Homework -- Bivariate Frequency Distribution
For each problem, identify the theory, the null hypothesis, answer the questions, and report the effect size
----------------------------------------------------------------------------------------
Problem 1: The students in the Corps of Cadets are faced with a challenge, a pass-fail assignment. Of the 56 people in the class, (comprised of 24 females and 32 males) 11 females failed and 23 males failed. Create the tables necessary to inform the community about the difference in the dependent variable (pass-fail) between the two levels of the independent variable (male and female cadets), and compute MOE to analyze if gender improved the predictability of pass fail rate among cadets.
----------------------------------------------------------------------------------------
What is the probability of randomly selecting a female cadet who failed from the population?
A. 20% B. 46% C. 21% D. 9% E. 0%
What is the best prediction of whether a cadet (male and female) passed the assignment?
A. Most cadets fail
B. Most Cadets pass
C. Impossible to tell from information given
D. Pass-Fail rate is pretty much even
E. None of the above
Compare male cadets who passed to female cadets who passed.
A. Females 54%, Males 28%
B. Males 54%, Females 46%
C. Males 46%, Females 72%
D. Males 100%, Females 0%
E. Males 46%, Females 54%
What is the magnitude of effect of gender?
A. 20%
B. 46%
C. 21%
D. 9%
E. 0%
----------------------------------------------------------------------------------------
Problem 2: A group of experienced police officers is under the impression that Latino drivers are less likely to exceed the speed limit than Caucasian drivers. They set up an observation point on a stretch of highway and observed 5600 drivers over a two-week period (3200 Caucasian and 2400 Latino). Of that total, 1100 drivers were ticketed for speeding (700 were Caucasian and 400 were Latino). Create a bivariate frequency distribution comparing the dependent variable (ticketed - not ticketed) between the two levels of the independent variable (Caucasian and Latino drivers), and compute MOE to determine if race improves the predictability of which drivers are more likely to speed between those two ethnicities.
----------------------------------------------------------------------------------------
What is the best prediction of whether a driver will be ticketed for speeding?
A. Most are ticketed B. Most are not ticketed C. No way to tell
What is the probability of randomly selecting a Latino driver who was ticketed for speeding?
A. 22% B. 17% C. 78% D. 83% E. 0%
Compare ticketing among Caucasian drivers and Latino drivers.
A. 78% Caucasian - 83% Latino
B. 29% Caucasian - 51% Latino
C. 33% Caucasian - 66% Latino
D. 22% Caucasian - 17% Latino
E. 100% Caucasian - 0% Latino
What is the magnitude of effect of race.
A. 12% B. 51% C. 62% D. 0%
Problem 3: A new infant behavior analysis device has demonstrated the ability to foretell the potential for criminality among newborns in suburban and inner city hospitals. Out of 78 suburb babies, 23 were positively identified, and 55 were negative. Inner-city positives numbered 90, with 51 negative. The sample totaled 219. Which neighborhood is more likely to produce criminals (dependent variable = Positive or Negative ID, independent variable = Inner-city or Suburbs)? Does knowledge about which neighborhood the baby is from improve the ability to predict whether or not a given individual will be a criminal as shown by MOE?
---------------------------------------------------------------------------------------
What is the best prediction of whether a newborn will be a criminal?
A. Most are not criminals
B. Most are criminals
C. No way to tell
What is the probability of randomly selecting a Suburban infant who is criminal?
A. 64% B. 29% C. 36% D. 71% E. 0%
Compare criminality among Inner-city and Suburban infants
A. Criminal 78%, Honest Citizen 83%
B. Criminal 41%, Honest Citizen 12%
C. Criminal 64%, Honest Citizen 29%
D. Criminal 82%, Honest Citizen 57%
E. None of the Above
What is the magnitude of effect of environment on criminality?
A. 30%
B. 70%
C. 21%
D. 54%
---------------------------------------------------------------------------------------
Problem 4: Two specialty units with the Memphis police department (Independent variable = the Narcotics unit and the Vice squad) are in competition to find which unit is more efficient at making arrests. Arrest efficiency is determined by whether or not a conviction is obtained (dependent variable = conviction or dismissed). The two divisions produced 926 arrests last year (vice = 447, Narc = 479). Vice scored 306 convictions, Narcotics scored 204).
---------------------------------------------------------------------------------------
What is the best prediction of whether an arrest will end up as a conviction?
A. Most are dismissed
B. Most are convicted
C. No way to tell
What is the probability of randomly selecting a vice arrest who got convicted?
A. 43% B. 68% C. 57% D. 32% E. 0%
Compare convictions among Narcotics and Vice?
A. Narcotics 57%, Vice 32%
B. Narcotics 23%, Vice 58%
C. Narcotics 41%, Vice 60%
D. Narcotics 43%, Vice 68%
What is the magnitude of effect of the arresting unit on conviction?
A. 20% B. 17% C. 10% D. 0%
Problem 5: Are Asian immigrant teens more likely to join gangs in America than Latino teens? A social worker believes so, and asks Asian immigrant and Latino teens if they had been solicited to become members of an ethnic gang (Dependent variable = solicited / not solicited, independent variable = Asian / Latino). Of 326 Asian immigrant teens 228 had been solicited. Latino teens numbered 78 in the affirmative category out of 185.
---------------------------------------------------------------------------------------
What is the best prediction of whether a youth will be solicited to join a gang?
A. Most are not solicited
B. Most are solicited
C. No way to tell
What is the probability of randomly selecting a Latino who got solicited to join a gang?
A. 70% B. 42% C. 58% D. 30% E. 0%
Compare solicitations among Asian and Latino teens?
A. Asians 79%, Latinos 62%
B. Asians 30%, Latinos 68%
C. Asians 27%, Latinos 74%
D. Asians 70%, Latinos 42%
Does knowing race improve ability to predict solicitations?
A. Yes
B. No
What is the magnitude of effect of race on gang solicitations?
A. 21%
B. 14%
C. 86%
D. 34%
----------------------------------------------------------------------------------------
Problem 6: The eligible driving population in millions (which were persons age 16 and over), counting only African Americans and Caucasians, was 181.8 (Blacks 24.5, White 157.3). The actual licensed drivers in that group was 158.9 (Black = 18.1, White = 140.8). Create bivariate distribution tables (freq and %) for two nominal variables (dependent variable = license / no license, independent variable = African American / Caucasian) and determine Lambda.
----------------------------------------------------------------------------------------
What is the best prediction of whether an eligible driver will be a licensed driver?
A. Most are not licensed
B. Most are licensed
C. No way to tell
What is the probability of randomly selecting a African Americans who will be a licensed driver?
A. 74% B. 90% C. 26% D. 10% E. 0%
Compare licensing among African Americans and Caucasians.
A. African American 74%, Caucasian 90%
B. African American 10%, Caucasian 90%
C. African American 24%, Caucasian 59%
D. African American 14%, Caucasian 30%
What is the magnitude of effect of race on licensing?
A. 0% B. 25% C. 18% D. 26%
----------------------------------------------------------------------------------------
Problem 7: Among African Americans and Caucasian drivers (Black = 18.1, White = 140.8), police accomplished 17.1 vehicle stops during the year (White = 14.9, Black = 2.2). The phenomenon commonly forwarded by sociologists is that police initiate vehicle stops for the crime “driving while Black” (DWB). Conduct a study to ascertain if the data supports or dispels the phenomenon. Create bivariate distribution tables (freq and %) for two nominal variables to measure the difference in the dependent variable (stopped / not stopped) between the two levels of the independent variable (African American / Caucasian) and compute MOE.
---------------------------------------------------------------------------------------
What is the best prediction of whether a driver will be stopped?
A. Most are not stopped
B. Most are stopped
C. No way to tell
Compare vehicle stops among African American drivers and Caucasian drivers?
A. African American 78%, Caucasian 51%
B. African American 32%, Caucasian 21%
C. African American 88%, Caucasian 19%
D. African American 12%, Caucasian 11%
What is the magnitude of effect of race on vehicle stops?
A. 22%
B. 63%
C. 0%
D. 9%
Problem 8: Of the total eligible drivers equaling 186.2 million (Male = 95.4, Female = 90.8), there were a total of 78.9 vehicle stops (Male = 50.4, Female = 28.5). You are in charge of determining if there is a disparity among the independent variable (males and females) concerning the dependent variable (traffic stops)? Create bivariate distribution tables (freq and %) for two nominal variables and compute MOE.
---------------------------------------------------------------------------------------
What is the best prediction of whether a male or female driver will be stopped?
A. Most are not stopped
B. Most are stopped
C. No way to tell
Compare vehicle stops among male and female drivers?
A. Female 11% Male 43%
B. Female 21% Male 59%
C. Female 31% Male 53%
D. Female 43% Male 63%
What is the magnitude of effect of gender on vehicle stops?
A. 27%
B. 17%
C. 70%
D. 7%
Group Study 2
----------------------------------------------------------------------------------------
It is widely believed that police dogs, although effective at apprehending criminals, leave the organization they are attached to at greater liability for lawsuits if the dog injures the arrestee during the apprehension process. Defenders of the canine unit typically site the number of law suits against units who do not use dogs, claiming that the presence of a dog does not necessarily make a lawsuit more likely than arrests without a dog. A random sample of departments nationwide reveal that 251 arrests made with a dog produced 100 lawsuits, 151 had no lawsuits. Out of 207 arrests with a no dog, 150 had lawsuits and 57 did not. Ascertain if the independent variable (dog / no dog) puts the unit at greater risk on the dependent variable (lawsuit / no lawsuit) and compute the Magnitude of Effect (MOE).
--------------------------------------------------------------------------------------