Exam 7: Nonparametric Measures of Association

arrow
  • Select Tags
search iconSearch Question
  • Select Tags

Take the distribution from problem 7 and compute the strength and direction of the relationship by computing lambda. Indicate whether lambda is a symmetric or asymmetric measure of association.

Free
(Essay)
4.9/5
(35)
Correct Answer:
Verified

E1=n modal category =12145=76E2=(4231)+(5632)+(2317)=11+24+6=41λPRE=E1E2E1=764176=0.4605\begin{array} { l } E _ { 1 } = n - \text { modal category } = 121 - 45 = 76 \\E _ { 2 } = ( 42 - 31 ) + ( 56 - 32 ) + ( 23 - 17 ) = 11 + 24 + 6 = 41 \\\\\lambda _ { \mathrm { PRE } } = \frac { E _ { 1 } - E _ { 2 } } { E _ { 1 } } = \frac { 76 - 41 } { 76 } = 0.4605\end{array}
Lambda is an asymmetric measure of association.

You want to find the strength of the relationship between one's political affiliation (Democrat/Republican) and one's opinion on gun control (yes, favor; no, opposed). Assume you have asked a random sample of 320 individuals about their political affiliation and their opinion on gun control. The results are to be found in the table below. Calculate the prediction error and interpret your results. Gun Democrats Republicans Totals Control Yes 109 33 142 No 57 121 178 Totals 166 154 320

Free
(Essay)
5.0/5
(34)
Correct Answer:
Verified

E1=n modal category =320178=142E2=(166109)+(154121)=57+33=90 PRE =E1E2E1=14290142=0.3662\begin{array} { c } E _ { 1 } = n - \text { modal category } = 320 - 178 = 142 \\E _ { 2 } = ( 166 - 109 ) + ( 154 - 121 ) = 57 + 33 = 90 \\\text { PRE } = \frac { E _ { 1 } - E _ { 2 } } { E _ { 1 } } = \frac { 142 - 90 } { 142 } = 0.3662\end{array}
The error in predicting favor for gun control is with knowledge of Democrats/Republicans opinion reduced by 36.62%.

Take the distribution from problem 7 and compute the strength and direction of the relationship by computing tau b. Is it even possible to utilize tau b as a measure of association? Explain.

Free
(Essay)
4.8/5
(32)
Correct Answer:
Verified

It is appropriate to utilize tau b as a measure of association because the number of rows and columns in the table are equal.
Ns = 3,043 (value from problem 7)
Nd = 348 (value from problem 7)
Ty = 1,475 (value from problem 9)
Tx for cell A = (31) × (8 + 3) = 341
Tx for cell B = (12) × (32 + 12) = 528
Tx for cell C = (2) × (4 + 17) = 42
Tx for cell D = (8) × (3) = 24
Tx for cell E = (32) × (12) = 384
Tx for cell F = (4) × (14) = 56
Total Tx = 1,375
b=NsNd(Ns+Nd+Ty)(Ns+Nd+Tx)=6043348(3043+348+1475)(3043+348+1375)=26954866×4766=26954815.74=0.5596\begin{array} { c } b = \frac { N _ { s } - N _ { d } } { \sqrt { \left( N _ { s } + N _ { d } + T _ { y } \right) \left( N _ { s } + N _ { d } + T _ { x } \right) } } = \frac { 6043 - 348 } { \sqrt { ( 3043 + 348 + 1475 ) ( 3043 + 348 + 1375 ) } } \\= \frac { 2695 } { \sqrt { 4866 \times 4766 } } = \frac { 2695 } { 4815.74 } = 0.5596\end{array}
Utilizing tau b, you find that the relationship between level of officer/population ratio and the level of overall job satisfaction is moderate (similar to the value you got utilizing Somer's d, but significantly lower than the value you got utilizing gamma).

What are the three questions that we have to ask ourselves when we want to assess association?

(Essay)
4.9/5
(39)

We know from research that the extent to which community policing is implemented (on a 1-10 scale) has an impact on perceived levels of neighborhood safety (on a 1-10 scale). Through statistical tests, you come to the conclusion that the relationship is positive. What does that mean?

(Essay)
4.8/5
(39)

The administration of a local prison is debating implementing a new program tailored to address the challenges of prisoners that score high on both risk and need assessment tools. Prior to implementation of the pilot program, the administration aims to determine the strength of the relationship between high scores on the risk and high scores on the need assessment tool to determine whether the implantation of one program (compared to another program) is even feasible. The results of a random sample of 10 inmates are presented in the table below. Both assessment tools range from 1 to 30, with 1 indicating the lowest level of need and risk and 30 indicating the highest level of need and risk. Compute Spearman's rank-order correlation (rho) and interpret your findings. Prisoner D Need assessment Risk assessment 1001 20 18 1002 10 14 1003 26 30 1004 12 11 1005 11 14 1006 9 13 1007 21 21 1008 30 30 1009 26 24 1010 3 7

(Essay)
4.9/5
(42)

In problem 6 you were interested in the strength of the relationship between one's political affiliation and one's opinion regarding gun control. a. What are the odds that a democrat is in favor of more stringent gun control laws? b. What are the odds that a republican is in favor of more stringent gun control laws? c. Compute the odds ratio and interpret your result. d. Compute Yule's Q and interpret your result.

(Essay)
4.8/5
(39)

We know from theory and research that there is a statistically significant relationship between age and the frequency of offending. You also know that this relationship is negative. Explain.

(Essay)
4.8/5
(31)

Take the distribution from problem 7 and compute the strength and direction of the relationship by computing Somer's d. Indicate whether Somer's d is a symmetric or asymmetric measure of association.

(Essay)
4.9/5
(34)

Assume you have conducted a study in which you were interested in the relationship between the number of police officers/population ratio and the level of overall job satisfaction within the individual police department. You arbitrarily divide the variable officer/population ratio into high, medium, and low (with high having the greatest number of officers per 1,000 inhabitants) and the variable job satisfaction into low, medium, and high. You are utilizing a random sample of 121 police departments from all over the United States. The results are to be found in the table below. Compute gamma and interpret your results. Indicate whether gamma is a symmetric or asymmetric measure of association. Officer/population ratio Job Low Medium High Totals satisfaction Low A B C 45 31 12 2 Medium 44 8 32 4 High 32 3 12 17 Totals 42 56 23 121

(Essay)
5.0/5
(39)

Assume you have been able to determine a statistically significant relationship between treatment while incarcerated and relapse within 1 year after release. Explain the meaning of a conditional distribution in this context.

(Essay)
4.8/5
(36)
close modal

Filters

  • Essay(0)
  • Multiple Choice(0)
  • Short Answer(0)
  • True False(0)
  • Matching(0)