Pure Guessing Research Articles

A Bayesian network for modelling the Lady tasting tea experiment.

A cup of tea can be made in one of the two ways: the milk or the tea infusion was first added to the cup. The Lady Tasting Tea experiment consists in mixing eight cups of tea, four in one way and four in the other, and presenting them to the Lady for judgment in a random order. This short article presents a Bayesian Network (BN) for modelling the Lady Tasting Tea experiment that provides a comprehensive perspective in inferential analysis of all the data samples possibly generated from the experiment. More specifically, with respect to a prior distribution of three possible levels (pure guessing, 75% sure, and 100% sure) of the Lady's ability in correctly deciding how a served cup of tea has been made, the proposed BN model enables us to calculate the posterior probabilities of any judgment outcomes possibly made by the Lady.

Encoding of continuous perceptual choices in human early visual cortex.

Research on the neural mechanisms of perceptual decision-making has typically focused on simple categorical choices, say between two alternative motion directions. Studies on such discrete alternatives have often suggested that choices are encoded either in a motor-based or in an abstract, categorical format in regions beyond sensory cortex. In this study, we used motion stimuli that could vary anywhere between 0° and 360° to assess how the brain encodes choices for features that span the full sensory continuum. We employed a combination of neuroimaging and encoding models based on Gaussian process regression to assess how either stimuli or choices were encoded in brain responses. We found that single-voxel tuning patterns could be used to reconstruct the trial-by-trial physical direction of motion as well as the participants' continuous choices. Importantly, these continuous choice signals were primarily observed in early visual areas. The tuning properties in this region generalized between choice encoding and stimulus encoding, even for reports that reflected pure guessing. We found only little information related to the decision outcome in regions beyond visual cortex, such as parietal cortex, possibly because our task did not involve differential motor preparation. This could suggest that decisions for continuous stimuli take can place already in sensory brain regions, potentially using similar mechanisms to the sensory recruitment in visual working memory.

Metacognitive Monitoring and Mathematical Abilities: Cognitive Diagnostic Model and Signal Detection Theory Approach

Besides various in-class assessments, there exist some standardized assessment tools that are administered in several countries, such as PISA (Programme for International Student Assessment) and TIMMS (Trends in International Mathematics and Science Study). The questions’ contents, type of responding, grading, and the analyses in these large-scale tests have been diversified in years. In this study, it was aimed to identify the abilities that are measured at PISA mathematics test in a single testing procedure and by utilizing the methods of analyses of Cognitive Diagnostic Model (CDM) as well as Signal Detection Theory (SDT), which have not been used so far in the assessment of these abilities. Therefore, a randomly selected sample of 6th-grade students ( N =230) in Izmir was tested with a PISA-equivalent 12-item mathematics test, where the items are graded dichotomously (correct vs. incorrect). CDM estimates were calculated by using the Deterministic Input Noisy Output and Gate (DINA) Model. The participants were asked to report whether they thought they could solve the question correctly, guess even if they thought they could not solve the question, and then, rate their confidence levels on the correctness of their answers in turn so as to allow us to measure their “metacognitive monitoring performance” with the SDT method, which refers to the ability to differentiate correct and incorrect responses. In short, a better metacognitive monitoring performance was obtained by measuring how well once could differentiate their correct and incorrect responses with the observation of they prefer reporting and then giving high confidence levels to the actually correct responses and prefer passing to give an answer yet rate lower confidence levels to the actually incorrect responses given as pure guesses. The results showed that CDM fits well to the assessment of PISA test and those who were better at the ability of “reasoning and developing strategies” in particular among four possible abilities detected with CDM (“representing and communicating”, “mathematization”, “reasoning and developing strategies”, “using symbolic and technical language”) had also better metacognitive monitoring performance. The present study, therefore, contributes to the research that investigates what features the ability of better differentiating correct and incorrect responses are actually linked. Based on the results, it is suggested that a better metacognitive monitoring ability is linked to having a better ability of “reasoning and developing strategies” in particular. Additionally, it is suggested that measuring metacognitive monitoring performance at PISA -or even any other possible tests- with the SDT calculation method, that has a relatively straightforward testing procedure, may yield various estimates for the students’ abilities measured at the test as well as their related higher-order abilities.

Quantum imaging using relativistic detectors

Imaging in quantum optics (QO) is usually formulated in the languages of quantum mechanics and Fourier optics. While relatively advanced fields, notions such as different reference frames and the degradation of entanglement due to acceleration do not usually feature. Here we propose the idea of using so-called Unruh-DeWitt (UDW) detectors to model the imaging process in QO. In particular, we first present a quantum field theory version of a state describing Spontaneous Parametric Down Conversion (SPDC), one of the principal processes employed to create entangled photons in the laboratory. This state, coupled to UDW detectors, is used to investigate a single-pixel ghost image under both inertial and non-inertial settings, and a two-pixel image under inertial conditions. The reconstructed images obtained for various possible inputs can be distinguished better than a pure guess, hence the formalism can be used to describe imaging between non-inertial frames. We briefly consider the origin of the correlations between the UDW detectors, which don't appear to arise from the usual notion of entanglement. Finally, we find that the contrast between the possible outcomes in the single-pixel case follows a curious coupling time dependent behaviour.

Anesthesiologists' Overconfidence in Their Perceived Knowledge of Neuromuscular Monitoring and Its Relevance to All Aspects of Medical Practice: An International Survey.

In patients who receive a nondepolarizing neuromuscular blocking drug (NMBD) during anesthesia, undetected postoperative residual neuromuscular block is a common occurrence that carries a risk of potentially serious adverse events, particularly postoperative pulmonary complications. There is abundant evidence that residual block can be prevented when real-time (quantitative) neuromuscular monitoring with measurement of the train-of-four ratio is used to guide NMBD administration and reversal. Nevertheless, a significant percentage of anesthesiologists fail to use quantitative devices or even conventional peripheral nerve stimulators routinely. Our hypothesis was that a contributing factor to the nonutilization of neuromuscular monitoring was anesthesiologists' overconfidence in their knowledge and ability to manage the use of NMBDs without such guidance. We conducted an Internet-based multilingual survey among anesthesiologists worldwide. We asked respondents to answer 9 true/false questions related to the use of neuromuscular blocking drugs. Participants were also asked to rate their confidence in the accuracy of each of their answers on a scale of 50% (pure guess) to 100% (certain of answer). Two thousand five hundred sixty persons accessed the website; of these, 1629 anesthesiologists from 80 countries completed the 9-question survey. The respondents correctly answered only 57% of the questions. In contrast, the mean confidence exhibited by the respondents was 84%, which was significantly greater than their accuracy. Of the 1629 respondents, 1496 (92%) were overconfident. The anesthesiologists surveyed expressed overconfidence in their knowledge and ability to manage the use of NMBDs. This overconfidence may be partially responsible for the failure to adopt routine perioperative neuromuscular monitoring. When clinicians are highly confident in their knowledge about a procedure, they are less likely to modify their clinical practice or seek further guidance on its use.

Using machine learning techniques to generate laboratory diagnostic pathways—a case study

Background: Diagnostic pathways are based on expert rules (“if…then…else”), which can be visualized as decision trees. Machine learning algorithms may be used to validate existing or to suggest potential new decision trees. Methods: We present and compare two machine learning algorithms, which automatically generate decision trees from laboratory data. The underlying functions (rpart and ctree) are included in the free statistical software environment R (www.r-project.org). Results: Using input data from a published study on hepatitis C patients, we demonstrate that both algorithms are easy to apply and produce plausible decision trees. These algorithms confirm common knowledge about the power of laboratory testing to detect liver fibrosis and cirrhosis. For example, we show that conventional measurands such as choline esterase or albumin can separate cases with no or slight fibrosis from those with severe fibrosis and cirrhosis, while intermediate stages are often misclassified. Addition of the newer ELF score to the test panel makes decision trees simpler and may improve the quality of classification. Validation of the automatic decision trees with the leave-one-out method shows that the accuracy of all models created by machine learning is significantly higher than pure guessing. However, small changes of the input data result in markedly deviant decision trees, which need to be validated further with respect to clinical plausibility. Conclusions: The machine learning algorithms presented here can support but do not replace the medical expert when designing decision trees for diagnostic pathways.

Promoting collaborative learning through regulation of guessing in clickers

Collaborative learning is a promising avenue in education research. Learning from others and with others can foster deeper learning at a multiple-choice assignment, but it is hard to control the level of students' pure guessing. This paper addresses the problem of promoting collaborative learning through regulation of guessing when students use clickers to answer multiple-choice questions of various levels of difficulty. The study is aimed at identifying how the difficulty of the task and students' levels of knowledge influence on the degree of partial guessing. To answer this research question, we developed two research models and validated them by testing 84 students with regard to the students' level of knowledge and the penalty announcement. The findings of this research reveal that: a) the announcement of penalty has a negative effect on promoting collaborative learning even if it leads to reducing pure guesses in test results; b) questions that require higher-order thinking skills promote collaborative learning to a greater extent; c) creating mixed level groups of students seems advisable to enhance learning from collaboration and, thus, to decrease the degree of pure guessing.

Discrimination accuracy between real and sham press needles in the hands.

To evaluate the blinding effectiveness of a modified blunt sham press needle on the basis of the ability of subjects to discriminate between real and sham acupuncture needles compared with their discrimination ability based on pure guessing, and to define differences between senses (touch and vision) in the rates of correctly identified needles. Sixty-three healthy students and staff members were recruited through convenience sampling. First, real or sham acupuncture was randomly administered to the left LI4 point while subjects could not observe the needle tip. A real or sham needle tip was then shown to the subjects. Finally, a random combination of real or sham acupuncture needles were randomly administered to the left and right LI4 points, this time with the subjects observing the procedure. In all conditions the subjects gave their judgement as Yes or No in response to questions asking them to identify the needle type. The proportion of correct judgements (P(C)) was computed for the last part of the trial in left and right LI4 points, and the rates of correctly identified needles for each trial were obtained. The subjects' accuracy of discrimination between the real and sham acupuncture needles in left and right LI4 points was not significantly different from that based on pure guess (P(C)=0.50 (chance level)), which indicates complete inability to discriminate between needles. The rates of correctly identified needles using touch, vision and a combination of both senses were not significantly different (p=0.807). The findings from this study show that this sham acupuncture device successfully blinded subjects to real and sham press needles, suggesting that it is effective for subject blinding in studies on acupuncture using press needles, and facilitating evaluation of the effects of acupuncture in placebo-controlled trials using a rigorous scientific research methodology.

Quantum State Discrimination with Prior Knowledge in Noisy Quantum Channels

Discrimination between two states of a qubit is investigated, which is performed under the influence of noisy quantum channels. When prior knowledge about on the quantum states is available, the detection probability of quantum measurement is compared with that of pure guessing during the irreversible time evolution. In the case of a Markovian channel, the superiority of quantum measurement to pure guessing is lost at finite time which is determined by the prior probability and the fidelity of the quantum states. For a non-Markovian channel, however, it is possible to recover the superiority of quantum measurement even if it is lost. The effect of a system-environment initial correlation on the quantum state discrimination is also investigated.

Order effects of ballot position without information-induced confirmatory bias

Candidate list positions have been shown to influence decision making when voters have limited candidate information. Here, a primacy advantage is observed owing to a greater number of positive arguments generated for early list candidates. The present study examined list position effects when an absence of information precludes such a confirmatory bias heuristic. We report the first large-scale low-information experimental election where candidate position is fully counterbalanced. Seven hundred and twenty participants voted in a mock election where the position of six fictitious and meaningless parties was counterbalanced across the electorate. Analysis by position revealed that significantly fewer votes were allocated to the terminal parties, and found preliminary evidence of an alphabetical bias. This positional bias was not present in a methodological replication using six genuine UK political parties. This suggests that in situations of pure guessing, the heuristic shifts from the primacy benefiting confirmatory bias to an alternative heuristic that prejudices the first and last parties. These findings suggest that while the UK general electoral process may be largely immune to positional prejudice, English local elections (in which there can be multiple candidates from the same party) and multiple preference ranking systems (Scottish Local Government and London Mayoral Elections) could be susceptible to both positional and alphabetical biases.

Discrimination Accuracy between Real and Sham Needles using the Park Sham Device in the Upper and Lower Limbs

To evaluate the blinding effectiveness of the Park sham acupuncture device using participants' ability to discriminate between the real and sham acupuncture needles in the (1) upper limb (TE points) compared with pure guessing and (2) lower limb (BL points) compared with pure guessing. 20 healthy acupuncture-naïve university students and staff were recruited through convenience sampling. Participants made Yes-No judgements on whether the real or sham needle was administered to four TE acupoints on the dominant upper limb, and four acupoints along the BL meridian on the dominant lower limb. The proportion of correct judgements, P(C), for each participant was calculated to indicate the discrimination accuracy of participants in distinguishing between the real and sham needles. Separate P(C) were computed for the upper limb acupoints and lower limb acupoints. The data were also pooled to calculate a P(C) for a combination of both body regions. The participants' discrimination accuracy between the real and sham needles was not statistically significant from P(C)=0.5 (chance level) for the lower limb alone and combined body regions' acupoint comparisons (lower limb: t(19)=0.00, unadjusted p=1.00; combined: t(19)=1.75, unadjusted p=0.10). However, the participants' discrimination accuracy was statistically significant from P(C)=0.5 for the upper limb acupoints alone comparison (t(19)=2.36, unadjusted p=0.03). This study showed that the Park sham device is more likely to blind participants in differentiating between the real and sham needles in the lower limb (BL meridian) acupoints than in the upper limb (TE meridian). However, the participants' ability to differentiate between the needle types for the upper limb acupoints was significantly different from chance levels.

Estimating the growth of internal evidence guiding perceptual decisions

Perceptual decision-making is thought to involve a gradual accrual of noisy evidence. Temporal integration of the evidence reduces the relative contribution of dynamic internal noise to the decision variable, thereby boosting its signal-to-noise ratio. We aimed to estimate the internal evidence guiding perceptual decisions over time, using a novel combination of external noise and the response signal methods. Observers performed orientation discrimination of patterns presented in external noise. We varied the contrast of the patterns and the delay at which observers were forced to signal their decision. Each test stimulus (patterns and noise sample) was presented twice. Across two experiments we varied the availability of the visual stimulus for processing. Observer model analyses of discrimination accuracy and response consistency to two passes of the same stimulus, suggested that there was very little growth in the internal evidence. The improvement in accuracy over time characterised by the speed-accuracy trade-off function predominantly reflected a decreasing proportion of non-visual decisions, or pure guesses. There was no advantage to having the visual patterns visible for longer than 80 ms, indicating that only the visual information in a short window after display onset was used to drive the decisions. The remarkable constancy of the internal evidence over time suggests that temporal integration of the sensory information was very limited. Alternatively, more extended integration of the evidence from memory could have taken place, provided that the dominant source of internal noise limiting performance occurs between-trials, which cannot be reduced by prolonged evidence integration.

Children's and Adults’ Judgments of the Certainty of Deductive Inferences, Inductive Inferences, and Guesses

ABSTRACT Children and adults rated their own certainty following inductive inferences, deductive inferences, and guesses. Beginning in kindergarten, participants rated deductions as more certain than weak inductions or guesses. Deductions were rated as more certain than strong inductions beginning in Grade 3, and fourth-grade children and adults differentiated strong inductions, weak inductions, and informed guesses from pure guesses. By Grade 3, participants also gave different types of explanations for their deductions and inductions. These results are discussed in relation to children's concepts of cognitive processes, logical reasoning, and epistemological development.

The distinction between intuition and guessing in the SRT task generation: A reply to Norman and Price

We (Fu, Dienes, & Fu, 2010) investigated the extent to which people could generate sequences of responses based on knowledge acquired from the Serial Reaction Time task, depending on whether it felt subjectively like the response was based on pure guessing, intuition, conscious rules or memories. Norman and Price (2010) argued that in the context of our task, intuition responses were the same as guessing responses. In reply, we argue that not only do subjects apparently claim to be experiencing different phenomenologies when saying intuition versus guess, but also intuition and guess responses are associated with different behaviors. We found that people could control the knowledge when generating responses felt to be based on intuition but not those felt to be pure guessing. We present further evidence here that triplets associated with intuition but not guessing were also processed fluently.

Children’s and Adults’ Evaluation of Their Own Inductive Inferences, Deductive Inferences, and Guesses

Adults’ and kindergarten through fourth-grade children’s evaluations and explanations of inductive inferences, deductive inferences, and guesses were assessed. Beginning in kindergarten, participants rated deductions as more certain than weak inductions or guesses. Beginning in third grade, deductions were rated as more certain than strong inductions. Adults and fourth-grade children differentiated strong inductions, weak inductions, and informed guesses from pure guesses. Adults’ and fourth-grade children’s evaluations of inferences were influenced by the sample of evidence provided and by population base-rate information. By third grade, participants also gave different types of explanations for their deductions and inductions. These results are discussed in relation to children’s concepts of cognitive processes, logical reasoning, and epistemological development. Learning to recognize and evaluate different patterns of reasoning is important for the development of social understanding, logical and scientific reasoning, critical thinking, and epistemological thought. Realizing that another person has made an inference can help children to assess the other person’s knowledge and beliefs, explain the other person’s actions, and respond appropriately during social interactions. Awareness of inferential processes also contributes to the development of children’s own reasoning and children’s evaluation of arguments. Explicit awareness of logic may facilitate sophisticated levels of reasoning (Moshman, 1990). Scientific

Chance guessing in a forced-choice recognition task and the detection of malingering.

Guessing occurs on forced-choice (FC) tests for which responses cannot be based on relevant knowledge. Its importance is in inverse relation to the level of knowledge being measured, so that it becomes an increasing component of test scores as the level of knowledge decreases. It is also used as a benchmark to detect simulated impairment. This investigation examined the role of guessing in a 2-alternative FC face recognition test. Chance groups shown only the test items were asked to guess which were the targets, to measure the variation in scores likely to be found with pure guessing. Controls performed normally, and two simulation groups tried to fake amnesia. Results suggested that simple guessing in the chance group produced variable scores that overlapped both low genuine and "malingering" performance. Low control scores were hidden by the guessing "chance bonus," which ameliorated the apparent level of decline in memory. Simulators told of their role at the outset (before presentation of the target items) produced more convincing "amnesic" scores than those told only just before the test, who produced the expected below-chance level of score. It is suggested that guessing variability should be taken into account in interpreting FC scores.

Parallel processing of stimulus features during RSVP: Evidence from the second response

When observers are asked to report a feature of a single target displayed in rapid serial visual presentation, they frequently make errors. Most frequently, a feature from the to-be-reported dimension pertaining to a stimulus presented near the target is reported. These migrations are so-called illusory conjunctions in the time domain. From parallel models, it is proposed that during the presentation of the series, the response features of the stimuli are extracted. If a high rate of presentation does not enable proper binding processes, the system could base its response on sophisticated guessing on the basis of the relative levels of activation of the available response features. However, the multiple extractions assumption has not received direct empirical support. We report two experiments in which the observers had to report their first and second response candidates. This double response paradigm makes it possible to test the assumption that more than one response feature is available for making a response. Furthermore, the application of the constant ratio rule (following Botella, Barriopedro, & Suero's, 2001, model) to the first responses allows predictions for the ratios between choices of the items for the second responses. The correlations between the observed and the predicted response proportions were .887 and .956 in the two experiments. This high predictive capacity indicates, first, that the observers have more than one response available, among which to choose, and second, that the choice among responses is determined largely in the same way for both, first and second responses. Nevertheless, the small errors in prediction are further reduced if a proportion of pure guesses is assumed in the second responses. These are probably due to memory losses, misidentifications of the features, and other factors impairing performance in second responses in comparison with first responses.

Proper and improper use and interpretation of Beta-binomial models in the analysis of replicated difference and preference tests

Abstract For the analysis of replicated difference and preference tests, the Beta-binomial (BB) model is frequently used in order to account for possible heterogeneity of the assessors. For two-sided tests, this approach can be quite useful. We argue, however, that a chance-corrected approach should be preferred in most situations employing a one-sided test like, e.g., the triangle test. In randomized one-sided tests, the success probability is bounded below by p 0 , the success probability from pure guessing. A further development of the BB model accounting for this constraint is supported. Furthermore, we show that heterogeneity found from either a standard BB or a chance-corrected BB model proves that the products are different, i.e., testing for heterogeneity is one way of testing for product differences. The parameters of any BB approach have therefore to be considered in parallel. A simulation study indicates a better performance of the chance-corrected BB model as compared to the standard model.

Termination of a visual search with large display size effects

The ability to locate an object in the visual field is a collaboration of at least three intermingled processes: scanning multiple locations, recognizing the object sought (the target), and ending the search in cases when the target is not found. In this paper, we focus on the termination rule. Using distribution analyses, it is possible to assess the probability of termination conditional on the number of locations examined. The results show that on some trials without target, the participants carried out more comparisons than there are objects in the display; in other conditions, they carried out fewer comparisons than objects. Because there were very few errors, the premature stops were not pure guesses. We present models to account for these findings. The distributions of terminations help determine the slopes of the functions relating response time to set size.

PRIORITY AS DOMINANCE IN DERIVED MEASUREMENT: INVARIANCE OF THE PRINCIPAL EIGENVECTOR

Ranking is a process of prioritization. Priorities, as measurement rather than pure guessing, can be derived from paired comparison judgments that generalize on ratios of actual measurements. Paired comparisons involve the selection of the smaller of the two objects being compared as the unit and estimating how many multiples of that unit the larger object is with respect to an attribute they share. In this paper, it is shown how priorities are derived as the principal right eigenvector of a pairwise comparison matrix and several examples are given to illustrate how the process works.