are biased by the availability of subjective qualities that they associate with effective performance when interviewing potential applicants.
Interviewer affect is influenced by superficial cues associated with a candidate and this influences their perceptions of the candidate.
Individuals who are extroverted, sociable, tall, attractive, and ingratiating are often considered as ones that represent the qualities of effective leaders even though they are far less predictive of performance than less observable traits such as conscientiousness and intelligence.
Managers tend to seek information that confirms the quality of their decision after hiring an applicant, which prevents them from learning. In addition, they never have a sample of performance from those who were interviewed but not hired to use as a comparison group.
Interviews may be difficult to justify based on the time that goes into them and the little predictive validity they have, but if the ratings of interviewers are combined into a linear model with measures of aptitude and past performance, hiring decisions could be improved.
One domain in which linear models could improve the decision-making process is admissions decisions.
In a manifestation of what is known as the correspondence bias, graduate school admissions tend to be biased in favor of individuals with high GPAs from lenient institutions because they fail to sufficiently discount the role of lenient grading.
One researcher developed a linear model to make graduate admissions decisions and compared it to the decisions of an admissions committee.
The model was capable of ruling out 55% of applicants who the committee also rejected.
The model also proved more effective at predicting the future ratings of applicants than the committee.
Overall, the effectiveness of this simple model suggests that the implementation of simple models using decision rules to make quick admissions decision can save universities hundreds of millions of dollars in time and resources saved from not having committees review applications.
Decision-analysis tools can also be of assistance in making hiring decisions.
Job interviews predict only 14% of the variability in employee performance. However, people continue to believe in them as an effective diagnostic tool.
Interviewers are biased by the availability of subjective qualities that they associate with effective performance when interviewing potential applicants.
Interviewer affect is influenced by superficial cues associated with a candidate and this influences their perceptions of the candidate.
Individuals who are extroverted, sociable, tall, attractive, and ingratiating are often considered as ones that represent the qualities of effective leaders even though they are far less predictive of performance than less observable traits such as conscientiousness and intelligence.
Managers tend to seek information that confirms the quality of their decision after hiring an applicant, which prevents them from learning. In addition, they never have a sample of performance from those who were interviewed but not hired to use as a comparison group.
Interviews may be difficult to justify based on the time that goes into them and the little predictive validity they have, but if the ratings of interviewers are combined into a linear model with measures of aptitude and past performance, hiring decisions could be improved.
7
