Data
ricci_vs_destefano

ricci_vs_destefano

active ARFF Public Domain (CC0) Visibility: public Uploaded 14-09-2020 by Florian Pfisterer
0 likes downloaded by 0 people , 0 total downloads 0 issues 0 downvotes
Issue #Downvotes for this reason By


Loading wiki
Help us complete this description Edit
DESCRIPTIVE ABSTRACT: The data set contains the oral, written and combined test scores for 2003 New Haven Fire Department promotion exams. The Race and Position for each test taker are also given. SOURCES: The data was obtained from the district court's decision, the briefs and appendices to them. Ricci v. DeStefeno, 554 F. Supp. 2d 142 (United States District Court for the district of Connecticut) VARIABLE DESCRIPTIONS: Columns 1 Race -- W = white, H=Hispanic, B=black 3-12 Position -- Captain or Lieutenant 14-18 Oral -- Oral exam score 20-21 Written -- Written exam score 23-28 Combine -- Weighted total score, with 60% written and 40% oral Values are centered and delimited by blanks. There are no missing values. STORY BEHIND THE DATA: In November and December of 2003, the New Haven Fire Department administered oral and written exams for promotion to Lieutenant and Captain. Under the contract between the City of New Haven and the firefighter's union, the written exam received a weight of 60% and oral exam received a weight of 40%. Applicants with a total score of 70% or above pass the exam and become eligible for promotion. A total of 118 firefighters took the exam. Among them, 77 took the Lieutenant exam, and 41 took the Captain exam. During the time of the exams, there were 8 Lieutenant and 7 Captain positions available. The City Charter of New Haven specifies that when "g" promotions are made, the Department must select them from the top "g+2" scorers. Consequently, top 10 Lieutenant scorers and top 9 Captain scorers are eligible for potential promotion. The City of New Haven decided not to certify the exam and promoted no one, because an insufficient number of minorities would receive a promotion to an existing position. Ricci and other test-takers who would be considered for promotion had the city certified the exam sued the city for reverse discrimination. The District Court decided that the plaintiffs did not have a viable disparate impact claim. The appeals court confirmed the district court's ruling in Feb, 2008. On June 29, 2009, the Supreme Court decided that the City's failure to certify the tests was a violation of Title VII of the Civil Rights Act of 1964. PEDAGOGICAL NOTES: There are three races in the data: white, Hispanic and black. The data can be used for one-way ANOVA on equality of average test scores for three races; two-way ANOVA for equlity of test scores for race and position; chi-square test on the equality of pass rates for three races; Fisher-Freeman-Halton test on the equality of the potential promotion rates. Instructors can also combine blacks and Hispanics as minority. Then the following analysis can be made: two-sample t-test on equality of average test scores for majority v. minority; chi-square test on the equality of pass rates for majority v. minority; Fisher's exact test on equality of the potential promotion rates for majority v.minority. Data obtained from https://github.com/algofairness/fairness-comparison

6 features

Promotion (target)string2 unique values
0 missing
Positionstring2 unique values
0 missing
Oralnumeric81 unique values
0 missing
Writtennumeric37 unique values
0 missing
Racestring3 unique values
0 missing
Combinenumeric113 unique values
0 missing

19 properties

118
Number of instances (rows) of the dataset.
6
Number of attributes (columns) of the dataset.
2
Number of distinct values of the target attribute (if it is nominal).
0
Number of missing values in the dataset.
0
Number of instances with at least one value missing.
3
Number of numeric attributes.
0
Number of nominal attributes.
0.05
Number of attributes divided by the number of instances.
50
Percentage of numeric attributes.
52.54
Percentage of instances belonging to the most frequent class.
0
Percentage of nominal attributes.
62
Number of instances belonging to the most frequent class.
47.46
Percentage of instances belonging to the least frequent class.
56
Number of instances belonging to the least frequent class.
0
Number of binary attributes.
0
Percentage of binary attributes.
0
Percentage of instances having missing values.
1
Average class difference between consecutive instances.
0
Percentage of missing values.

2 tasks

0 runs - estimation_procedure: 50 times Clustering
0 runs - estimation_procedure: 50 times Clustering
Define a new task