An algorithmic approach to construct D-optimal saturated designs for logistic model

In this paper, locally D-optimal saturated designs for a logistic model with one and two continuous input variables have been constructed by modifying the famous Fedorov exchange algorithm. A saturated design not only ensures the minimum number of runs in the design but also simplifies the row exchange computation. The basic idea is to exchange a design point with a point from the design space. The algorithm performs the best row exchange between design points and points form a candidate set representing the design space. Naturally, the resultant designs depend on the candidate set. For gain in precision, intuitively a candidate set with a larger number of points and the low discrepancy is desirable, but it increases the computational cost. Apart from the modification in row exchange computation, we propose implementing the algorithm in two stages. Initially, construct a design with a candidate set of affordable size and then later generate a new candidate set around the points of design searched in the former stage. In order to validate the optimality of constructed designs, we have used the general equivalence theorem. Algorithms for the construction of optimal designs have been implemented by developing suitable codes in R.     

Parameter estimation in semi-linear models using a maximal invariant likelihood function

In this paper, we consider the problem of estimation of semi-linear regression models. Using invariance arguments, Bhowmik and King [2007. Maximal invariant likelihood based testing of semi-linear models. Statist. Papers 48, 357–383] derived the probability density function of the maximal invariant statistic for the non-linear component of these models. Using this density function as a likelihood function allows us to estimate these models in a two-step process. First the non-linear component parameters are estimated by maximising the maximal invariant likelihood function. Then the non-linear component, with the parameter values replaced by estimates, is treated as a regressor and ordinary least squares is used to estimate the remaining parameters. We report the results of a simulation study conducted to compare the accuracy of this approach with full maximum likelihood and maximum profile-marginal likelihood estimation. We find maximising the maximal invariant likelihood function typically results in less biased and lower variance estimates than those from full maximum likelihood.     

Minimally changed run sequences in factorial experiments

Randomization of run sequences in factorial experiments may result in large number of changes in factor levels which will make the experimentation expensive, time-consuming and difficult. Experiments in which it is difficult to change the levels of factor(s) use of minimally changed run sequences may often be preferable to a random run sequence. In the present paper, we have developed method for obtaining minimally changed run sequences for factorial experiments. The general expression of factor-wise number of level changes for the developed minimally changed run sequences has also been obtained. A relationship has been established between the time count effect of a lower order factorial with minimally changed run sequences and that of a higher order factorial with minimally changed run sequences obtained through the lower order minimally changed run sequences. For providing a readymade solution to the end users, a SAS macro has also been developed for generating these minimally changed run sequences along with its parameters.     

Optimal Block Designs With Interference Effects From Neighboring Units Under a Non Additive Model

Here, the optimality of block design with interference effect from neighboring unit under a general non additive model is investigated, which allows for the presence of interactions among the treatments applied in the adjacent plots. A non additive model with interference × direct effects of treatments is considered as these effects contribute significantly to the response. A class of complete block designs balanced for interference effects from left neighboring unit is shown to be universally optimal for the estimation of direct and interference effects of treatments and two such series of designs have been constructed. Furthermore, considering direct treatment × block non additivity with interference effects, the optimality is studied and the optimal designs are obtained.     

Optimal block designs with non additive mixed effects interference

This paper deals with optimality aspects of block designs balanced for interference effects from neighboring units on both sides under a general non additive model along with random block effects. Here, a class of complete, circular block designs strongly balanced for interference effects has been shown to be universally optimal for the estimation of direct effects and interference effects (left and right) of treatments under a non additive mixed effects model.     

A note on circular neighbor balanced designs

ABSTRACT

This paper describes some methods of constructing circular neighbor balanced and circular partially neighbor balanced block designs for estimation of direct and neighbor effects of the treatments. A class of circular neighbor balanced block designs with unequal block sizes is also proposed.     

Maximal invariant likelihood based testing of semi-linear models

In this paper, we use a maximal invariant likelihood (MIL) to construct two likelihood ratio (LR) tests in the context of a semi-linear regression model. The first involves testing for the inclusion of a non-linear regressor and the second involves testing a linear regressor against the alternative of a non-linear regressor. We report the results of a Monte Carlo experiment that compares the size and power properties of the traditional LR tests with those of our proposed MIL based LR tests. Our simulation results show that in both cases, the MIL based tests have more accurate asymptotic critical values and better behaved (i.e., better centred) power curves than their classical counterparts.     

The Impact of Training Videos on Attitudes About Parenting After Divorce

Many U.S. states require divorcing parents to take education classes about the impact of divorce on parents and children. As educators, psychiatrists, social workers, and others create these classes, it is important to evaluate what elements of the curriculum are effective in achieving targeted outcomes. Successful Coparenting After Divorce (https://coparenting.fsu.edu) is a free online divorce education course that focuses on issues such as the emotional impact of divorce, conflict reduction, and skills for parental cooperation. The course also includes videos covering topics including examples of negative and positive parental behavior, and testimonials from children. Pilot testing of the videos and the overall course impact with divorcing parents (n = 218) suggested that the videos’ utility were significantly related to the positive change in parents’ perceptions of their relationship with their former spouse, and their child-focused attitudes. Implications for practitioners who design or provide divorce education to parents are discussed.     

Education for all – but not Hong Kong’s ethnic minority students

School participation rates in Hong Kong are acknowledged to be high but not always for the city’s ethnic minority students. Case study was used to portray the ‘out of school’ phenomenon for ethnic minority students at a secondary school in Hong Kong through the experiences of one such ethnic minority student. Morshed, a Pakistani student, participated with the researchers in an unstructured in-depth interview. In addition, interviews were also conducted with the principal, two teachers, and one support staff from the school. The study showed the complex interaction of school contexts and personal life issues that influenced Morshed’s dropping out of school. Implications are drawn for providing better school support that can facilitate more positive schooling experiences for Hong Kong’s ethnic minority students.     

Experimental designs for open pollination in polycross trials

A polycross is the pollination by natural hybridization of a group of genotypes, generally selected, grown in isolation from other compatible genotypes in such a way to promote random open pollination. A particular practical application of the polycross method occurs in the production of a synthetic variety resulting from cross-pollinated plants. Laying out these experiments in appropriate designs, known as polycross designs, would not only save experimental resources but also gather more information from the experiment. Different situations may arise in polycross nurseries where accordingly different polycross designs may be used. For situations in which some genotypes interfere in the growth or production of other genotypes, but have to be grown together, neighbour-restricted design is a better option. Furthermore, when the topography of the nursery is such that a known wind system in a certain direction may prevail, then designs balanced for neighbour effects of genotypes only in the direction of wind are appropriate which may help in saving experimental resources to a great extent. Also, when genotypes are planted in a small area without leaving much space between rows, designs balanced for neighbour effects from all possible eight directions are useful to have equal chance of pollinating and being pollinated by every other genotype. Here, polycross designs have been obtained to match above-mentioned three situations. SAS Macros have also been developed to generate these proposed designs.