### Citations

3021 |
Principles and practice of structural equation modeling (2nd ed
- Kline
- 2005
(Show Context)
Citation Context ...f data requires that the process be stable (not trending over time) and have achieved an equilibrium – conditions that not only may be unrealistic, but for which there is not a good statistical test (=-=Kline, 2005-=-). Structural equations models have other limitations as well. For example, identifiability requirements may limit the specifications that can be empirically tested, and it is possible for several dif... |

2034 |
Limited-Dependent and Qualitative Variables in Econometrics
- Maddala
- 1983
(Show Context)
Citation Context ...ood), we observe TB in either case, not only if RC = 1. A more general form of the sample selection model, referred to as a switching regression model with endogenous switching, is needed (Lee, 1983; =-=Maddala, 1983-=-; Heckman, 1990): RC fRðBE; X ; ZÞþeR; RC = 1 (urban neighborhood chosen) if and only if RC * P 0; otherwise RC = 0 (suburban neighborhood chosen), TBU fUðBE; X ; Y ÞþeU; and ð7Þ TBS fSðBE; X ; ... |

1690 | Instrumental variables regression with weak instruments
- Staiger, Stock
- 1997
(Show Context)
Citation Context ...ng device to identify a suitable set of instruments, finding that higher relevance is likely to be achieved at the expense of lower exogeneity. Nevertheless, several authors (e.g. Bound et al., 1995; =-=Staiger and Stock, 1997-=-) urge that R 2 s and/or Fstatistics from the first-stage regression of instruments against endogenous explanatory variables be routinely reported, as a basis for judging instrument relevance.(2) Hav... |

1485 |
and Telephone Surveys: The Total Design Method
- Dillman, Mail
- 1978
(Show Context)
Citation Context ...orms. Of course, these biases are also possible with the design of the self-administered questionnaires from which the data for quantitative analyses are often collected. However, some scholars (e.g. =-=Dillman, 1978-=-) suggest that all else equal, the extent of at least the latter three biases could be more severe in the case of direct questioning, where the body language and tone of the interviewer can offer addi... |

1194 |
Problems with Instrumental Variables Estimation when the Correlation between the Instruments and the Endogenous Explanatory Variable Is Weak’.
- Bound, Jaeger, et al.
- 1995
(Show Context)
Citation Context ...uments’’ 2 occurs quite often with this technique, and has a number of (related) potential deleterious consequences: (1) The standard error of the coefficient of c BE in Eq. (6) is likely to be high (=-=Bound et al., 1995-=-; Shea, 1997), reflecting the imprecision with which the true effect of BE on TB is being captured. In that case, finding cBE to be insignificant may not reflect a true lack of influence after control... |

1146 |
Discrete Choice Analysis: Theory And Application To Travel Demand
- Ben-Akiva, Lerman
- 1985
(Show Context)
Citation Context ...l as on mode choice. 5 This category can be further subdivided into two: ‘‘sequential’’ and simultaneous models. The ‘‘sequential’’ approach is represented by the multidimensional nested logit model (=-=Ben-Akiva and Lerman, 1985-=-), where both choices are treated as nominal, and in which one choice (most naturally, TB) is conditioned on the other (RC) so that the joint probability of an (RC, TB) bundle being chosen is modeled ... |

1084 |
Identification and Estimation of Local Average Treatment Effects,” Econometrica
- Imbens, Angrist
- 1994
(Show Context)
Citation Context ...h characteristics x, ifZ changes from z0 to z1. That is, loosely speaking, what is the effect for those who needed an incentive to move from a suburban neighborhood to an urban one? This is the LATE (=-=Imbens and Angrist, 1994-=-; Imbens, 2001); it focuses on the effectiveness of the incentive by eliminating those who would have moved anyway (but see Winship and Morgan, 1999 p. 685 for some problems of the LATE approach). In ... |

289 |
Instrument relevance in multivariate linear models: a simple measure
- Shea
- 1997
(Show Context)
Citation Context ...ite often with this technique, and has a number of (related) potential deleterious consequences: (1) The standard error of the coefficient of c BE in Eq. (6) is likely to be high (Bound et al., 1995; =-=Shea, 1997-=-), reflecting the imprecision with which the true effect of BE on TB is being captured. In that case, finding cBE to be insignificant may not reflect a true lack of influence after controlling for sel... |

270 | Generalized econometric models with selectivity.
- Lee
- 1983
(Show Context)
Citation Context ...) neighborhood), we observe TB in either case, not only if RC = 1. A more general form of the sample selection model, referred to as a switching regression model with endogenous switching, is needed (=-=Lee, 1983-=-; Maddala, 1983; Heckman, 1990): RC fRðBE; X ; ZÞþeR; RC = 1 (urban neighborhood chosen) if and only if RC * P 0; otherwise RC = 0 (suburban neighborhood chosen), TBU fUðBE; X ; Y ÞþeU; and ð7Þ TB... |

225 |
Varieties of selection bias‘,
- Heckman
- 1990
(Show Context)
Citation Context ...e TB in either case, not only if RC = 1. A more general form of the sample selection model, referred to as a switching regression model with endogenous switching, is needed (Lee, 1983; Maddala, 1983; =-=Heckman, 1990-=-): RC fRðBE; X ; ZÞþeR; RC = 1 (urban neighborhood chosen) if and only if RC * P 0; otherwise RC = 0 (suburban neighborhood chosen), TBU fUðBE; X ; Y ÞþeU; and ð7Þ TBS fSðBE; X ; Y ÞþeS; where T... |

192 |
Structural equation modeling with LISREL: Essentials and advances
- Hayduk
- 1988
(Show Context)
Citation Context ... interpretation of a coefficient as representing the marginal effect on the target variable of changing another variable by one unit, holding all other variables constant, is essentially meaningless (=-=Hayduk, 1987-=-). Changing BE would change TB, which would change AT – so AT could not be held constant. For the most robust SEMs, then (i.e. those allowing multiple directions of causality), it is quite difficult t... |

177 |
The Distribution of the Instrumental Variable Estimator and its t Ratio When the Instrument is a Poor One,"
- NELSON, STARTZ
- 1990
(Show Context)
Citation Context ...t in Eq. (6), even when the true effect of BE is zero, and ‘‘[t]hus it is in the cases where least squares is a poor estimator that instrumental variables with a poor instrument will be even worse’’ (=-=Nelson and Startz, 1990-=- p. S125; also see Hall et al., 1996). Finally, independently of the weak instruments problem, special account needs to be taken of the sampling variance in c BE, or else incorrect statistical inferen... |

157 | Qualitative Choice Analysis: Theory, Econometrics, and an Application to Automobile Demand - Train - 1986 |

148 | Structural equations, treatment effects and econometric policy evaluation”.
- Heckman, Vytlacil
- 2005
(Show Context)
Citation Context ...retely measured new medical treatments or public health-promotion programs or policies. The recent theoretical and applied econometric literature on policy evaluation (e.g., Winship and Morgan, 1999; =-=Heckman and Vytlacil, 2005-=-) offers a deepening understanding of this framework. The basic scenario in this literature is that there is a discrete treatment (e.g. a new policy), which is chosen by or applied to some of the popu... |

139 | The decomposition of effects in path analysis.
- Alwin, Hauser
- 1975
(Show Context)
Citation Context ...and the (3) direct and (4) indirect effects of BE on TB (for an example, see the blockrecursive model analyzed by Wolfle, 1980 and let his X 1 represent our TB, his X 3 be BE, and X 4 be AT; also see =-=Alwin and Hauser, 1975-=-). In Fig. 2 there are no indirect effects of BE on TB (i.e. those occurring through the impact of BE on an intervening variable that then affects TB), so the total effect, which is the sum of the dir... |

136 |
A micro-analysis of land use and travel in five neighborhoods in the San Francisco Bay Area.
- Kitamura, Mokhtarian, et al.
- 1997
(Show Context)
Citation Context ...work. If so, we are likely to overestimate the influence of built environment elements on travel behavior when we use land use policies to try to reduce travel, fuel consumption, and emissions (e.g., =-=Kitamura et al., 1997-=-). In the past few years, this complex issue has been addressed in a variety of ways. This paper describes and critiques the various methodological approaches adopted to date to assess the causal impa... |

131 |
The estimation of causal effects from observational data
- Winship, Morgan
- 1999
(Show Context)
Citation Context ... an influence of the BE that is purged of the self-selection attitudinal component. Thus, the statistical control and IV methods represent opposite strategies in dealing with the endogeneity problem (=-=Winship and Morgan, 1999-=-): whereas the object of the former method is to identify variables that are maximally correlated with e to use as controls, the object of the latter method is to find variables that are minimally cor... |

130 |
Basic Principles of Structural Equation Modelling,
- Mueller
- 1996
(Show Context)
Citation Context ... were modeled as affecting AT (consistent with Fig. 1b) as well as the converse, computation of the total effect of BE on TB becomes more complex, and also includes the effect of AT on BE (see, e.g., =-=Mueller, 1996-=-). In such cases, the familiar regression-model interpretation of a coefficient as representing the marginal effect on the target variable of changing another variable by one unit, holding all other v... |

125 | Travel by Design: The Influence of Urban Form on Travel. - Boarnet, Crane - 2001 |

119 |
Introductory Econometrics with Applications.
- Ramanathan
- 1995
(Show Context)
Citation Context ...lues. Furthermore, the conventionally-estimated standard errors of the estimated coefficients will also be biased, which renders invalid the usual hypothesis-testing on the significance of variables (=-=Ramanathan, 2002-=-). Endogeneity bias can occur in two conceptually distinct ways, either of which could arise in our context. Simultaneity bias is produced when an ‘‘explanatory’’ variable is simultaneously a function... |

100 | Travel and the Built Environment: A Synthesis - Ewing, Cervero - 2001 |

95 |
Does neighborhood design influence travel?: A behavioral analysis of travel diary and GIS data.
- Crane, Crepeau
- 1998
(Show Context)
Citation Context ...o-traditional’’, or ‘‘new urbanist’’) neighborhoods tend to walk more and drive less than do inhabitants of lower-density, single-use residential (‘‘suburban’’) areas (e.g., Cervero and Duncan, 2003; =-=Crane and Crepeau, 1998-=-; Frank et al., 2006). What is less well understood is the extent to which the observed patterns of travel behavior can be attributed to the residential built environment itself, as opposed to the pri... |

95 | Residential Relocation and Changes in Urban Travel: Does Neighborhood-Scale Urban Form - Krizek - 2003 |

93 | Examining the impacts of residential self-selection on travel behavior: a focus on methodologies.
- Mokhtarian, Cao
- 2008
(Show Context)
Citation Context ...ior. A companion paper (Cao et al., submitted for publication) focuses more heavily on the empirical studies employing each approach, describing and evaluating their key findings. A companion report (=-=Cao et al., 2006-=-) contains the essential content of both papers, plus considerable additional detail that could not be incorporated into journal-length articles. In particular, it includes a table summarizing the 28 ... |

80 | A Comprehensive Analysis of Built Environment Characteristics on Household Residential Choice and Auto Ownership Levels
- Bhat, Guo
- 2007
(Show Context)
Citation Context ...ce in c BE, or else incorrect statistical inferences on the significance of its coefficient in the TB model may result. The corrections needed are especially tedious when the TB variable is discrete (=-=Bhat and Guo, 2007-=-). 3.4. Sample selection models P.L. Mokhtarian, X. Cao / Transportation Research Part B 42 (2008) 204–228 213 The basic idea behind this approach is to explicitly model the prior selection into (or p... |

80 | Local Shopping as a Strategy for Reducing Automobile Travel.
- Handy, Clifton
- 2001
(Show Context)
Citation Context ...oward certain travel modes and land use configurations. For example, residents who prefer walking may consciously choose to live in neighborhoods conducive to walking, and thus walk more (as found by =-=Handy and Clifton, 2001-=-). Therefore, the observed differences in pedestrian behavior in those two types of neighborhoods may be more a matter of residential choice than travel choice. In other words, residential self-select... |

72 | The built environment as a determinant of walking behavior: analyzing non-work pedestrian travel in Portland, - Greenwald, Boarnet - 2001 |

72 |
Latent Variable Structural Equation Modeling with Categorical Data
- MUTHÉN
- 1983
(Show Context)
Citation Context ...types of simultaneous models is possible – that is, a system of structural equations with discrete (binary or ordinal) endogenous variables (see, e.g., Lee, 1981; Lewbel, 2004; Louviere et al., 2005; =-=Muthén, 1983-=-) – applications involving such systems are still relatively rare and we are not aware of any in the present context. 3.5.1. Joint discrete choice models In joint discrete choice models, the observed ... |

71 |
Self-selection in the relationship between the built environment and walking.
- Handy, Cao, et al.
- 2006
(Show Context)
Citation Context ...ich has numerous influences both systematic and idiosyncratic. For the contribution of the BE to TB to be small would not render pointless any attempt to reshape the BE – as discussed elsewhere (e.g. =-=Handy et al., 2006-=-), there are many reasons for improving the BE beyond influencing travel behavior (such as increasing the diversity of available housing options), and even small contributions can be useful at the mar... |

69 | The impact of residential neighborhood type on travel behaviour: A structural equations modeling approach. - Bagelym, Mokhtarian - 2002 |

67 | Many Pathways from Land Use to Health: Associations Between Neighborhood Walkability and Active Transportation, Body Mass Index and Air Quality.
- Frank, Sallis, et al.
- 2006
(Show Context)
Citation Context ... urbanist’’) neighborhoods tend to walk more and drive less than do inhabitants of lower-density, single-use residential (‘‘suburban’’) areas (e.g., Cervero and Duncan, 2003; Crane and Crepeau, 1998; =-=Frank et al., 2006-=-). What is less well understood is the extent to which the observed patterns of travel behavior can be attributed to the residential built environment itself, as opposed to the prior self-selection of... |

60 |
Model specification: Procedures, strategies, and related issues
- MacCallum
- 1995
(Show Context)
Citation Context ... Transportation Research Part B 42 (2008) 204–228 specifications – representing substantively distinct behavioral processes, with different policy implications – to fit the data roughly equally well (=-=MacCallum, 1995-=-; although the same can be true for nested logit models as well – see, e.g., Forinash and Koppelman, 1993). Further, structural equations models are not well-suited to situations where one or more end... |

59 |
Simultaneous Equation Models with Discrete and Censored Dependent Variables. In
- Lee
- 1981
(Show Context)
Citation Context ...lthough in principle a combination of these two types of simultaneous models is possible – that is, a system of structural equations with discrete (binary or ordinal) endogenous variables (see, e.g., =-=Lee, 1981-=-; Lewbel, 2004; Louviere et al., 2005; Muthén, 1983) – applications involving such systems are still relatively rare and we are not aware of any in the present context. 3.5.1. Joint discrete choice mo... |

57 | Bicycling, and Urban Landscapes: Evidence From the San Francisco Bay Area. - Walking - 2003 |

53 |
Analyzing Censored and Sample-Selected Data with Tobit and Heckit Models. Political Analysis,
- Sigelman, Zeng
- 1999
(Show Context)
Citation Context ... dealing with res4 At the other extreme, for the first formulation of the selection model above, if the selection and participation equations are identically specified, the Tobit model results (e.g., =-=Sigelman and Zeng, 1999-=-).P.L. Mokhtarian, X. Cao / Transportation Research Part B 42 (2008) 204–228 215 idential self-selection for several decades. For example, Horowitz (1986) reported on a 1976 multinomial logit estimat... |

52 | Travel behavior in neotraditional neighborhood developments: A case study
- Khattak, Rodriguez
- 2005
(Show Context)
Citation Context ...context as the ‘‘participation’’ or ‘‘selection’’ equation(s). The latter equation is typically a discrete choice model in the sample selection context, but the same can be true for an IV model (e.g. =-=Khattak and Rodriguez, 2005-=-), although they are more often linear regressions on continuous dependent variables. However, whereas in a classic IV model the instrumental predictors of BE should be variables that are not expected... |

38 | Land Use Impacts on Transport: How Land Use Factors Affect Travel Behavior, Victoria Transport Policy Institute, - Litman, Steele - 2011 |

31 |
Do Changes in Neighborhood Characteristics Lead to Changes in Travel Behavior? A Structural Equations Modeling Approach.” Transportation (forthcoming).
- Cao, Mokhtarian, et al.
- 2007
(Show Context)
Citation Context ...e sum of the direct and all indirect effects, is simply the direct effect. This is captured by the coefficient a of BE in the equation for TB, and represents the true effect of BE on TB (however, see =-=Cao et al., 2007-=- for a discussion of the case in which c is insignificant). For nonrecursive models, e.g. if in Fig. 2, TB were modeled as affecting AT (consistent with Fig. 1b) as well as the converse, computation o... |

31 |
Econometric Analysis, third ed
- Greene
- 1997
(Show Context)
Citation Context ...ich is known to result in coefficient estimates for that variable that are inconsistent and biased toward zero, and coefficient estimates for the other variables in the equation that are also biased (=-=Greene, 1997-=-). (3) The asymptotic properties of IV estimators no longer hold, even for very large samples. The resulting coefficient estimators can be extremely biased and statistically inconsistent, i.e. differi... |

29 |
What if you live in the wrong neighborhood? The impact of residential neighborhood type dissonance on distance traveled. Transportation Research Part D
- Schwanen, Mokhtarian, et al.
(Show Context)
Citation Context ..., one incorporating attitudes directly (e.g. Kitamura et al., 1997), and the other incorporating an attitudinal-based measure of dissonance between one’s preferred and actual neighborhood types (e.g. =-=Schwanen and Mokhtarian, 2005-=-). In the first case, TB is simply modeled as a function of AT as well as BE: TB f3ðBE; AT; X Þþn; ð5Þ which removes AT from the e of Eqs. (3) and (4), and thereby presumably eliminates any correlat... |

28 | Choosing Between and Interpreting the Heckit and Two-Part Models for Corner Solutions, - Dow, Norton - 2003 |

28 | Trade-offs in Residential Location Decisions: Transportation versus Other Factors. Transport Policy and Decision Making 1(1 - Weisbrod, Lerman, et al. - 1980 |

26 |
Qualitative methods in travel behaviour research,” in Transport survey quality and innovation
- Clifton, Handy
(Show Context)
Citation Context ...velopment of survey instruments, the identification of appropriate model specifications and/or market segments having different decision-making processes, and the validation of multivariate analyses (=-=Clifton and Handy, 2003-=-; Pendyala, 1998). Nevertheless, used on its own it has several limitations. To begin with, the sample size is generally small and may not be representative of the population of interest. Moreover, di... |

26 | Modeling residential sorting effects to understand the impact of the built environment on commute mode choice - Pinjari, Pendyala, et al. - 2007 |

25 |
Application and interpretation of nested logit models of intercity mode choice
- Forinash, Koppelman
- 1993
(Show Context)
Citation Context ...y distinct behavioral processes, with different policy implications – to fit the data roughly equally well (MacCallum, 1995; although the same can be true for nested logit models as well – see, e.g., =-=Forinash and Koppelman, 1993-=-). Further, structural equations models are not well-suited to situations where one or more endogenous variables are multinomial, a case of considerable interest in travel behavior research (particula... |

22 |
Accessibility in a metropolis: Toward a better understanding of land use and travel
- Kitamura, Akiyama, et al.
(Show Context)
Citation Context ...e of transit mode. However, since most previous studies have employed multivariate analysis and accounted for the sorting effect of sociodemographic characteristics (e.g., Abreu e Silva et al., 2006; =-=Kitamura et al., 2001-=-), we focus this review on the issue of attitude-induced self-selection. Unless explicitly indicated, residential self-selection in the remainder of the paper refers only to that resulting from attitu... |

22 | The use of the life-style concept in travel demand models. Environment and Planning A 15: 623–638
- Salomon, Benakiva
- 1983
(Show Context)
Citation Context ...cessary, condition for nested logit to be a potentially appropriate model structure). The present context is one such natural application, since residential choice has long and widely been held (e.g. =-=Salomon and Ben-Akiva, 1983-=-) to be a longer-term choice which is antecedent to short-term choices related to individual trips. But it must be emphasized that finding such a structure to fit the data well cannot be taken as conf... |

21 |
Joint models of attitudes and behavior in evaluation of the San Diego I-15 congestion pricing project.
- Golob
- 2001
(Show Context)
Citation Context ...n). There is a sizable literature in transportation (and other fields) on the mutual causality between attitudes and behavior, with ample evidence for impacts in both directions (e.g., Tardiff, 1977; =-=Golob, 2001-=-). Thus, improving the realism of the model even further suggests the need for multiple interrelated equations, reflecting the multiple likely directions of causality. Specifically, one could postulat... |

21 | An exploration of the relationship between mode choice 22 and complexity of trip chaining patterns. - Ye, Pendyala, et al. - 2007 |

20 |
Tests for stability in linear structural equation systems
- Bentler, Freeman
- 1983
(Show Context)
Citation Context ...ty), it is quite difficult to isolate the true effect of BE on TB (that is, the separate BE effect remaining after the influence of AT is accounted for). The stability index of a nonrecursive system (=-=Bentler and Freeman, 1983-=-) gives a mathematical indication of whether the infinite loops of impacts converge or diverge, but gives no advice on how to isolate the true effect of BE from the portion of BE’s total effect that i... |

20 | Onmeasures of explained variance in nonrecursive structural equation models,”
- Raykov
- 2000
(Show Context)
Citation Context ...tuation again differs depending on whether the model in question is recursive or nonrecursive. For recursive models, computing equationby-equation R 2 s is appropriate and relatively straightforward (=-=Bentler and Raykov, 2000-=-), with interpretation (proportion of total variance in the ‘‘left-hand side’’ variable explained by the model) identical to that of the R 2 for a single-equation regression model (see, e.g., Mueller,... |

19 |
The Impact of Urban Form on Automobile Travel: Disentangling Causation from Correlation
- Vance, Hedel
- 2007
(Show Context)
Citation Context ...udes are explicitly controlled for. f Strong if attitudes are measured at each wave; otherwise, a DBE that precedes a DTB may itself be preceded and caused by a DAT. g However, a very recent article (=-=Vance and Hedel, 2007-=-) applies a related technique, the two-part model (2PM), to control for the endogeneity of urban form measures in a model of 210 P.L. Mokhtarian, X. Cao / Transportation Research Part B 42 (2008) 204–... |

18 |
Causal inferences involving transportation attitudes and behavior
- Tardiff
- 1976
(Show Context)
Citation Context ...ntial relocation). There is a sizable literature in transportation (and other fields) on the mutual causality between attitudes and behavior, with ample evidence for impacts in both directions (e.g., =-=Tardiff, 1977-=-; Golob, 2001). Thus, improving the realism of the model even further suggests the need for multiple interrelated equations, reflecting the multiple likely directions of causality. Specifically, one c... |

17 | The determinants of daily travel-activity patterns: Relative location and sociodemographic factors - Hanson - 1982 |

17 |
Cars and the City: An investigation of transportation and residential location choices in New
- Salon
- 2006
(Show Context)
Citation Context ...atistical significance of the BE after self-selection has been controlled for, and several comment that the BE is only one (type) of a number of variables influencing travel behavior, only one study (=-=Salon, 2006-=-) directly quantified the true contribution of the BE to the explained variation in travel behavior. This is despite the fact that it is relatively easy to assess the proportionate magnitude of the im... |

16 |
The impact of college selectivity on income for men and women.
- Mueller
- 1988
(Show Context)
Citation Context ...ov, 2000), with interpretation (proportion of total variance in the ‘‘left-hand side’’ variable explained by the model) identical to that of the R 2 for a single-equation regression model (see, e.g., =-=Mueller, 1988-=- for an application). Thus, it is reasonable to suggest that, for recursive models, the title question of this subsection can be answered in the same way as for the statistical controls method. With r... |

15 |
Microeconomics: Methods and Applications.
- Cameron, Trivedi
- 2005
(Show Context)
Citation Context ...hly correlated with the endogenous explanatory variable (BE) they are predicting (‘‘relevance’’), but not be significantly correlated with the error term (e) of the original equation (‘‘exogeneity’’; =-=Cameron and Trivedi, 2005-=-; Hall et al., 1996).The problem is that BE must be substantially correlated with e in order for endogeneity bias to be a problem; small correlations between observed and unobserved variables are tole... |

13 | Simple Estimators For Hard Problems: Endogeneity in Discrete Choice Related Models. Working Papers in Economics No
- Lewbel
- 2004
(Show Context)
Citation Context ...principle a combination of these two types of simultaneous models is possible – that is, a system of structural equations with discrete (binary or ordinal) endogenous variables (see, e.g., Lee, 1981; =-=Lewbel, 2004-=-; Louviere et al., 2005; Muthén, 1983) – applications involving such systems are still relatively rare and we are not aware of any in the present context. 3.5.1. Joint discrete choice models In joint ... |

12 |
Effects of land use characteristics on residence and employment location and travel behavior of urban adult workers. Transportation Research Record,
- SILVA, GOLOB, et al.
- 2006
(Show Context)
Citation Context ...uence on their choice of transit mode. However, since most previous studies have employed multivariate analysis and accounted for the sorting effect of sociodemographic characteristics (e.g., Abreu e =-=Silva et al., 2006-=-; Kitamura et al., 2001), we focus this review on the issue of attitude-induced self-selection. Unless explicitly indicated, residential self-selection in the remainder of the paper refers only to tha... |

10 | Modeling choices of residential location and mode of travel to work. Chapter 9 - Horowitz - 1986 |

10 | The application and economic interpretation of selectivity models
- CL, Raunikar, et al.
- 1991
(Show Context)
Citation Context ...st general case, to account for the appearance of BE in both the participation and the outcome equations. Respectively, the effects of interest have the following interpretations in this context (see =-=Huang et al., 1991-=- and Vance and Geoghegan, 2004 for applications in different contexts): • oETB S Š=oBE is the effect of increasing BE by one unit on the expected potential travel behavior of a randomly selected pers... |

10 | Approaches to Social Research (Fourth ed - Singleton, Straits - 1999 |

10 |
Strategies of path analysis
- Wolfle
(Show Context)
Citation Context ...correlations between the predetermined variables AT and sociodemographics themselves, and the (3) direct and (4) indirect effects of BE on TB (for an example, see the blockrecursive model analyzed by =-=Wolfle, 1980-=- and let his X 1 represent our TB, his X 3 be BE, and X 4 be AT; also see Alwin and Hauser, 1975). In Fig. 2 there are no indirect effects of BE on TB (i.e. those occurring through the impact of BE on... |

9 |
Causal Analysis in Travel Behaviour Research: A Cautionary Note. Travel Behaviour Research: Updating the State of Play. Pergamon
- Pendyala
- 1998
(Show Context)
Citation Context ...uments, the identification of appropriate model specifications and/or market segments having different decision-making processes, and the validation of multivariate analyses (Clifton and Handy, 2003; =-=Pendyala, 1998-=-). Nevertheless, used on its own it has several limitations. To begin with, the sample size is generally small and may not be representative of the population of interest. Moreover, direct questioning... |

9 |
Travel demand forecasting with the nested multinomial logit model
- Sobel
- 1980
(Show Context)
Citation Context ...OjRC] by PrWLjRCŠ P PrWL; RC; AOŠ), and (2) that calculated aPrWL; AOjRCŠ). It is important to realize that although such a model can represent a temporal sequence of choices, it need not do so (=-=Sobel, 1980-=-). Mathematically, the nested logit model simply represents a particular structure for the correlations of unobserved variables across sets of alternatives; the choices themselves theoretically could ... |

9 | Modeling the determinants of Semi-subsistent and Commercial Land uses in an Agricultural Frontier of Southern Mexico: A switching regression approach
- Vance, Geoghegan
(Show Context)
Citation Context ...ients will be inconsistent and inefficient. Doing so leads to a form of omitted variables bias, where the omitted variable is one that corrects for the sample selectivity (see the equations in, e.g., =-=Vance and Geoghegan, 2004-=-). 3 See Train (1986, Chapter 5) for an informative discussion of this model when the RC variable is multinomial rather than binary.214 P.L. Mokhtarian, X. Cao / Transportation Research Part B 42 (20... |

8 |
Blocked-error-R2: A conceptually improved definition of the proportion of explained variance in models containing loops or correlated residuals
- Hayduk
- 2006
(Show Context)
Citation Context ...les (including AT) and relationships, to the incremental change when AT and BE are added together. For nonrecursive models, not surprisingly, the situation is more complex. An important recent paper (=-=Hayduk, 2006-=-) offers a meaningful definition of R 2 for an endogenous variable in either a recursive or nonrecursive SEM, 6 but decomposing such an R 2 into components due to specific explanatory variables has no... |

8 | household travel, and household lifestyles - Krizek |

8 | Recent progress on endogeneity in choice modeling
- Louviere, Train, et al.
- 2005
(Show Context)
Citation Context ...mbination of these two types of simultaneous models is possible – that is, a system of structural equations with discrete (binary or ordinal) endogenous variables (see, e.g., Lee, 1981; Lewbel, 2004; =-=Louviere et al., 2005-=-; Muthén, 1983) – applications involving such systems are still relatively rare and we are not aware of any in the present context. 3.5.1. Joint discrete choice models In joint discrete choice models,... |

7 |
Some Remarks on Instrumental Variables
- Imbens
- 2001
(Show Context)
Citation Context ...hanges from z0 to z1. That is, loosely speaking, what is the effect for those who needed an incentive to move from a suburban neighborhood to an urban one? This is the LATE (Imbens and Angrist, 1994; =-=Imbens, 2001-=-); it focuses on the effectiveness of the incentive by eliminating those who would have moved anyway (but see Winship and Morgan, 1999 p. 685 for some problems of the LATE approach). In general, we ex... |

6 | Departure-time choice and mode choice for nonwork trips: Alternative formulations of joint model systems - Tringides, Ye, et al. - 2004 |

2 |
The RESIDE (RESIDential Environments) project. research/hpreg/hpreg/programs/parg/reside> (accessed
- Giles-Corti
- 2006
(Show Context)
Citation Context ...and is only quasi-longitudinal in that ‘‘prior’’ measures are obtained only retrospectively, and do not include attitudes – though current attitudes are measured), a project is underway in Australia (=-=Giles-Corti, 2006-=-) which aims to do exactly that. Unfortunately, given the various limitations discussed throughout this paper, we are unable at this point to confidently specify the nature and extent of the causality... |