Paired Comparison Design Research Articles

Subjective Values of Quality of Life Dimensions in Elderly People. A SEM Preference Model Approach

This article proposes a Thurstonian model in the framework of Structural Equation Modelling (SEM) to assess preferences among quality of life dimensions for the elderly. Data were gathered by a paired comparison design in a sample comprised of 323 people aged from 65 to 94 years old. Five dimensions of quality of life were evaluated: Health, Autonomy, Family and Social Support, Social Activities and Home Conditions. An unrestricted model was estimated and subjective preference values of dimensions were obtained. According to the subjective preferences, four groups of dimensions were established: Health-Autonomy-Home and Support-Social Activities. No differences among gender and age were found in the preference values.

Approximate and exact optimal designs for paired comparisons of partial profiles when there are two groups of factors

For paired comparison experiments involving pairs of multifactor options differing in a specified number of factors the problem of finding optimal designs is considered, when only main effects are to be estimated. It is presumed that the set of factors can be partitioned into two groups such that the number of levels is constant within each group. The optimal designs for this frequently encountered case are also optimal for the corresponding choice experiments under the hypothesis that the parameters in the multinomial logit model are equal to zero.

Assessing vocational interests in the Basque Country using paired comparison design

This article proposes the Thurstonian paired comparison model to assess vocational preferences and uses this approach to evaluate the Realistic, Investigative, Artistic, Social, Enterprise, and Conventional (RIASEC) model in the Basque Country (Spain). First, one unrestricted model is estimated in the Structural Equation Modelling framework using one paired comparison design. Secondly, the RIASEC circumplex structure is evaluated upon the estimated latent preferences correlation matrix. The results show good fit between the Thurstonian model and the data, but do not confirm the RIASEC structure in the Basque population.

Intervention Research and Program Evaluation in the School Setting: Issues and Alternative Research Designs

This article discusses the difficulties in conducting intervention research or evaluating intervention programs in a school setting. In particular, the problems associated with randomization and obtaining control groups are examined. The use of quasi-experimental designs, specifically a paired comparison design using the individual as his or her own control in combination with a Separate-Sample Pretest-Posttest Control Group Design, are suggested as one means of increasing internal and external validity while overcoming the problems associated with implementing a true experimental design.

Can Self-declared Personal Values be Used to Identify those with Family Medicine Career Aspirations?

Self-declaration of personal values has been suggested as a means of identifying students with greater predilection for future primary care careers. While statistically significant differences have been demonstrated, absolute differences between those interested in primary care and those interested in specialist careers tend to be small. This study's objective was to determine whether limited differences can nevertheless be used to identify those with particular interest in primary care. Schwartz' values were submitted to a paired comparison design in which respondents were asked to identify the value they held most dearly for possible pairings of values. 88 medical school applicants, 57 first year medical students, 78 final year medical students, and 34 admissions interviewers participated. Applicant and medical student subgroups were analyzed as a whole and as a function of self-identified career interest (i.e., primarily interested in Family medicine or other specialty careers). The values statements were remarkably consistent (r > 0.90) between groups, regardless of which subgroups were analyzed. Despite apparent differences in the literature between those interested in primary care and those interested in other specialist careers, the differences are small and do not correlate with career aspirations in a way that could inform admissions decisions.

Optimal Designs for Asymmetric Linear Paired Comparisons with a Profile Strength Constraint

Within the linear model framework the problem of determining optimal designs for paired comparisons of alternatives which are described by a set of discrete attributes is considered under the constraint that the alternatives in a pair are only allowed to differ with regard to a certain number of attributes. Whereas in previous treatments of this problem it was assumed that all attributes possess the same number of levels, here the general asymmetric case is discussed. We provide a characterization of optimal designs and demonstrate how this can be used to derive a solution of the design problem for many situations of interest.

Statistics for experimenters: design, innovation, and discovery

Statistics for experimenters: design, innovation, and discovery

The role of texture in shape from shading: are humans biased towards seeing relief textures?

When a textured surface is modulated in depth and illuminated, the level of illumination varies across the surface, producing coarse-scale luminance modulations (LM) and amplitude modulation (AM) of the fine-scale texture. If the surface has an albedo texture (reflectance variation) then the LM and AM components are always in-phase, but if the surface has a relief texture the phase relation between LM and AM varies with the direction and nature of the illuminant. We showed observers sinusoidal luminance and amplitude modulations of a binary noise texture, in various phase relationships, in a paired-comparisons design. In the first experiment, the combinations under test were presented in different temporal intervals. Observers indicated which interval contained the more depthy stimulus. LM and AM in-phase were seen as more depthy than LM alone which was in turn more depthy than LM and AM in anti-phase, but the differences were weak. In the second experiment the combinations under test were presented in a single interval on opposite obliques of a plaid pattern. Observers were asked to indicate the more depthy oblique. Observers produced the same depth rankings as before, but now the effects were more robust and significant. Intermediate LM/AM phase relationships were also tested: phase differences less than 90 deg were seen as more depthy than LM-only, while those greater than 90 deg were seen as less depthy. We conjecture that the visual system construes phase offsets between LM and AM as indicating relief texture and thus perceives these combinations as depthy even when their phase relationship is other than zero. However, when different LM/AM pairs are combined in a plaid, the signals on the obliques are unlikely to indicate corrugations of the same texture, and in this case the out-of-phase pairing is seen as flat. [Supported by the Engineering and Physical Sciences Research Council (EPSRC)].

The validity and reliability of the global index of safety (GIS)

Objective: The global index of safety (GIS) is an adverse event (AE) based instrument designed to evaluate the safety profile of drugs. This paper presents the evaluation of the inter-rater reliability and validity of a 94-item GIS for antipsychotics through Rasch analysis.Research design and methods: A total of 194 psychiatrists participating in an outpatient pharmacoepidemiologic study of olanzapine in schizophrenia rated the severity that each AE would have on a 5-point scale. Reliability wasdetermined through a paired comparison design involving the new independent ratings of 101 different psychiatrists participating in another study of olanzapine in acute inpatient units. Spearman's, Pearson's and Intra-class correlation (ICC) coefficients were used to estimate the interrater reliability of the AE weights. Validity was analyzed through the Rasch rating scale model.Results: Reliability coefficient estimates were excellent (Spearman = 0.99, Pearson = 0.99, ICC = 0.98), supporting the inter-rater reliability of the item weights. Through goodness-of-fit statistics and the investigation of the hierarchy of item calibrations, Rasch analysis confirmed the validity of the instrument.Conclusion: The data presented here on interrater reliability estimates of adverse events related to antipsychotic drugs indicate that GIS is a promising alternative for the evaluation of the safety profile of drugs.

実験デザインに基づく一対比較データの解析

Paired comparison is a useful method for assessing ranks among several objects, and it enables us to obtain more reliable data than assessing objects one by one. But paired comparison principally provides information only about the ranks of the objects. On the other hand, experimental design provides a framework for elucidating causal associations. If we could analyze paired comparison data by the experimental design framework, it would be a very effective method. But experimental design, in its original form, is not readily applicable to paired comparison data. However, if we adopt the perspective of structural equation modeling (SEM), we can deal with paired comparison and experimental design in a unified way, because they are both submodels of SEM. The purpose of this study is to provide a new method to analyze causal connection of paired comparison data by using SEM. Here, two actual numerical examples are shown, one of which is obtained by within-subject design and the other is obtained by between-subject design.

Adaptive paired comparison design

An important aspect of paired comparison experiments is the decision of how to form pairs in advance of collecting data. A weakness of typical paired comparison experimental designs is the difficulty in incorporating prior information, which can be particularly relevant for the design of tournament schedules for players of games and sports. Pairing methods that make use of prior information are often ad hoc algorithms with little or no formal basis. The problem of pairing objects can be formalized as a Bayesian optimal design. Assuming a linear paired comparison model for outcomes, we develop a pairing method that maximizes the expected gain in Kullback–Leibler information from the prior to the posterior distribution. The optimal pairing is determined using a combinatorial optimization method commonly used in graph-theoretic contexts. We discuss the properties of our optimal pairing criterion, and demonstrate our method as an adaptive procedure for pairing objects multiple times. We compare the performance of our method on simulated data against random pairings, and against a system that is currently in use in tournament chess.

A comparative evaluation of accelerated solvent extraction and Polytron extraction for quantification of lipids and extractable organochlorine in fish

Optimal conditions for extraction methods using an accelerated solvent extractor (ASE) and a Polytron homogenizer were established for quantification of (crude) lipids and extractable organochlorine (EOCl) in fish. The two methods, an ASE consecutive extraction at 55 and 100 °C for ground, freeze-dried filets, and a Polytron extraction at room temperature with four repetitions for ground, wet filets and with two repetitions for wet fish livers were evaluated in terms of reproducibility and comparability. With respect to lipid measurement, the two methods have comparable reproducibility and the experimental errors are relatively small compared to natural variations of fish to fish within a sampling site. For EOCl measurement, the relative standard deviation tends to increase with decreasing levels of EOCl in specimens. An important factor for the relative standard deviation of measurements of analytes present at trace levels in fish matrices is sampling error in taking aliquots. Statistical analysis using a paired comparison design showed that measurement values (crude lipids or EOCl) obtained from these two extraction methods are, in general, not equivalent. There is a good correlation between the two methods for lipid measurement; whereas for EOCl measurements, high degrees of correlation exist within the exposed fish group, but not in the reference group.

Optimal paired comparison designs for first-order interactions

In many fields of applications paired comparisons are used in which either full or partial profiles of the alternatives are presented. For this situation we introduce an appropriate model and derive optimal designs in the presence of interactions when all attributes have the same number of levels.

Perception of taste intensity in solutions of random-coil polysaccharides above and below c∗

Sensory paired comparison tests were used to study differences in taste intensity in solutions of hydroxypropylmethyl cellulose (HPMC) at concentrations above (1.0% w/w) and below (0.2% w/w) c ∗, the coil-overlap concentration (the point at which viscosity changes abruptly with increasing thickener). The sweetness intensities of aspartame (250 ppm), sucrose (5% w/w), fructose (4.5% w/w) and neohesperidin dihydrochalcone (39 ppm) and the saltiness of sodium chloride (0.35%) were all found to be significantly reduced in the more viscous HPMC solution. There was no significant effect of HPMC concentration on the acidity of citric acid (600 ppm) or the bitterness of quinine hydrochloride (26 ppm). The sweetness intensities of sucrose and aspartame were likewise investigated in two further hydrocolloid solutions, guar gum and λ-carrageenan. Experiments were designed so that the ratios of the thickener concentrations (above and below c ∗) to their measured c ∗ values remained constant. The sweetness of sucrose was found to be significantly reduced in the more viscous guar gum solution ( P<0.05) and that of aspartame was reduced in the λ-carrageenan above c ∗ ( P<0.001). A multiple paired comparison design was used to show that the perceived sweetness of 6.5% sucrose in 1.0% HPMC did not differ significantly from that of 5% sucrose in 0.2% HPMC. The magnitude of effect with aspartame was broadly analogous.

D-optimal weighted paired comparison designs

As a generalization of the paired comparison designs, the experimental design which considers the weighted difference between responses is studied. D-optimal designs are constructed for a model with main effects and first-order interactions. The experimental region is a hypercube or a hypersphere.

OPTIMAL DESIGNS FOR 2 k PAIRED COMPARISON EXPERIMENTS

In this paper we establish the form of the optimal paired comparison design when there are k attributes, each with two levels, for testing for main effects, for main effects and two factor interactions and for main effects and two and three factor interactions. In all cases we assume that all pairs with the same number of attributes different appear equally often. In this setting the D and A optimal designs for main effects are the foldover pairs and those for main effects and two factor interactions have pairs in which about half the attributes are different.

Attention and recognition memory in the 1st year of life: A longitudinal study of preterm and full-term infants.

Several aspects of visual attention and their implications for recognition memory were examined in a longitudinal sample of full-term and preterm (birth weight < 1,750 g) infants seen at 5, 7, and 12 months of age. At all 3 ages, full-terms had shorter look durations, faster shift rates, less off-task behavior, and higher novelty scores than preterms. Both groups followed similar developmental trajectories, with older infants having shorter looks and more shifts. Infants were consistent in attentional style across problems of the same type, across problems that used different types of stimuli (faces and patterns), and across the familiarization and test phases of this paired-comparison design; there was also modest cross-age stability. Shorter looks and higher shift rates during familiarization were related to better recognition memory, with shift rate adding to prediction independently of either peak or mean look. These findings underscore the importance of attention to infant information processing.

On the combined use of accented and unaccented diphones in speech synthesis

In this paper, three experiments are reported that were run in order to assess the quality of Dutch synthetic speech using accented and unaccented diphones, i.e., diphones extracted from accented and unaccented syllables, respectively. In a paired-comparison design, subjects were asked to evaluate the naturalness and fluency of different versions of an utterance. The results of the first two experiments, in which isolated polysyllabic words were used, indicate that the use of accented or unaccented diphones has a perceptual effect on phonologically long vowels only. In a third experiment, the use of the different diphone types in short sentences with a fixed temporal structure was evaluated. Results suggest that using unaccented diphones in unaccented or secondarily accented syllables does not result systematically in more natural-sounding speech.

Recurrent pericarditis. Relief with colchicine.

Recurrence is one of the major complications of pericarditis. Treatment of recurrence is often difficult, and immunosuppressive drugs or surgery may be necessary. We conducted an open-label prospective study of nine patients (seven men and two women; age, 18-64 years; mean age, 41.7 +/- 13.7 years). Patients were treated with colchicine (1 mg/day) to prevent recurrences. All patients had suffered at least three relapses despite treatment with acetylsalicylic acid, indomethacin, prednisone, or a combination. Pericarditis was classified as idiopathic in five patients, postpericardiotomy in two, post-myocardial infarction in one, and associated with disseminated lupus erythematosus in one. For statistical analysis, we conducted a paired comparison design (Student's t test). All patients treated with colchicine responded favorably to therapy. Prednisone was discontinued in all patients after 2-6 weeks (mean, 26.33 +/- 10.9 days), and colchicine alone was continued. After a mean follow-up of 24.3 months (minimum, 10 months; maximum, 54 months), no recurrences were observed in any patient; there was a significant difference between the symptom-free periods before and after treatment with colchicine (p less than 0.002). Our study suggests that colchicine may be useful in avoiding recurrence of pericarditis, although these results need to be confirmed in a larger, double-blind study.

Reduction of the number of pairs in paired comparison designs and exact designs for quadratic models

The construction of designs in paired comparisons experiments is considered. In Berkum (1987) D-optimal designs are given under the assumption of no treatment differences. These designs can not be used in most practical applications since they entail different weights. Also, the number of pairs is (too) large. In this paper D-optimal designs are given with a reduced number of pairs (approximate theory) and exact designs are given with a high efficiency. This is done for a quadratic model for 2, 3, 4 and 5 dimensions and a hypercube as experimental region.