Looking for Validity or Testing It? The Perils of Stepwise Regression, Extreme-Scores Analysis, Heteroscedasticity, and Measurement Error

When researchers introduce a new test they have to demonstrate that it is valid, using unbiased designs and suitable statistical procedures. In this article we use Monte Carlo analyses to highlight how incorrect statistical procedures (i.e., stepwise regression, extreme scores analyses) or ignoring regression assumptions (e.g., heteroscedasticity) contribute to wrong validity estimates. Beyond these demonstrations, and as an example, we re-examined the results reported by Warwick, Nettelbeck, and Ward (2010) concerning the validity of the Ability Emotional Intelligence Measure (AEIM). Warwick et al. used the wrong statistical procedures to conclude that the AEIM was incrementally valid beyond intelligence and personality traits in predicting various outcomes. In our re-analysis, we found that the reliability-corrected multiple correlation of their measures with personality and intelligence was up to .69. Using robust statistical procedures and appropriate controls, we also found that the AEIM did not predict incremental variance in GPA, stress, loneliness, or well-being, demonstrating the importance for testing validity instead of looking for it.

An overall statistic for testing symmetry in social interactions

The present work focuses the attention on the skew-symmetry index as a measure of social reciprocity. This index is based on the correspondence between the amount of behaviour that each individual addresses to its partners and what it receives from them in return. Although the skew-symmetry index enables researchers to describe social groups, statistical inferential tests are required. The main aim of the present study is to propose an overall statistical technique for testing symmetry in experimental conditions, calculating the skew-symmetry statistic (Φ) at group level. Sampling distributions for the skew- symmetry statistic have been estimated by means of a Monte Carlo simulation in order to allow researchers to make statistical decisions. Furthermore, this study will allow researchers to choose the optimal experimental conditions for carrying out their research, as the power of the statistical test has been estimated. This statistical test could be used in experimental social psychology studies in which researchers may control the group size and the number of interactions within dyads.

Social reciprocity measures based on dyadic discrepancies: Methodological review and software development

This paper examines statistical analysis of social reciprocity at group, dyadic, and individual levels. Given that testing statistical hypotheses regarding social reciprocity can be also of interest, a statistical procedure based on Monte Carlo sampling has been developed and implemented in R in order to allow social researchers to describe groups and make statistical decisions.

Testing of Old Reinforced Concrete Bridges, HR-390, 1997

The objective of this research project was to service load test a representative sample of old reinforced concrete bridges (some of them historic and some of them scheduled for demolition) with the results being used to create a database so the performance of similar bridges could be predicted. The types of bridges tested included two reinforced concrete open spandrel arches, two reinforced concrete filled spandrel arches, one reinforced concrete slab bridge, and one two span reinforced concrete stringer bridge. The testing of each bridge consisted of applying a static load at various locations on the bridges and monitoring strains and deflections in critical members. The load was applied by means of a tandem axle dump truck with varying magnitudes of load. At each load increment, the truck was stopped at predetermined transverse and longitudinal locations and strain and deflection data were obtained. The strain data obtained were then evaluated in relation to the strain values predicted by traditional analytical procedures and a carrying capacity of the bridges was determined based on the experimental data. The response of a majority of the bridges tested was considerably lower than that predicted by analysis. Thus, the safe load carrying capacities of the bridges were greater than those predicted by the analytical models, and in a few cases, the load carrying capacities were found to be three or four times greater than calculated values. However, the test results of one bridge were lower than those predicted by analysis and thus resulted in the analytical rating being reduced. The results of the testing verified that traditional analytical methods, in most instances, are conservative and that the safe load carrying capacities of a majority of the reinforced concrete bridges are considerably greater than what one would determine on the basis of analytical analysis alone. In extrapolating the results obtained from diagnostic load tests to levels greater than those placed on the bridge during the load test, care must be taken to ensure safe bridge performance at the higher load levels. To extrapolate the load test results from the bridges tested in this investigation, the method developed by Lichtenstein in NCHRP Project 12-28(13)A was used.

Refinement of Rapid PCC Coarse Aggregate Testing Program, HR-1071, 1999

The characterization and categorization of coarse aggregates for use in portland cement concrete (PCC) pavements is a highly refined process at the Iowa Department of Transportation. Over the past 10 to 15 years, much effort has been directed at pursuing direct testing schemes to supplement or replace existing physical testing schemes. Direct testing refers to the process of directly measuring the chemical and mineralogical properties of an aggregate and then attempting to correlate those measured properties to historical performance information (i.e., field service record). This is in contrast to indirect measurement techniques, which generally attempt to extrapolate the performance of laboratory test specimens to expected field performance. The purpose of this research project was to investigate and refine the use of direct testing methods, such as X-ray analysis techniques and thermal analysis techniques, to categorize carbonate aggregates for use in portland cement concrete. The results of this study indicated that the general testing methods that are currently used to obtain data for estimating service life tend to be very reliable and have good to excellent repeatability. Several changes in the current techniques were recommended to enhance the long-term reliability of the carbonate database. These changes can be summarized as follows: (a) Limits that are more stringent need to be set on the maximum particle size in the samples subjected to testing. This should help to improve the reliability of all three of the test methods studied during this project. (b) X-ray diffraction testing needs to be refined to incorporate the use of an internal standard. This will help to minimize the influence of sample positioning errors and it will also allow for the calculation of the concentration of the various minerals present in the samples. (c) Thermal analysis data needs to be corrected for moisture content and clay content prior to calculating the carbonate content of the sample.

Development of an Economic Dust Palliative for Limestone Surfaced Secondary Roads, HR-297, 1990

This research project was directed at laboratory and field evaluation of sodium montmorillonite clay (bentonite) as a dust palliative for limestone surfaced secondary roads. It had been postulated that the electrically charged surfaces of the clay particles could interact with the charged surfaces of the limestone and act as a bonding agent to agglomerate fine (-#200) particulates and also to band the fine particulates to larger (+#200) limestone particles. Laboratory testing using soda ash dispersed bentonite treatment of limestone fines indicated significant improvement of compressive strength and slaking characteristics. It was recommended that the project proceed to field trials and test roads were constructed in Dallas and Adair counties in Iowa. Soda ash dispersed bentonite solutions can be field mixed and applied with conventional spray distribution equipment. A maximum of 1.5% bentonite(by weight of aggregate)can be applied at one time. Higher applications would have to be staged allowing the excess moisture to evaporate between applications. Construction of higher application treatments can be accomplished by adding dry bentonite to the surfacing material and then by dry road mixing. The soda ash water solution can then be spray applied and the treated surfacing material wet mixed by motor graders to a consistency of 3 to 4 inch slump concrete. Two motor graders working in tandem can provide rapid mixing for both methods of construction. Calcium and magnesium chloride treatments are 2 to 3 times more effective in dust reduction in the short term (3-4 months) but are prone to washboarding and potholing due to maintenance restrictions. Bentonite treatment at the 2-3% level is estimated to provide a 30-40% dust reduction over the long term(18-24 months). Normal maintenance blading operations can be used on bentonite treated areas. Vehicle braking characteristics are not adversely affected up to the 3.0% treatment level. The bentonite appears to be functioning as a banding agent to bind small particulates to larger particles and is acting to agglomerate fine particles of limestone. This bonding capability appears recoverable from environmental effects of winter, and from alternating wet and dry periods. The bentonite appears to be able to interact with new applications of limestone maintenance material and maintains a dust reduction capability. Soda ash dispersed bentonite treatment is approximately 10 times more cost effective per percent dust reduction than conventional chloride treatments with respect to time. However,the disadvantage is that there is not the initial dramatic reduction in dust generation as with the chloride treatment. Although dust is reduced 30-40% after treatment there is still dust being generated and the traveling public or residents may not perceive the reduction.

Bentonite Treatment for Economical Dust Reduction on Limestone Surfaced Secondary Roads, HR-351, 1995

This research project was directed at laboratory and field evaluation of sodium montmorillonite clay (Bentonite) as a dust palliative for limestone surfaced secondary roads. It was postulated that the electrically charged surfaces (negative) of the clay particles could interact with the charged surfaces (positive) of the limestone and act as a bonding agent to agglomerate fine (-#200) particulates, and also to bond the fine particulates to larger (+#200) limestone particles. One mile test roads were constructed in Tama, Appanoose, and Hancock counties in Iowa using Bentonite treatment levels (by weight of aggregate) ranging from 3.0 to 12.0%. Construction was accomplished by adding dry Bentonite to the surfacing material and then dry road mixing. The soda ash/water solution (dispersing agent) was spray applied and the treated surfacing material wet mixed by motor graders to a consistency of 2 to 3 inch slump concrete. Two motor graders working in tandem provided rapid mixing. Following wet mixing the material was surface spread and compacted by local traffic. Quantitative and qualitative periodic evaluations and testing of the test roads was conducted with respect to dust generation, crust development, roughness, and braking characteristics. As the Bentonite treatment level increased dust generation decreased. From a cost/benefit standpoint, an optimum level of treatment is about 8% (by weight of aggregate). For roads with light traffic, one application at this treatment level resulted in a 60-70% average dust reduction in the first season, 40-50% in the second season, and 20-30% in the third season. Crust development was rated at two times better than untreated control sections. No discernible trend was evident with respect to roughness. There was no evident difference in any of the test sections with respect to braking distance and braking handling characteristics, under wet surface conditions compared to the control sections. Chloride treatments are more effective in dust reduction in the short term (3-4 months). Bentonite treatment is capable of dust reduction over the long term (2-3 seasons). Normal maintenance blading operations can be used on Bentonite treated areas. Soda ash dispersed Bentonite treatment is estimated to be more than twice as cost effective per percent dust reduction than conventional chloride treatments, with respect to time. However, the disadvantage is that there is not the initial dramatic reduction in dust generation as with the chloride treatment. Although dust is reduced significantly after treatment there is still dust being generated. Video evidence indicates that the dust cloud in the Bentonite treated sections does not rise as high, or spread as wide as the cloud in the untreated section. It also settles faster than the cloud in the untreated section. This is considered important for driving safety of following traffic, and for nuisance dust invasion of residences and residential areas. The Bentonite appears to be functioning as a bonding agent.

Paternity test in "Mangalarga-Marchador" equines by DNA-fingerprinting

GC-rich molecular minisatellite probes isolated from the human genome have presented a poor ability for individualization in horses. In this study new DNA sequences were isolated which could be used in paternity tests in horses. Genomic DNA from "Mangalarga-Marchador" horses was treated with restriction enzymes that preferentially digest non-repetitive sequences, so preserving the structure where mini and microsatellites are located. Four clones (S01, S05, S07 and S09) selected from a genomic library screened with a (TG)n oligonucleotide showed similar hybridization profiles generating bands of DNA-fingerprinting type. Using these probes the individualization power obtained was 10-8, which is 10(5)fold higher than that obtained with M13, another GC-rich type probe. All clones were efficient in parentage detection in crossbreedings and presented a 27 bp consensus sequence, GTTTCATTTATTATTCTTTGGAAGAAA, which was repeated 12, 18, 11 and 21 times in clones S01, S05, S07 and S09, respectively.

Soil Tracking Paint Test, MLR-89-13, 1989

A new paint testing device was built to determine the resistance of paints to darkening due to road grime being tracked onto them. The device consists of a tire rotating on a sample drum. Soil was applied to the tire and then tracked onto paint samples which were attached to the drum. A colorimeter was used to measure the lightness of the paints after being tracked. Lightness is measured from 0 (absolute black) to 100 (absolute white). Four experiments were run to determine the optimum time length to track a sample, the reproducibility, the effects of different soils, and the maximum acceptable level for darkening of a paint. The following conclusions were reached: 1) the optimum tracking time was 10 minutes; 2) the reproducibility had a standard deviation of 1.5 lightness units; 3) different soils did not have a large effect on the amount of darkening on the paints; 4) a maximum acceptable darkness could not be established based on the limited amount of data; and 5) a correlation exists between the paints which were darkening in the field and the paints which were turning the darkest on the tracking wheel.

Accelerated Testing Track, HR-7, 1951

In May 1950 a proposal for a research project was submitted to the newly formed Iowa Highway Research Board for consideration and action. This project, designated RPSl by the Board, encompassed the study, development, preparation of preliminary plans and specifications for the construction of a wheel track to be used in the accelerated testing of highway pavements. The device envisioned in the proposal was a circular track about seventy-five feet in diameter equipped with a suitable automobile-tired device to test pavements about five feet in width laid into the track under regular construction practices by small scale construction equipment. The Board, upon review, revised and expanded the basic concepts of the project. The project as revised by the Board included a study of the feasibility of developing, constructing and operating an accelerated testing track in which pavements, bases and subgrades may be laid one full lane, or at least ten feet, in width by full size construction equipment in conformity with usual construction practices. The pavements so laid are to be subjected, during test, to conditions as nearly simulating actual traffic as possible.

FastEpistasis: a high performance computing solution for quantitative trait epistasis.

Motivation: Genome-wide association studies have become widely used tools to study effects of genetic variants on complex diseases. While it is of great interest to extend existing analysis methods by considering interaction effects between pairs of loci, the large number of possible tests presents a significant computational challenge. The number of computations is further multiplied in the study of gene expression quantitative trait mapping, in which tests are performed for thousands of gene phenotypes simultaneously. Results: We present FastEpistasis, an efficient parallel solution extending the PLINK epistasis module, designed to test for epistasis effects when analyzing continuous phenotypes. Our results show that the algorithm scales with the number of processors and offers a reduction in computation time when several phenotypes are analyzed simultaneously. FastEpistasis is capable of testing the association of a continuous trait with all single nucleotide polymorphism ( SNP) pairs from 500 000 SNPs, totaling 125 billion tests, in a population of 5000 individuals in 29, 4 or 0.5 days using 8, 64 or 512 processors.

Testing extreme value copulas to estimate the quantile

Testing weather or not data belongs could been generated by a family of extreme value copulas is difficult. We generalize a test and we prove that it can be applied whatever the alternative hypothesis. We also study the effect of using different extreme value copulas in the context of risk estimation. To measure the risk we use a quantile. Our results have motivated by a bivariate sample of losses from a real database of auto insurance claims. Methods are implemented in R.

Bruce Treadmill Test in Children : First Normal Values for Children from Western Europe : P67

Introduction: One of the main goals for exereise testing in children is evaluation of exercise capacity. There are many testing protocols, but the Bruce treadmill protocol is widely used among pediatrie cardiology centers. Thirty years ago, Cuming et al. were the first to establish normal values for children from North America (Canada) aged 4 to 18 years old. No data was ever published for children from Western Europe. Our study aimed to assess the validity of the normal values from Cuming et al. for children from Western Europe in the 21 st century. Methods: It is a retrospective cohort study in a tertiary care children's hospital. 144 children referred to our institution but finally diagnosed as having a normal heart underwent exercise stress testing using the Bruce protocol between 1999 and 2006. Data from 59 girls and 85 boys aged 6 to 18 were reviewed. Mean endurance time (ET) for each age category and gender was compared with the mean normal values fram Cumming et al by an unpaired t-test. Results: Mean ET increases with age until 15 years old in girls and then decreases. Mean endurance time increases continuouslY'from 6 to 18 years old in boys. The increase is more pronounced in boys than girls. In our study, a significant higher mean ET was found for boys in age categories 10 to 12, 13 to 15 and 16 to 18. No significant difference was found in any other groups. Conclusions: Some normal values from Cuming et al. established in 1978 for ET with the Bruce protocol are probably not appropriate any more today for children from Western Europe. Our study showed that mean ET is higher for boys from 10 to 18 years old. Despite common beliefs, cardiovascular conditioning doesn't seem yet reduced in children from Western Europe. New data for Bruce treadmill exercise. testing for healthy children, 4 to 18 years old, living in Western Europe are required. .

The QUASAR reproducibility study, Part II: Results from a multi-center Arterial Spin Labeling test-retest study.

Arterial Spin Labeling (ASL) is a method to measure perfusion using magnetically labeled blood water as an endogenous tracer. Being fully non-invasive, this technique is attractive for longitudinal studies of cerebral blood flow in healthy and diseased individuals, or as a surrogate marker of metabolism. So far, ASL has been restricted mostly to specialist centers due to a generally low SNR of the method and potential issues with user-dependent analysis needed to obtain quantitative measurement of cerebral blood flow (CBF). Here, we evaluated a particular implementation of ASL (called Quantitative STAR labeling of Arterial Regions or QUASAR), a method providing user independent quantification of CBF in a large test-retest study across sites from around the world, dubbed "The QUASAR reproducibility study". Altogether, 28 sites located in Asia, Europe and North America participated and a total of 284 healthy volunteers were scanned. Minimal operator dependence was assured by using an automatic planning tool and its accuracy and potential usefulness in multi-center trials was evaluated as well. Accurate repositioning between sessions was achieved with the automatic planning tool showing mean displacements of 1.87+/-0.95 mm and rotations of 1.56+/-0.66 degrees . Mean gray matter CBF was 47.4+/-7.5 [ml/100 g/min] with a between-subject standard variation SD(b)=5.5 [ml/100 g/min] and a within-subject standard deviation SD(w)=4.7 [ml/100 g/min]. The corresponding repeatability was 13.0 [ml/100 g/min] and was found to be within the range of previous studies.

Optimization of Snow Drifting Mitigation and Control Methods for Iowa Conditions Part II, TR-626, 2015

Blowing and drifting of snow is a major concern for transportation efficiency and road safety in regions where their development is common. One common way to mitigate snow drift on roadways is to install plastic snow fences. Correct design of snow fences is critical for road safety and maintaining the roads open during winter in the US Midwest and other states affected by large snow events during the winter season and to maintain costs related to accumulation of snow on the roads and repair of roads to minimum levels. Of critical importance for road safety is the protection against snow drifting in regions with narrow rights of way, where standard fences cannot be deployed at the recommended distance from the road. Designing snow fences requires sound engineering judgment and a thorough evaluation of the potential for snow blowing and drifting at the construction site. The evaluation includes site-specific design parameters typically obtained with semi-empirical relations characterizing the local transport conditions. Among the critical parameters involved in fence design and assessment of their post-construction efficiency is the quantification of the snow accumulation at fence sites. The present study proposes a joint experimental and numerical approach to monitor snow deposits around snow fences, quantitatively estimate snow deposits in the field, asses the efficiency and improve the design of snow fences. Snow deposit profiles were mapped using GPS based real-time kinematic surveys (RTK) conducted at the monitored field site during and after snow storms. The monitored site allowed testing different snow fence designs under close to identical conditions over four winter seasons. The study also discusses the detailed monitoring system and analysis of weather forecast and meteorological conditions at the monitored sites. A main goal of the present study was to assess the performance of lightweight plastic snow fences with a lower porosity than the typical 50% porosity used in standard designs of such fences. The field data collected during the first winter was used to identify the best design for snow fences with a porosity of 50%. Flow fields obtained from numerical simulations showed that the fence design that worked the best during the first winter induced the formation of an elongated area of small velocity magnitude close to the ground. This information was used to identify other candidates for optimum design of fences with a lower porosity. Two of the designs with a fence porosity of 30% that were found to perform well based on results of numerical simulations were tested in the field during the second winter along with the best performing design for fences with a porosity of 50%. Field data showed that the length of the snow deposit away from the fence was reduced by about 30% for the two proposed lower-porosity (30%) fence designs compared to the best design identified for fences with a porosity of 50%. Moreover, one of the lower-porosity designs tested in the field showed no significant snow deposition within the bottom gap region beneath the fence. Thus, a major outcome of this study is to recommend using plastic snow fences with a porosity of 30%. It is expected that this lower-porosity design will continue to work well for even more severe snow events or for successive snow events occurring during the same winter. The approach advocated in the present study allowed making general recommendations for optimizing the design of lower-porosity plastic snow fences. This approach can be extended to improve the design of other types of snow fences. Some preliminary work for living snow fences is also discussed. Another major contribution of this study is to propose, develop protocols and test a novel technique based on close range photogrammetry (CRP) to quantify the snow deposits trapped snow fences. As image data can be acquired continuously, the time evolution of the volume of snow retained by a snow fence during a storm or during a whole winter season can, in principle, be obtained. Moreover, CRP is a non-intrusive method that eliminates the need to perform man-made measurements during the storms, which are difficult and sometimes dangerous to perform. Presently, there is lots of empiricism in the design of snow fences due to lack of data on fence storage capacity on how snow deposits change with the fence design and snow storm characteristics and in the estimation of the main parameters used by the state DOTs to design snow fences at a given site. The availability of such information from CRP measurements should provide critical data for the evaluation of the performance of a certain snow fence design that is tested by the IDOT. As part of the present study, the novel CRP method is tested at several sites. The present study also discusses some attempts and preliminary work to determine the snow relocation coefficient which is one of the main variables that has to be estimated by IDOT engineers when using the standard snow fence design software (Snow Drift Profiler, Tabler, 2006). Our analysis showed that standard empirical formulas did not produce reasonable values when applied at the Iowa test sites monitored as part of the present study and that simple methods to estimate this variable are not reliable. The present study makes recommendations for the development of a new methodology based on Large Scale Particle Image Velocimetry that can directly measure the snow drift fluxes and the amount of snow relocated by the fence.