Hana Vonkova

How Sensitive are Retirement Decisions to Financial Incentives: A Stated Preference Analysis

We study effects of financial incentives on the retirement age using stated preference data. Dutch survey respondents were given hypothetical retirement scenarios describing age(s) of (partial and full) retirement and replacement rate(s). A structural model is estimated in which utility is the discounted sum of within period utilities that depend on employment status and income. Parameters of the utility function vary with observed and unobserved characteristics. Simulations show that the income and substitution effects of pensions as a function of the retirement age are substantial and larger than according to studies using data on actual retirement decisions.

Anchoring Vignettes and Response Consistency

The use of anchoring vignettes to correct for differential item functioning rests upon two identifying assumptions: vignette equivalence and response consistency. To test the second assumption the authors conduct an experiment in which respondents in an Internet panel are asked to both describe their health in a number of domains and rate their health in these domains. In a subsequent interview respondents are shown vignettes that are in fact descriptions of their own health. Under response consistency and some auxiliary assumptions with regard to the validity of the experiment, there should be no systematic differences between the evaluation of these vignettes in the second interview and the self-evaluations in the first interview. They analyze data for five health domains: sleep, mobility, concentration, breathing and affect. Although descriptively the vignettes and the self-evaluations are similar for a number of domains, their nonparametric analysis suggests that response consistency is satisfied for the domain of sleep, while it indicates rejection of either the auxiliary assumptions or response consistency for the other domains of health. Parametric analysis suggests that the auxiliary assumptions may be most problematic. The analysis points at the need for a systematic experimental approach to the design of anchoring vignettes before using them in practice.

Cannabis use in children with individualized risk profiles: Predicting the effect of universal prevention intervention

AIM To study the effect of a universal prevention intervention targeting cannabis use in individual children with different risk profiles. METHODS A school-based randomized controlled prevention trial was conducted over a period of 33 months (n=1874 sixth-graders, baseline mean age 11.82). We used a two-level random intercept logistic model for panel data to predict the probabilities of cannabis use for each child. Specifically, we used eight risk/protective factors to characterize each child and then predicted two probabilities of cannabis use for each child if the child had the intervention or not. Using the two probabilities, we calculated the absolute and relative effect of the intervention for each child. According to the two probabilities, we also divided the sample into a low-risk group (the quarter of the children with the lowest probabilities), a moderate-risk group, and a high-risk group (the quarter of the children with the highest probabilities) and showed the average effect of the intervention on these groups. RESULTS The differences between the intervention group and the control group were statistically significant in each risk group. The average predicted probabilities of cannabis use for a child from the low-risk group were 4.3% if the child had the intervention and 6.53% if no intervention was provided. The corresponding probabilities for a child from the moderate-risk group were 10.91% and 15.34% and for a child from the high-risk group 25.51% and 32.61%. School grades, thoughts of hurting oneself, and breaking the rules were the three most important factors distinguishing high-risk and low-risk children. CONCLUSIONS We predicted the effect of the intervention on individual children, characterized by their risk/protective factors. The predicted absolute effect and relative effect of any intervention for any selected risk/protective profile of a given child may be utilized in both prevention practice and research.

Testing Parametric Models Using Anchoring Vignettes Against Nonparametric Alternatives

Comparing assessments of health, job satisfaction, etc. on a subjective scale across countries or socio-economic groups is often hampered by differences in response scales across groups. Anchoring vignettes help to correct for such differences, either in parametric models (CHOPIT and extensions) or nonparametrically, comparing rankings of vignette ratings and self-assessments across groups. We construct specification tests of parametric models, comparing non-parametric rankings with rankings using the parametric estimates. Applied to six domains of health, the test always rejects standard CHOPIT, but an extended CHOPIT performs better. This implies a need for more flexible (parametric or semi-parametric) models than standard CHOPIT.

How Students Report Dishonest Behavior in School: Self-Assessment and Anchoring Vignettes

ABSTRACT The authors have studied heterogeneity in reporting behavior and its impact on the analysis of self-reports about students’ dishonest behavior in schools. Two hundred sixty-five randomly chosen, seventh-grade students (typically 12 years old) from lower secondary schools in Prague 6, a district in the capital of the Czech Republic, participated in this survey. The results of the self-reports, adjusted for heterogeneity, are highly related to students’ levels of academic achievement and their parents’ education and partly related to their gender, while unadjusted self-reports are only slightly related to the level of parents’ education. The authors also show differences in the reporting behavior across diverse subdomains of school behavior and suggest using anchoring vignettes closely related to the domain described in the self-reports.

Enhancing the Cross-Cultural Comparability of Self-Reports Using the Overclaiming Technique: An Analysis of Accuracy and Exaggeration in 64 Cultures

The overclaiming technique (OCT) is a novel way of measuring how socially desirable responding influences survey responses. It has the potential to enhance the cross-cultural comparability of survey data. It allows the identification of respondents’ knowledge accuracy and exaggeration by comparing their assessments of familiarity with existing and nonexisting concepts in a particular field of knowledge. We aim to compare the response patterns of countries and world regions based on their OCT accuracy and exaggeration-index values and validate these OCT scores using external variables. We also introduce a general model for the categorization of respondents based on their OCT indices values. We use the Programme for International Student Assessment (PISA) 2012 data from 64 countries (N = 275,904). We found considerable differences in response patterns across world regions: high accuracy values in East Asia, low accuracy values in Southern and Central America, high exaggeration values in Southern Europe, and low exaggeration values in Western Europe. Furthermore, we show that familiarity with math concepts changes substantially after adjustment using the OCT. The correlation between unadjusted math familiarity and math test scores is weak and nonsignificant (.13) whereas after adjustment the correlation becomes strong and significant (.66). Concerning other indicators such as gross domestic product (GDP), public expenditure in education, and Corruption Perceptions Index (CPI), before adjustment the correlation is negative and even significant for CPI (–.11, –.22, and –.45) whereas after adjustment the correlation becomes, though nonsignificant, positive (.05, .19, and .07). We also discuss the OCT in the context of other methods indicative of culturally preferred scale usage.