Search results

We extend Vuong’s (1989) model-selection statistic to allow for complex survey samples. As a further extension, we use an M-estimation setting so that the tests apply to general estimation problems – such as linear and nonlinear least squares, Poisson regression and fractional response models, to name just a few – and not only to maximum likelihood settings. With stratified sampling, we show how the difference in objective functions should be weighted in order to obtain a suitable test statistic. Interestingly, the weights are needed in computing the model-selection statistic even in cases where stratification is appropriately exogenous, in which case the usual unweighted estimators for the parameters are consistent. With cluster samples and panel data, we show how to combine the weighted objective function with a cluster-robust variance estimator in order to expand the scope of the model-selection tests. A small simulation study shows that the weighted test is promising.

This paper uses a sample of school age children from the Nepal Demographic Health Survey (NDHS) to examine the relationship between maternal education and child schooling in Nepal. Taking advantage of the two-stage stratified sample design, we estimate a sample selection model controlling for cluster fixed effects. These results are then compared to OLS and Tobit models. Our analysis shows that being male significantly increases the likelihood of attending school and for those children attending school, it also affects the years of schooling. Parental education has a similarly positive effect on child school, but interestingly we find maternal education having a relatively greater effect on the schooling of girls. Our results also point to household wealth as having a positive effect on both the probability of schooling and the years of schooling in all our models, with the magnitude of these effects being similar for male and female children. Finally, a comparison of our results with a model ignoring cluster fixed effects produces results that are statistically different both in signs and in the levels of significance.

Sampling units for the 2013 Methods-of-Payment survey were selected through an approximate stratified two-stage sampling design. To compensate for nonresponse and noncoverage and ensure consistency with external population counts, the observations are weighted through a raking procedure. We apply bootstrap resampling methods to estimate the variance, allowing for randomness from both the sampling design and raking procedure. We find that the variance is smaller when estimated through the bootstrap resampling method than through the naive linearization method, where the latter does not take into account the correlation between the variables used for weighting and the outcome variable of interest.

Purpose — This study proposes an optimal survey design method for multi-day and multi-period panels that maximizes the statistical power of the parameter of interest under the conditions that non-linear changes in response to a policy intervention over time can be expected.

Design/methodology/approach — The proposed method addresses balances among sample size, survey duration for each wave and frequency of observation. Higher-order polynomial changes in the parameter are also addressed, allowing us to calculate optimal sampling designs for non-linear changes in response to a given policy intervention.

Findings — One of the most important findings is that variation structure in the behaviour of interest strongly influences how surveys are designed to maximize statistical power, while the type of policy to be evaluated does not influence it so much. Empirical results done by using German Mobility Panel data indicate that not only are more data collection waves needed, but longer multi-day periods of behavioural observations per wave are needed as well, with the increase in the non-linearity of the changes in response to a policy intervention.

Originality/value — This study extends previous studies on sampling designs for travel diary survey by dealing with statistical relations between sample size, survey duration for each wave, and frequency of observation, and provides the numerical and empirical results to show how the proposed method works.

Large-scale continuous mobility surveys have some advantages over less frequent (usually every 10 years), even larger-scale cross-sectional surveys; these advantages have been well documented in previous papers (Ampt & Ortúzar, 2004).

In this paper we first define what we mean by ‘ongoing mobility surveys’. We then describe the state of practice in this context, briefly reviewing the state of affairs in all the cases that we are aware of. We then discuss some problems encountered in practice and offer ideas for improvement. In particular, we discuss a wide range of issues that are likely to act as barriers to a high quality and sustainable implementation and suggest approaches for improvement. Issues covered include sampling frames and sampling methods, survey methods, respondent burden, weighting processes and expansion, and the increased importance of developing and maintaining field staff motivation. We also touch briefly on the practical/political issue of securing ongoing funding. Throughout, we advance some thoughts to try and explain why this method has not gained wider acceptance, particularly in the Northern Hemisphere where there are more examples of travel surveys in general.

The paper also raises some ideas and issues about the way in which ongoing mobility surveys can best collect data for the environmental accounting of travel. Finally, we raise questions about the environmental impact of the survey methods themselves as a stimulus for further consideration.