1,721,011 research outputs found
Path analysis for binary random variables
No full textThe decomposition of the overall effect of a treatment into direct and indirect effects is here investigated with reference to a recursive system of binary random variables. We show how, for the single mediator context, the marginal effect measured on the log odds scale can be written as the sum of the indirect and direct effects plus a residual term that vanishes under some specific conditions. We then extend our definitions to situations involving multiple mediators and address research questions concerning the decomposition of the total effect when some mediators on the pathway from the treatment to the outcome are marginalized over. Connections to the counterfactual definitions of the effects are also made. Data coming from an encouragement design on students’ attitude to visit museums in Florence, Italy, are reanalyzed. The estimates of the defined quantities are reported together with their standard errors to compute p values and form confidence intervals
On the estimation of a binary response model in a selected population
No full textA generalization of the Probit model is presented, with the extended skew-normal
cumulative distribution as a link function, which can be used for modelling a binary
response variable in the presence of selectivity bias. The estimate of the parameters via
ML is addressed, and inference on the parameters expressing the degree of selection is
discussed. The assumption underlying the model is that the selection mechanism
influences the unmeasured factors and does not affect the explanatory variables. When
this assumption is violated, but other conditional independencies hold, then the model
proposed here is derived. In particular, the instrumental variable formula still applies
and the model results at the second stage of the estimating procedure
On the estimation of a binary response model in a selected population
No full textNell’articolo si propone una generalizzazione del modello probit attraverso l’applicazione della funzione di ripartizione della variabile casuale Extended Skew Normal come funzione link. Tale modello si può utilizzare per variabili di risposta binarie quando si è in presenza di un meccanismo di distorsione selettiva. I parametri del modello vengono stimati attraverso il metodo della massima verosimiglianza, e si considera l’inferenza sui parametri del livello di selezione. Il modello viene analizzato nel caso in cui il meccanismo di selezione influenza i fattori non osservati ma non ha alcun effetto sulle variabili esplicative tuttavia continuano a valere altre condizioni di indipendenza. In particolare si dimostra che la formula della variabile strumentale continua a valere e il modello risulta al secondo stadio della procedura di stimaA generalization of the probit model is presented, with the extend skew normal distribution as a link function, which can be used for modeling a binary response variable in the presence of selectivity bias. The estimate of the parameters via ML is addressed, and inference on the parameters expressing the degree of selection is discussed. The assumption underlying the model is that the selection mechanism influences the unmeasured factors and does not affect the explanatory variables. When this assumption is violated, but other conditional independecies hold, then the model proposed here is derived. In particular, the instrumental variable formula still applies and the model results at the second stage of the estimating procedure
A multiple-record systems estimation method that takes observed and unobserved heterogeneity into account
No full textWe present a model to estimate the size of an unknown population from a number of lists that applies when the assumptions of (a) homogeneity of capture probabilities of individuals and (b) marginal independence of lists are violated. This situation typically occurs in epidemiological studies, where the heterogeneity of individuals is severe and researchers cannot control the independence between sources of ascertainment. We discuss the situation when categorical covariates are available and the interest is not only in the total undercount, but also in the undercount within each stratum resulting from the cross-classification of the covariates. We also present several techniques for determining confidence intervals of the undercount within each stratum using the profile log likelihood, thereby extending the work of Cormack (1992, Biometrics48, 567–576)
Role of antioxidant in the formulation of sun filter
No full textAntioxidant substances can be used in the prevention and treatment of cutaneous pathologies, such as photoageing damage and some tumours. Topical
use is considered most appropriate since antioxidants concentrate first in the horny layer, a structure very exposed to oxidative stress. The antioxidant
activity of most products is due to the association of several principles. In this study, we tested a mixture of three antioxidant substances: tocopherol acetate,
ascorbic acid and lycopene (extract of Lycopersicon esculentum). The mixture was incorporated in a standard low-protection sun filter and was evaluated
‘in vivo’ by means of a well-tested experimental device. Different quantities of the product were applied to skin areas which were then irradiated with
a sun simulator. The magnitude of the reaction was evaluated with a three-stimulus chromometer. The experimental product provided greater protection
than the standard control product, confirming the results of our previous investigations and proving that antioxidants increase the protective capacity of
sun filters
Modelling censored data with the skew-normal distribution
No full textSulla base di un set di dati reali, relativo ad un test clinico per confrontare quattro tipi di trattamento per fratture ossee, si propone un'estensione del modello normale per dati troncati, noto anche come modello di Tobit, al fine di poter tener conto dell'asimmetria della distribuzione dei residui. Il modello proposto utilizza la famiglia delle distribuzioni normali asimmetriche che include il modello normale come caso particolare. Nel lavoro vengono analizzati gli aspetti teorici del modello attraverso l'analisi delle funzioni score e la formulazione di una midura basata sull'R quadro per valutare la capacità del modello di descrivere i dati.Motivated by a real data example, coming from a clinical trial to
compare four treatments on severely sprained ankles, we extend the normal model
for censored data, also known as Tobit model, to accomodate asymmetry in the
distribution of the residuals. The proposed model makes use of the skew-normal
distribution, a distribution that includes the normal one. The theoretical features
of the model are investigated. These involve the analysis of the behaviour of the
score functions and the formulation of a R-squared type of measure to evaluate
the capacity of the model to represent the data
- …
