Influence measures based on confidence ellipsoids in general linear regression model with correlated regressors

In this paper, we investigated the Andrews–Pregibon (AP), COVRATIO and Cook–Weisberg (CW) statistics to determine the influential observations on the confidence ellipsoids in linear regression model with correlated errors and correlated regressors. A real example and a Monte Carlo simulation study are given to detect the effects of autocorrelation coefficient and ridge parameter on the AP, COVRATIO and CW statistics.     

Raising the legal gambling age in Finland: problem gambling prevalence rates in different age groups among past-year gamblers pre- and post-implementation

This article reports on the frequency of problem gambling, measured with the Lie/Bet instrument, in different age groups among Finnish past-year gamblers in 2011 (n = 2984) and 2014 (n = 2326). The data highlights the situation before, and three years after, the implementation of a raised minimum age limit for gambling from 15 to 18 years. The difference in problem gambling frequency when comparing all age groups was statistically significant in 2011, but not in 2014. A significantly lower frequency of problem gambling was found among 18–19-year-olds in 2014 (3.4%), compared to 2011 (16.3%). The results regarding problem gambling prevalence among 15–17-year-olds (8.0% in 2011, 0.0% in 2014) are somewhat inconclusive as the number of respondents fulfilling the criteria for problem gambling was zero in 2014, thus affecting the analysis. No statistically significant difference in problem gambling frequency was found among 20–21-year-olds (a group less affected by the policy implementation) – or other older age groups – between the survey years. While the findings should be viewed with caution, they do support recommendations regarding a minimum gambling age of 18 years or higher as an effective harm-minimization measure.     

An Extension of Sampford's Method for Unequal Probability Sampling

Abstract. Sampford's unequal probability sampling method is extended to the case that the inclusion probabilities do not sum to an integer. In this case, the sampling outcome is left open for exactly one randomly chosen unit and that unit gets a new inclusion probability. Three applications are presented. Two of them challenge traditional sampling routines. The simple Pareto sampling design, which was introduced by Rosén in 1997, is also extended. The extended Pareto design is shown to be close to the extended Sampford design.     

Immigrants and the public sector budget – accounting exercises for Sweden

Net contributions to the public sector budget in Sweden are investigated using large samples of foreign born and native born. The accounts build on various assumptions including that expenditures on public consumption are allocated according to the age of the person. The results indicate that during the period 1983 to 1992 net contributions of immigrants deteriorated. Upon arrival to Sweden, immigrants on average place a burden on the public sector budget but after a few years this no longer applies. Refugees initially put a larger burden on the public sector budget than other immigrants, but such a difference declines with years since immigration. Received: 11 March 1998/Accepted: 31 May 2000     

The influence function of penalized regression estimators

To perform regression analysis in high dimensions, lasso or ridge estimation are a common choice. However, it has been shown that these methods are not robust to outliers. Therefore, alternatives as penalized M-estimation or the sparse least trimmed squares (LTS) estimator have been proposed. The robustness of these regression methods can be measured with the influence function. It quantifies the effect of infinitesimal perturbations in the data. Furthermore, it can be used to compute the asymptotic variance and the mean-squared error (MSE). In this paper we compute the influence function, the asymptotic variance and the MSE for penalized M-estimators and the sparse LTS estimator. The asymptotic biasedness of the estimators make the calculations non-standard. We show that only M-estimators with a loss function with a bounded derivative are robust against regression outliers. In particular, the lasso has an unbounded influence function.     

SOCIAL DRINKING VERSUS ADMINISTERING ALCOHOL

Alcohol consumption tends to make some people (unwillingly) tell the truth, hence social drinking can serve as a signal in social contact games. We provide empirical evidence which shows that social drinking can serve as a trust facilitating mechanism. Liver cirrhosis mortality and the rate of abstainers are used to construct a novel index of moderate alcohol consumption, which correlates with trust levels in a cross‐country analysis.(JEL L14, D82, D71)