The inclusion of open-ended questions on quantitative surveys of children: Dealing with unanticipated responses relating to child abuse and neglect

Web surveys have been shown to be a viable, and relatively inexpensive, method of data collection with children. For this reason, the Kids’ Life and Times (KLT) was developed as an annual online survey of 10 and 11 year old children. Each year, approximately 4,000 children participate in the survey. Throughout the six years that KLT has been running, a range of questions has been asked that are both policy-relevant and important to the lives of children. Given the method employed by the survey, no extremely sensitive questions that might cause the children distress are included. The majority of questions on KLT are closed yielding quantitative data that are analysed statistically; however, one regular open-ended question is included at the end of KLT each year so that the children can suggest questions that they think should be asked on the survey the following year. While most of the responses are innocuous, each year a small minority of children suggest questions on child abuse and neglect. This paper reports the responses to this question and reflects on how researchers can, and should, deal with this issue from both a methodological and an ethical perspective.

Predicting forest attributes In southeast Alaska using artificial neural networks

Artificial neural network (ANN) methods are used to predict forest characteristics. The data source is the Southeast Alaska (SEAK) Grid Inventory, a ground survey compiled by the USDA Forest Service at several thousand sites. The main objective of this article is to predict characteristics at unsurveyed locations between grid sites. A secondary objective is to evaluate the relative performance of different ANNs. Data from the grid sites are used to train six ANNs: multilayer perceptron, fuzzy ARTMAP, probabilistic, generalized regression, radial basis function, and learning vector quantization. A classification and regression tree method is used for comparison. Topographic variables are used to construct models: latitude and longitude coordinates, elevation, slope, and aspect. The models classify three forest characteristics: crown closure, species land cover, and tree size/structure. Models are constructed using n-fold cross-validation. Predictive accuracy is calculated using a method that accounts for the influence of misclassification as well as measuring correct classifications. The probabilistic and generalized regression networks are found to be the most accurate. The predictions of the ANN models are compared with a classification of the Tongass national forest in southeast Alaska based on the interpretation of satellite imagery and are found to be of similar accuracy.

Machine learning for transient discovery in Pan-STARRS1 difference imaging

Efficient identification and follow-up of astronomical transients is hindered by the need for humans to manually select promising candidates from data streams that contain many false positives. These artefacts arise in the difference images that are produced by most major ground-based time-domain surveys with large format CCD cameras. This dependence on humans to reject bogus detections is unsustainable for next generation all-sky surveys and significant effort is now being invested to solve the problem computationally. In this paper, we explore a simple machine learning approach to real-bogus classification by constructing a training set from the image data of similar to 32 000 real astrophysical transients and bogus detections from the Pan-STARRS1 Medium Deep Survey. We derive our feature representation from the pixel intensity values of a 20 x 20 pixel stamp around the centre of the candidates. This differs from previous work in that it works directly on the pixels rather than catalogued domain knowledge for feature design or selection. Three machine learning algorithms are trained (artificial neural networks, support vector machines and random forests) and their performances are tested on a held-out subset of 25 per cent of the training data. We find the best results from the random forest classifier and demonstrate that by accepting a false positive rate of 1 per cent, the classifier initially suggests a missed detection rate of around 10 per cent. However, we also find that a combination of bright star variability, nuclear transients and uncertainty in human labelling means that our best estimate of the missed detection rate is approximately 6 per cent.

Enchytraeid worms retard polycyclic aromatic hydrocarbon degradation in a coniferous forest soil

In this study the fate of naphthalene, fluorene and pyrene were investigated in the presence and absence of enchytraeid worms. Microcosms were used, which enabled the full fate of ¹⁴C-labelled PAHs to be followed. Between 60 and 70% of naphthalene was either mineralised or volatilised, whereas over 90% of the fluorene and pyrene was retained within the soil. Mineralisation and volatilisation of naphthalene was lower in the presence of enchytraeid worms. The hypothesis that microbial mineralisation of naphthalene was limited by enchytraeids because they reduce nutrient availability, and hence limit microbial carbon turnover in these nutrient poor soils, was tested. Ammonia concentrations increased and phosphorus concentrations decreased in all microcosms over the 56 d experimental period. The soil nutrient chemistry was only altered slightly by enchytraeid worms, and did not appear to be the cause of retardation of naphthalene mineralisation. The results suggest that microbial availability and volatilisation of naphthalene is altered as it passes through enchytraeid worms due to organic material encapsulation. © 2004 Elsevier Ltd. All rights reserved.