Multiple Response Model Research Articles

An analysis of complex multiple‐choice science–technology–society items: Methodological development and preliminary results

AbstractThe scarce attention to the assessment and evaluation in science education research has been especially harmful for teaching science–technology–society (STS) issues, due to the dialectical, tentative, value‐laden, and polemic nature of most STS topics. This paper tackles the methodological difficulties of the instruments that monitor views related to STS topics and rationalizes a quantitative methodology and an analysis technique to improve the utility of an empirically developed multiple‐choice item pool, the Questionnaire of Opinions on STS. This methodology embraces an item‐scaling psychometrics based on the judgments by a panel of experts, a multiple response model, a scoring system, and the data analysis. The methodology finally produces normalized attitudinal indices that represent the respondent's reasoned beliefs toward STS statements, the respondent's position on an item that comprises several statements, or the respondent's position on an entire STS topic that encompasses a set of items. Some preliminary results show the methodology's ability to evaluate the STS attitudes in a qualitative and quantitative way and for statistical hypothesis testing. Lastly, some applications for teacher training and STS curriculum development in science classrooms are discussed. © 2006 Wiley Periodicals, Inc. Sci Ed 90:681–706, 2006

A response-modeling alternative to surrogate models for support in computational analyses

Often, the objectives in a computational analysis involve characterization of system performance based on some function of the computed response. In general, this characterization includes (at least) an estimate or prediction for some performance measure and an estimate of the associated uncertainty. Surrogate models can be used to approximate the response in regions where simulations were not performed. For most surrogate modeling approaches, however, (1) estimates are based on smoothing of available data and (2) uncertainty in the response is specified in a point-wise (in the input space) fashion. These aspects of the surrogate model construction might limit their capabilities. One alternative is to construct a probability measure, G( r), for the computer response, r, based on available data. This “response-modeling” approach will permit probability estimation for an arbitrary event, E( r), based on the computer response. In this general setting, event probabilities can be computed: prob( E)=∫ r I( E( r))d G( r) where I is the indicator function. Furthermore, one can use G( r) to calculate an induced distribution on a performance measure, pm. For prediction problems where the performance measure is a scalar, its distribution F pm is determined by: F pm(z)=∫ r I(pm( r)⩽z)d G( r). We introduce response models for scalar computer output and then generalize the approach to more complicated responses that utilize multiple response models.

Developmental expression of the major human hepatic CYP3A enzymes.

The human cytochrome P4503A forms show expression patterns subject to developmental influence. CYP3A7 and CYP3A4 are generally classified as the major fetal and adult liver forms, respectively. However, characterization of CYP3A4, -3A5, and -3A7 developmental expression has historically been confounded by the lack of CYP3A isoform-specific antibodies or marker enzyme activities. Therefore, the objective of this study was to characterize the developmental expression of hepatic CYP3A forms from early gestation to 18 years of age using up to 212 fetal and pediatric liver samples. Based on immunoquantitation, CYP3A5 protein expression was found to be highly variable, generally independent of age, and more frequently observed for African-American individuals. For differentiation of CYP3A4 and -3A7 levels, dehydroepiandrosterone metabolite patterns for expressed CYP3A forms were characterized and used for simultaneous quantitation of protein levels within liver microsome samples. The major metabolite formed by CYP3A4, 7beta-hydroxy-dehydroepiandrosterone, was identified based on cochromatography and mass spectra matching with the authentic standard. Kinetic analysis showed a 34-fold greater intrinsic clearance of 7beta-hydroxy-dehydroepiandrosterone by CYP3A4 versus -3A7, whereas CYP3A7 showed the highest 16alpha-hydroxy-dehydroepiandrosterone intrinsic clearance. Metabolite profiles for the expressed enzymes were fit to a multiple response model and CYP3A4 and -3A7 levels in fetal and pediatric liver microsome samples were calculated. Fetal liver microsomes showed extremely high CYP3A7 levels (311-158 pmol/mg protein) and significant expression through 6 months postnatal age. Low CYP3A4 expression was noted for fetal liver (< or =10 pmol/mg), with mean levels increasing with postnatal age.

Effect of high pressure processing on rennet coagulation properties of milk

The effect of high pressure (HP) treatment of milk on rennet coagulation properties were evaluated. Three coagulation parameters—lag time, mean coagulation rate and inflexion time (time for reaching the point of maximum coagulation rate) were evaluated as a function of pressure (200–400 MPa), temperature (3–21 °C) and holding time (10–110 min) using a response surface methodology. A central composite rotatable design and a second order multiple response model were used to investigate the effect of process variables (pressure, temperature and holding time) and relate them to coagulation parameters. In general, with an increase in pressure or holding time, or a decrease in temperature, lag and inflexion times associated with the rennet coagulation showed a decreasing trend while the mean coagulation rate showed the opposite. Canonical analyses indicated that the stationary point was a saddle point for lag time and mean coagulation rate whereas a minimum for inflexion time. At the stationary point, the predicted values were: lag time, 7.7 min (85 min HP treatment at 296 MPa and 5 °C); inflexion time, 11.3 min (50 min treatment at 370 MPa and 6 °C) and mean coagulation rate, 0.0034 min −1 (72 min at 275 MPa and 14 °C).

Water-holding capacity and gel strength of rennet curd as affected by high-pressure treatment of milk

The effect of high pressure (HP) treatment on two gel characteristics, water-holding capacity (WHC) and gel-strength (GS), of rennet curd was evaluated as a function of pressure (200–400 MPa), temperature (3–21°) and holding time (10–110min) using a response surface methodology. A central composite design was used to investigate the effect of process variables and a second order multiple response model was used to relate the pressure, temperature and holding time to WHC and GS. In general, with a decrease in pressure level, temperature and holding time, there was a decrease in water-holding capacity and an increase in the gel-strength of the rennet curds. The conditions of minimum of WHC (40%) were: pressure, 280 MPa; temperature, 9°C, and holding time, 42 min which also resulted in a high gel strength of 0.47N slightly below the maximum of 0.52N.

Rheology of tapioca starch

Rheological properties of tapioca starch solutions were evaluated using a computer controlled rotational viscometer in relation to four factors: temperature, 20–80°C; concentration, 2–6%; pH, 4–8 and cooking time, 5–35 min. All test solutions showed power-law flow behavior. A second order composite design was used to investigate the effects of concentration, temperature, pH value and cook time on rheological property parameters (consistency coefficient, m, and flow behavior index, n), and develop second order multiple response models of parameters related to concentration, temperature, pH value and cook time.

Response and scoring models for the 'Views on Science-Technology-Society' instrument

The understanding of science, technology, and their mutual relationships in and with society is a growing goal in educational systems. Researchers and educators are concerned about the assessment of STS topics. The 'Views on Science-Technology-Society' instrument (VOSTS) is an empirically developed multiple-choice item pool (114) that offers a wide coverage of STS topics and a methodology to assess them. However, the unique response model usually applied to answer VOSTS items does not allow the use of inferential statistics and does not take advantage of the large volume of available information. This paper addresses both issues, suggesting new response and scoring models to improve VOSTS implementation. First, the paper analyses the scaling properties of some scoring procedures within the unique response model and suggests ways to produce 'best fidelity' measurements. Second, the paper presents a multiple response model that could take advantage of all the information contained in each item, and puts forward a rating scale procedure to establish the model. This paper provides the rationale to justify the proposals: further empirical work is needed to test the application of the model.

Generalized least squares estimation of multivariate nonlinear models with missing data

The method of estimated generalized least squares estimation of multiple response models is extended to the randomly missing date case. This estimation procedure is computationally simply when there are many missing data but the number of distinct patterns of missing data for the response vectors is small. The consistency and asymptotic normality of the proposed estimators are established.

The Role of Learning in Chronic Pain States

SUMMARY This chapter provides a discussion of the role of learning and conditioning in the initiation, development and maintenance of clinical pain states, with emphasis on chronic pain. The chapter begins by providing a brief overview of conditioning/learning theory, concluding that an interactional/mixed theory of learning which accommodates respondent, operant and observational learning models offers the greatest degree of application. A multiple response model of clinical pain is then presented, with pain being defined as an interacting cluster of individualized overt, covert and neurophysiological responses capable of being produced by relevant tissue damage/irritation, as well as produced and maintained by other antecedent and consequent stimulus conditions. The point is established that pain responses, like all other behaviours, are subject to the principles and laws of conditioning/ learning. Given this, prevalent antecedent and consequent stimulus conditions typically seen in acute and chronic pain states are outlined, as well as primary and secondary pain responses typically encountered across these states. In addition to tissue damage/irritation, the observation is made that major environmental stressors can also play an important role in initiating and maintaining acute pain responses. Likewise, pain responses can be significantly influenced during the acute state by reinforcing consequences (i.e. the presentation of rewards or the removal of aversive conditions). Such conditions set the occasion for development of chronic pain states, which may be only secondarily influenced and maintained by tissue damage and irritation. A list of indicators which suggests that learning effects are making a significant contribution to pain responses is provided, with emphasis on identifying the presence of controlling antecedent or consequent stimulus conditions in addition to tissue damage/irritation. The chapter concludes by focusing on the need for incorporation of behavioural assessment and treatment methods in acute and chronic pain states that adequately evaluate and treat learning effects.

Implicit Multifunctional Nonlinear Regression Analysis

The least squares estimation of parameters in algebraically implicit, nonlinear, multiple response models having only one ezperimentally accessible response variable is treated within the context of a Gauss–Newton—Newton iteration. The algorithm, derived through application of the implicit function theorem to the model, is sufficiently general to cover Bayesian estimation of parameters for multiresponse data.

Implicit Multifunctional Nonlinear Regression Analysis

The least squares estimation of parameters in algebraically implicit, nonlinear, multiple response models having only one ezperimentally accessible response variable is treated within the context of a Gauss–Newton—Newton iteration. The algorithm, derived through application of the implicit function theorem to the model, is sufficiently general to cover Bayesian estimation of parameters for multiresponse data.

Design of experiments for parameter estimation in multiresponse situations

Abstract : A criterion is developed for the choice of n experimental runs in multiresponse situations, after N runs are already available. The criterion is applied to three examples involving two particular non-linear multiple response models and the results are discussed.

語認知に及ぼす情動効果 (V)

Our previous expriments showed that, when 72 words of similar structure and familiarity were each presented by tachistoscope, emotional stimuli given 5sec. prior to the presentation affected the word recognition.Now, in order to explore the above results more explicitly, it was first tested whether emotional stimuli will differentially affect word recognition or the results are attributed to the differeces in receptor organs, secondly whether chlorpromazine (a kind of tranquilizers) will repress the effect of emotion on word recognition.The following results were obtained:1) Difference in receptor organs did not affect word recongnition. Thus our previous results showed clearly that the differential emotions affect word recognition.2) Use of chlorpromazine repressed significantly the effect of emotion on word recognition.The obtained facts that emotional stimuli are effective on verbal response are not in accordance with Eriksen's theory on multiple concurrent response model. Besides, his theory does not seem to explain perceptual defense and sensitization phenomena. Thus, following Lazarus, we may assume that “subception” is a function of the autonomic system.