Dorian Allen

November 14, 2022

DECS 431
Professor Deserranno

DECS431 Assignment 7
Part 1:
Question 1:
a) 1182.9666 HP, [1029.1019 HP, 1336.8313 HP]

. reg lnhp enginesize cylinders rpm manual manuales

Source | SS df MS Number of obs = 82

-------------+---------------------------------- F(5, 76) = 196.13

Model | 9.67711682 5 1.93542336 Prob > F = 0.0000

Residual | .749989535 76 .009868283 R-squared = 0.9281

-------------+---------------------------------- Adj R-squared = 0.9233

Total | 10.4271064 81 .128729708 Root MSE = .09934

------------------------------------------------------------------------------

lnhp | Coefficient Std. err. t P>|t| [95% conf. interval]

-------------+----------------------------------------------------------------

enginesize | .2195166 .0334462 6.56 0.000 .1529027 .2861306

cylinders | .0485263 .0192086 2.53 0.014 .010269 .0867835

rpm | .2948873 .0231289 12.75 0.000 .248822 .3409526

manual | -.6735393 .0917133 -7.34 0.000 -.8562022 -.4908764

manuales | .2745843 .0304751 9.01 0.000 .2138878 .3352807

_cons | 2.564791 .1567421 16.36 0.000 2.252612 2.87697

------------------------------------------------------------------------------

lnhp = 2.564791 + 0.2195166*enginesize + 0.0485263*cylinders + 0.2948873*rpm -

0.6735393*manual + 0.2745843*manueles
lnhp = 2.564791 + 0.2195166*3 + 0.0485263*4 + 0.2948873*4000 - 0.6735393*0 +
0.2745843*(0*3)
. di 2.564791 + 0.2195166*3 + 0.0485263*4 + 0.2948873*4000 - 0.6735393*0 + 0.2745

> 843*(0*3)

1182.9666

. kpredint _b[_cons] + _b[enginesize]*3 + _b[cylinders]*4 + _b[rpm]*4000 +

_b[manual]*0 + _b[manuales]*0, level (90)

Estimate: 1182.9666

Standard Error of Individual Prediction: 92.402818

Individual Prediction Interval (90%): [ 1029.1019 , 1336.8313 ]

t-ratio: 12.802279

If Ha: < then Pr(T < t) = 1

If Ha: not = then Pr(|T| > |t|) = 0

If Ha: > then Pr(T > t) = 0

b) Manual: 1184.2991 HP
Automatic: 1183.5998 HP

. kpredint _b[_cons] + _b[enginesize]*5 + _b[cylinders]*8 + _b[rpm]*4000 +

_b[manual]*0 + _b[manuales]*0

Estimate: 1183.5998

Standard Error of Individual Prediction: 92.409543

Individual Prediction Interval (95%): [ 999.55021 , 1367.6493 ]

t-ratio: 12.808198

If Ha: < then Pr(T < t) = 1

If Ha: not = then Pr(|T| > |t|) = 0

If Ha: > then Pr(T > t) = 0

. kpredint _b[_cons] + _b[enginesize]*5 + _b[cylinders]*8 + _b[rpm]*4000 +

_b[manual]*1 + _b[manuales]*1*5

Estimate: 1184.2991

Standard Error of Individual Prediction: 92.399846

Individual Prediction Interval (95%): [ 1000.2689 , 1368.3294 ]

t-ratio: 12.817112

If Ha: < then Pr(T < t) = 1

If Ha: not = then Pr(|T| > |t|) = 0

If Ha: > then Pr(T > t) = 0

Part II:

1.
Using a fixed effect model, we estimate that hiring 10 additional nurses and holding caseweight,
residency, and percent of patients on medicare constant will increase I recommend hiring the 10
additional nurses will increase mortality by 0.00965800666% and decrease the hospital’s bonus
by $48,290.03. We are 95% confident that the change in mortality will be between -0.1039917%
and 0.1233077% and the change in bonus will be between -$519958.5 and $616538.6.
Because of the estimated increase in mortality and decrease in bonus, along with the widespread
95% confidence intervals for mortality and bonus, we recommend that the hospital does not hire
10 more nurses.
. xi: regress mortality nsr caseweight residency pctMedicare i.hospname

i.hospname _Ihospname_1-42 (_Ihospname_1 for hos~e==AC HOSPITAL omitte

> d)

Source | SS df MS Number of obs = 420

-------------+---------------------------------- F(45, 374) = 49.18

Model | 1060.91986 45 23.5759968 Prob > F = 0.0000

Residual | 179.282149 374 .479364034 R-squared = 0.8554

-------------+---------------------------------- Adj R-squared = 0.8380

Total | 1240.20201 419 2.95990932 Root MSE = .69236

-------------------------------------------------------------------------------

mortality | Coefficient Std. err. t P>|t| [95% conf. interval]

--------------+----------------------------------------------------------------

nsr | .1448701 .8669695 0.17 0.867 -1.559876 1.849616

caseweight | .2974003 .1272929 2.34 0.020 .0471009 .5476997

residency | -.2710708 .3474306 -0.78 0.436 -.954233 .4120914

pctMedicare | .1062218 .0163883 6.48 0.000 .0739971 .1384465

_Ihospname_2 | .3037307 .3460192 0.88 0.381 -.3766562 .9841176

_Ihospname_3 | -1.355884 .3401894 -3.99 0.000 -2.024808 -.6869602

_Ihospname_4 | -.4989155 .5099315 -0.98 0.329 -1.501608 .5037766

_Ihospname_5 | -.4372591 .3354039 -1.30 0.193 -1.096773 .2222548

_Ihospname_6 | .165895 .5629181 0.29 0.768 -.9409862 1.272776

_Ihospname_7 | .3994051 .3514217 1.14 0.256 -.291605 1.090415

_Ihospname_8 | 2.128898 .3475411 6.13 0.000 1.445518 2.812277

_Ihospname_9 | .9930904 .4409635 2.25 0.025 .1260119 1.860169

_Ihospname_10 | .9996087 .4060764 2.46 0.014 .2011296 1.798088

_Ihospname_11 | .5629445 .3147577 1.79 0.075 -.0559722 1.181861

_Ihospname_12 | 2.210086 .5138345 4.30 0.000 1.199719 3.220453

_Ihospname_13 | 1.604368 .327361 4.90 0.000 .9606687 2.248066

_Ihospname_14 | .8758878 .3142792 2.79 0.006 .257912 1.493864

_Ihospname_15 | 1.185085 .4668106 2.54 0.012 .2671829 2.102987

_Ihospname_16 | 1.939404 .5292903 3.66 0.000 .8986459 2.980162

_Ihospname_17 | -1.024885 .3364378 -3.05 0.002 -1.686432 -.363338

_Ihospname_18 | 1.598328 .3205018 4.99 0.000 .9681161 2.228539

_Ihospname_19 | -.3150782 .3705605 -0.85 0.396 -1.043721 .413565

_Ihospname_20 | 1.385918 .3688801 3.76 0.000 .6605791 2.111257

_Ihospname_21 | 2.224351 .4784183 4.65 0.000 1.283624 3.165078

_Ihospname_22 | -3.364465 .4720691 -7.13 0.000 -4.292707 -2.436223

_Ihospname_23 | 1.389077 .3331095 4.17 0.000 .7340748 2.044079

_Ihospname_24 | -1.661441 .3487237 -4.76 0.000 -2.347146 -.9757365

_Ihospname_25 | -2.691871 .3665034 -7.34 0.000 -3.412536 -1.971205

_Ihospname_26 | 1.741766 .4818951 3.61 0.000 .7942028 2.68933

_Ihospname_27 | -2.169044 .3869555 -5.61 0.000 -2.929925 -1.408163

_Ihospname_28 | -1.001063 .4410777 -2.27 0.024 -1.868366 -.1337599

_Ihospname_29 | -.5832321 .3695564 -1.58 0.115 -1.309901 .1434366

_Ihospname_30 | .230334 .5285832 0.44 0.663 -.8090335 1.269701

_Ihospname_31 | .6990217 .4861926 1.44 0.151 -.2569921 1.655035

_Ihospname_32 | .185398 .4858674 0.38 0.703 -.7699762 1.140772

_Ihospname_33 | -3.367132 .3810545 -8.84 0.000 -4.11641 -2.617855

_Ihospname_34 | -1.058751 .3264561 -3.24 0.001 -1.70067 -.4168315

_Ihospname_35 | 1.843225 .4893121 3.77 0.000 .881077 2.805372

_Ihospname_36 | -3.08848 .3834834 -8.05 0.000 -3.842533 -2.334426

_Ihospname_37 | .0977259 .3549504 0.28 0.783 -.6002226 .7956745

_Ihospname_38 | .1946616 .4241953 0.46 0.647 -.6394451 1.028768

_Ihospname_39 | 1.252776 .3557511 3.52 0.000 .5532527 1.952299

_Ihospname_40 | .8152586 .3778435 2.16 0.032 .0722946 1.558223

_Ihospname_41 | -3.314349 .4594495 -7.21 0.000 -4.217777 -2.410921

_Ihospname_42 | 1.255065 .3130669 4.01 0.000 .6394735 1.870657

_cons | 2.501244 .5114636 4.89 0.000 1.495539 3.506949

-------------------------------------------------------------------------------

. areg mortality nsr caseweight residency pctMedicare, absorb(hospname)

Linear regression, absorbing indicators Number of obs = 420

Absorbed variable: hospname No. of categories = 42

F(4, 374) = 17.71

Prob > F = 0.0000

R-squared = 0.8554

Adj R-squared = 0.8380

 Root MSE = 0.6924

------------------------------------------------------------------------------

mortality | Coefficient Std. err. t P>|t| [95% conf. interval]

-------------+----------------------------------------------------------------

nsr | .1448701 .8669695 0.17 0.867 -1.559876 1.849616

caseweight | .2974003 .1272929 2.34 0.020 .0471009 .5476997

residency | -.2710708 .3474306 -0.78 0.436 -.954233 .4120914

pctMedicare | .1062218 .0163883 6.48 0.000 .0739971 .1384465

_cons | 2.557183 .5549132 4.61 0.000 1.466043 3.648324

------------------------------------------------------------------------------

F test of absorbed indicators: F(41, 374) = 40.748 Prob > F = 0.000

Mortality = 2.557183 + 0.1448701*nsr + 0.2974003*caseweight – 0.2710708*residency +

0.1062218*pctMedicare
Δmortality = 0.1448701*Δnsr + 0.2974003*Δcasewight – 0.2710708*Δresidency +
0.1062218*Δpctmedicare
Δmortality = 0.1448701*(10/150)
Δmortality = 0.00965800666
Bonus = $5 million*mortality
ΔBonus = $5 million*Δmortality
ΔBonus = $5 million*0.00965800666
ΔBonus = $48,290.0333333

. klincom _b[nsr]*(10/150)

( 1) .0666667*nsr = 0

------------------------------------------------------------------------------

mortality | Coefficient Std. err. t P>|t| [95% conf. interval]

-------------+----------------------------------------------------------------

(1) | .009658 .057798 0.17 0.867 -.1039917 .1233077

------------------------------------------------------------------------------
If Ha: < then Pr(T < t) = .566

If Ha: not = then Pr(|T| > |t|) = .867

If Ha: > then Pr(T > t) = .434

. klincom _b[nsr]*(10/150)*5000000

( 1) 333333.3*nsr = 0

------------------------------------------------------------------------------

mortality | Coefficient Std. err. t P>|t| [95% conf. interval]

-------------+----------------------------------------------------------------

(1) | 48290.05 288989.8 0.17 0.867 -519958.5 616538.6

------------------------------------------------------------------------------

If Ha: < then Pr(T < t) = .566

If Ha: not = then Pr(|T| > |t|) = .867

If Ha: > then Pr(T > t) = .434

2.
a. Within hospitals: As nsr increases, on average, caseweight increases as well. We gather this by
looking at the coefficient for nsr (0.1448701) with a fixed effects regression model that holds
hospitals constant.
b. Across hospitals: As nsr increases, on average, caseweight decreases. We gather this by
looking at the coefficient for nsr (-6.292554) for a regression model that does not hold hospitals
constant.

. areg mortality nsr caseweight residency pctMedicare, absorb(hospname)

Linear regression, absorbing indicators Number of obs = 420

Absorbed variable: hospname No. of categories = 42

 F(4, 374) = 17.71

Prob > F = 0.0000

R-squared = 0.8554

Adj R-squared = 0.8380

Root MSE = 0.6924

------------------------------------------------------------------------------

mortality | Coefficient Std. err. t P>|t| [95% conf. interval]

-------------+----------------------------------------------------------------

nsr | .1448701 .8669695 0.17 0.867 -1.559876 1.849616

caseweight | .2974003 .1272929 2.34 0.020 .0471009 .5476997

residency | -.2710708 .3474306 -0.78 0.436 -.954233 .4120914

pctMedicare | .1062218 .0163883 6.48 0.000 .0739971 .1384465

_cons | 2.557183 .5549132 4.61 0.000 1.466043 3.648324

------------------------------------------------------------------------------

F test of absorbed indicators: F(41, 374) = 40.748 Prob > F = 0.000

. regress mortality nsr caseweight residency pctMedicare

Source | SS df MS Number of obs =

> 420

-------------+---------------------------------- F(4, 415) =

> 27.53

Model | 260.066352 4 65.0165879 Prob > F =

> 0.0000

Residual | 980.135654 415 2.36177266 R-squared =

> 0.2097

-------------+---------------------------------- Adj R-squared =

> 0.2021

Total | 1240.20201 419 2.95990932 Root MSE =

> 1.5368

---------------------------------------------------------------------

> ---------
 mortality | Coefficient Std. err. t P>|t| [95% conf.

> interval]

-------------+-------------------------------------------------------

> ---------

nsr | -6.292554 .7977454 -7.89 0.000 -7.86068

> -4.724429

caseweight | .6231484 .1189139 5.24 0.000 .3893997

> .8568971

residency | 1.08284 .1757261 6.16 0.000 .737416

> 1.428264

pctMedicare | .0812151 .0145813 5.57 0.000 .0525526

> .1098776

_cons | 5.929582 .432559 13.71 0.000 5.079303

> 6.779862

---------------------------------------------------------------------

> ---------