Statistical Solutions for Expenditures, Credit Scores, and Test Scores, Study notes of Economic statistics

Solutions to various statistical problems, including calculating the budget level that bob spends below 90% of the time, the credit score value that 75% of applicants exceed, and statistical properties of test scores. It uses the normal distribution and z-tables to find the answers.

Typology: Study notes

Pre 2010

Uploaded on 02/13/2009

koofers-user-ms1
koofers-user-ms1 🇺🇸

10 documents

1 / 2

Toggle sidebar

This page cannot be seen from the preview

Don't miss anything!

bg1
Solutions PS III
1. Let Ebe expenditures
E= 80
E= 10 )2
E= 100
EsN(80;100)
Let Ebe the target budget level. We want Esuch that:
P(EE) = 0:90
)P(EE
EEE
E) = P(zE80
10 ) = 0:90:Let z=E80
10
From the ztable, the value of zthat corresponds to this probability is
z= 1:285
)z=E80
10 = 1:285
)E= 1:285(10) + 80 = 92:85
With this budget, Bob will spend less that this amount 90% of the time.
2. Let the credit score SsN(600;1002)
(i) Let Sbe the value of the credit score that 75% of the customers score
above. We want Ssuch that
P(SS) = 0:75
)P(SS
SS600
100 ) = 0:75
)P(zz) = 0:75;where z=S600
100
)1P(zz) = 1 0:75 = 0:25
The corresponding value of zfrom the ztable is 0:68
)S600
100 =0:68
)S= 600 68 = 532
75% of the applicants will have a credit score exceeding 532.
(ii) For this problem, we consider 80% of the area symmetric about the mean.
P(SLSSU) = 0:80
)P(SSU)P(SSL) = 0:80
)P(zSU600
100 )P(zSL600
100 ) = 0:80
Now since we are looking at an 80% area symmetric about the mean, and
since the normal is a symmetric distribution,
P(zSU600
100 ) = P(zSL600
100 ) = 0:10
Working with the lower limit, let SL600
100 =z:From the ztables, the cor-
responding value of zsuch that P(zz) = 0:10 is 1:28
)z=SL600
100 =1:28
)SL= 600 128 = 472
For the upper limit:
P(zSU600
100 ) = 0:10 )P(zSU600
100 ) = 0:90
Let z=SU600
100 . Then the corresponding value of zfrom the table is 1:28
)SL= 600 + 128 = 728
3. Let Tbe the score
E(T) = 25; (T) = 6; 2
Tn= 36
1
pf2

Partial preview of the text

Download Statistical Solutions for Expenditures, Credit Scores, and Test Scores and more Study notes Economic statistics in PDF only on Docsity!

Solutions PS III

  1. Let E be expenditures E = 80 E = 10 )  2 E = 100 E s N (80; 100) Let E^ be the target budget level. We want E^ such that: P (E  E) = 0: 90

) P (

EE E

EE E ) = P (z  E 80 10 ) = 0: 90 : Let z^ = E 80 10 From the z table, the value of z^ that corresponds to this probability is

z  = 1: 285 ) z  = E 80 10

) E

 = 1:285(10) + 80 = 92: 85 With this budget, Bob will spend less that this amount 90% of the time.

  1. Let the credit score S s N (600; 100 2 ) (i) Let S  be the value of the credit score that 75% of the customers score

above. We want S  such that P (S  S  ) = 0: 75

) P (

SS S ^

S 600 100 ) = 0:^75 ) P (z  z  ) = 0: 75 ; where z  = S 600 100 ) 1 P (z  z  ) = 1 0 :75 = 0: 25 The corresponding value of z  from the z table is 0 : 68 ) S

 600 100

) S^ = 600 68 = 532

75% of the applicants will have a credit score exceeding 532.

(ii) For this problem, we consider 80% of the area symmetric about the mean. P (SL  S  SU ) = 0: 80 ) P (S  SU ) P (S  SL) = 0: 80 ) P (z  SU 600 100 ) P (z  SL 600 100

Now since we are looking at an 80% area symmetric about the mean, and

since the normal is a symmetric distribution,

P (z  SU 600 100 ) =^ P^ (z^ ^

SL 600 100 ) = 0:^10 Working with the lower limit, let SL 600 100 = z: From the z tables, the cor-

responding value of z^ such that P (z  z) = 0: 10 is 1 : 28

) z^ = SL 600 100

) SL = 600 128 = 472

For the upper limit: P (z  SU 600 100 ) = 0: 10 ) P (z  SU 600 100

Let z^ = SU 600 100

. Then the corresponding value of z^ from the table is 1 : 28 ) SL = 600 + 128 = 728

  1. Let T be the score E(T ) = 25; (T ) = 6; ^2 T n = 36

Then for Hasbin:

E(H) = E(T + 50) = 25 + E(50) = 75

 2 (H) =V ar(T + 25) = V ar(T ) + 0 = 36

(H) = 6 = (T )

For Upstart:

E(H) = E(3T ) = 3E(T ) = 3(25) = 75

^2 (U ) = V ar(3T ) = 9V ar(T ) = 9  36 = 324

(U ) = 18

  1. C = a + bF

F;C =

Cov(F;C) p V ar(F )

p V ar(C)

: Consider the numerator

Cov(F; C) = Cov(F; a + bF ) = Cov(a; F ) + Cov(bF; F )

= 0 + Cov(F; bF ) = E(F  bF ) E(F )E(bF ) (by the covariance formula)

= E(bF 2 ) E(F )  bE(F )

bE(F 2 ) b(E(F )) 2 = b[E(F 2 ) (E(F )) 2 ] = bV ar(F )

Now consider the denominatorp

V ar(F )

p V ar(C) =

p V ar(F )

p V ar(a + bF )

=

p V ar(F )

p V ar(a) + b^2 V ar(F )

Since a constant has zero variance,

=

p V ar(F )

p b^2 V ar(F ) = b[

p V ar(F )] 2 = bV ar(F )

So the numerator and denominator are identical. QED:

5. E(T ) = E(M + V ) = E(M ) + E(V ) = 510 + 475 = 985

V ar(T ) = V ar(M + V ) = V ar(M ) + V ar(V ) + 2Cov(M; V )

M V = Cov(M;V ) M V

Cov p (M;V ) 750

p 610 ) Cov(M; V ) = 270: 552

V ar(T ) = 750 + 610 + 2(270:552) = 1901: 10