Subject Index
917
interpretation of regression
equation, 191
multiple coefficient of correlation, 198
multiple coefficient of determination,
196–197
notation/assumptions, 188–190
partial
regression coefficients, 191–192
standardized variables, regression on,
199–200
Multiple regression analysis, 21
Multiple regression model, 14
Multiple-equation model, 3
Multiplication, matrix, 841–843
Multiplicative effect, 470
Multiplicative form, 287
Mutual fund advisory feeds, 530–531
Mutually exclusive events, 802
MWD test, 260–261
N
N
(number of observations), 21
National Bureau of Economic Research
(NBER), 900
National Trade Data Bank, 901
Natural logarithms, 184, 185
Nature of
X
variables (assumption 7), 68
NBER (National
Bureau of Economic
Research), 900
N.e.d. (normal equivalent deviate), 568
Negative correlation, 66
Neo-classical linear regression model
(NLRM), 63
Nested models, 487
Newey–West method, 441, 447–448
Newton–Raphson iterative method, 530
Newton’s law of gravity, 19
NID (normally and independently
distributed), 98
NLLS (nonlinear least squares), 527
NLRM (
see
Nonlinear regression models)
NLRM (neo-classical linear regression
model), 63
No autocorrelation between disturbances
(assumption 5), 66–67
Nominal level of significance, 475–476
Nominal
regressand, 542
Nominal scale, 28
Nonexperimental data, 25, 27
Nonlinear least squares (NLLS), 527
Nonlinear regression models (NLRM), 38,
39, 525–535
direct optimization, 529
direct search method, 529
estimation of, 527
examples, 530–534
iterative linearization method, 530
linear vs., 525–526
trial-and-error method, 527–529
Non-nested
F
test, 488–489
Non-nested hypotheses tests, 488–492
Davidson–MacKinnon
J
test, 490–492
discerning approach, 488–492
discrimination approach, 488
non-nested
F
test, 488–489
Non-nested models, 487
Non-normal error distribution, 509–510
Non-normality, of disturbances, 544
Nonparametric
statistical methods, 758
Nonparametric tests, 432n
Nonresponse, 27
Nonsense regression, 737
Nonsingular matrix, 844
Nonstationary stochastic processes, 741–744
Nonstationary time series, 741, 760–762
Nonsystematic component, 40
Normal distribution, 143–144, 816–819
Normal equations, 58, 527, 875
Normal equivalent deviate (N.e.d.), 568
Normal probability plot (NPP), 131, 132
Normality (assumption 10), 233–234
for disturbances, 98
properties of OLS estimators under,
100–101
reasons for using, 99–100
of
stochastic distribution, 315, 318
Normality tests, 130–132
histogram of residuals, 130–131
Jarque–Bera test, 131, 132
normal probability plot, 131, 132
Normally and independently distributed
(NID), 98
Normit, 568
Normit model (
see
Probit model)
Not statistically significant, 114
NPP (
see
Normal probability plot)
Nuisance parameters, 596
Nuisance variables, 598
Null hypothesis, 113, 120, 121, 235n, 831
Null matrix, 840
Null vector, 840
Number crunching, 475
Numerator
degrees of freedom, 144
Numerical properties, of estimators, 59
NYSE price changes example, 794–795
O
Observational data:
assumption about, 67–68
experimental vs., 2
quantity of, 67–68
Occam’s razor, 42
Odds ratio, 554
Ohm’s law, 19
OLS (
see
Ordinary least squares)
OLS estimation, 853–858
and autocorrelation, 418–427
and heteroscedasticity, 370–371,
374–376
illustration, 855–856
properties of OLS vector
β
,
858
variance-covariance matrix of
β
,
856–857
OLS estimators, 192–196
derivation of, 227–228
inconsistency of, 679–682
multicollinearity
and variance of,
328–330
properties, 100–101
properties of, 195–196
sensitivity of, 331–332
variances and standard errors of,
194–195
OLS standard-error correction, 447–448
OLS vector, 858
Omission, of relevant variable, 469, 471–473
Omitted category, 281
Omitted variables, 477–482
One-sided hypothesis, 115
One-tail hypothesis test, 115
One-tail test of significance, 117, 118
One-way
fixed effects, 598
Order, 838
Order condition of identifiability, 699–700
Ordinal models, 580
Ordinal regressand, 542
Ordinal scale, 28
Ordinary least squares (OLS), 55–85
(
See also
OLS estimation; OLS
estimators)
assumptions, 61–69
BLUE property of, 875–876
examples of, 81–83
Gauss–Markov theorem, 71–73
GLS vs., 373–374
goodness of fit, 73–78
method of, 55–61
and Monte Carlo experiments, 83–84
precision/standard errors, 69–71
and recursive models, 712–714
Orthogonal polynomials, 346
Orthogonal variables, 355
Outliers, 367, 496–498
Out-of-sample
forecasting, 491
Overall significance testing:
ANOVA, 238–240
F
test, 240–241
incremental contribution of explanatory
variable, 243–246
individual vs. joint, 241
in multiple regression, 237–246
R
2
and
F
relationship, 241–242
in terms of
R
2
, 242–243
Overdifferencing, 761
Overfitting, of model, 473–474
Overidentification, 697–698
Overidentified equation, 718–721
guj75772_index.qxd 05/09/2008 11:14 AM Page 917
