|
on Discrete Choice Models |
| By: | Katja Möhring (Universität zu Köln); Alexander Schmidt (Universität zu Köln) |
| Abstract: | The Stata package “multilevel tools” (mlt) includes a range of ado-files for postestimation after multilevel models (xtmixed/xtmelogit). Up to now, it contains three commands (more ado-files will be added in the future): mltrsq gives the Boskers/Snijders R-square and the Bryk/Raudenbusch R-square values. mltcooksd gives the influence measures Cook’s D and DFBETAs for the higher-level units in hierarchical mixed models. mltshowm presents how the model looks if those cases detected as influential are excluded from the sample. In our presentation, we will discuss the issue of influential cases in multilevel modeling. We will use some research examples to stress the importance of considering influential cases, particularly in multilevel analysis. We will show how the influence measures for second-level units are defined and how we calculate them. |
| Date: | 2012–06–04 |
| URL: | http://d.repec.org/n?u=RePEc:boc:dsug12:06&r=dcm |
| By: | Dirk Enzmann (University of Hamburg); Ulrich Kohler (WZB Berlin) |
| Abstract: | Because of the scaling of the unobserved latent dependent variable in logistic and probit multilevel models, the lowest level residual variance is always pi^2/3 (logistic regression) or 1.0 (probit regression). As a consequence, a change of regression coefficients and variance components between hierarchically nested models cannot be interpreted unambiguously. To overcome this issue, rescaling of the unobserved latent dependent variable of nested models to the scale of the intercept-only model has been proposed (Hox 2010). In this talk, we demonstrate the use of the program meresc, which implements this procedure to rescale the results of mixed nonlinear probability models such as xtmelogit, xtlogit, or xtprobit. |
| Date: | 2012–06–04 |
| URL: | http://d.repec.org/n?u=RePEc:boc:dsug12:04&r=dcm |
| By: | Richard Norman (CHERE, University of Technology, Sydney); Paula Cronin (CHERE, University of Technology, Sydney); Rosalie Viney (CHERE, University of Technology, Sydney) |
| Abstract: | Purpose: To estimate an Australian algorithm for the newly developed 5-level version of the EQ-5D, for use in the economic evaluation of health and healthcare interventions. Methods: A discrete choice experiment (DCE) was run in an online Australian-representative sample. A random-effects probit model was estimated, and converted to a zero to one scale for use in economic evaluation. Results: At least one choice set was completed by 944 respondents, of which 932 completed all ten choice sets. The mean and median completion times were 17.9 and 9.4 minutes respectively, demonstrating a highly skewed pattern. Respondents were slightly younger and better-educated than the general Australian population. The regression results broadly reflect the monotonic nature of the EQ-5D-5L. Utility increases in life expectancy, and decreases in higher levels in each dimension of the instrument. A high proportion of respondents found the task clear and relatively easy to complete. Conclusions: DCEs are a valuable approach in the estimation of utility weights for multi-attribute utility instruments such as the EQ-5D-5L. |
| Keywords: | Discrete choice experiments, EQ-5D, multi-attribute utility measurement |
| JEL: | I1 I19 |
| Date: | 2012–01 |
| URL: | http://d.repec.org/n?u=RePEc:her:chewps:2012/01&r=dcm |
| By: | M. T. Aparicio (University of Zaragoza); I. Villanúa (University of Zaragoza) |
| Abstract: | The binary model selection procedures are the purpose of this paper. It has been very often studied in the case of linear and nested competing models, but not too much in the framework of non linear and non nested models. Using the classification of Vuong (1989) in nested, overlapping and strictly non-nested models, we focus on the overlapping models. The special situation that the competing models don’t include the true model is studied together with the usual case of the true model included in the compared models. We carry out the analysis both theoretically and through a Monte Carlo experiment. |
| Keywords: | logit and probit, selection criteria, overlapping models, true model, convergence, asymptotic behaviour. |
| JEL: | C15 C25 C44 |
| Date: | 2012–01 |
| URL: | http://d.repec.org/n?u=RePEc:zar:wpaper:dt2012-01&r=dcm |