Automatic Item Generation Research Articles

Algorithmic Bias in BERT for Response Accuracy Prediction: A Case Study for Investigating Population Validity

AbstractPretrained large language models (LLMs) have gained popularity in recent years due to their high performance in various educational tasks such as learner modeling, automated scoring, automatic item generation, and prediction. Nevertheless, LLMs are black box approaches where models are less interpretable, and they may carry human biases and prejudices because historical human data have been used for pretraining these large‐scale models. For these reasons, the prediction tasks based on LLMs require scrutiny to ensure that the prediction models are fair and unbiased. In this study, we used BERT—a pretrained encoder‐only LLM for predicting response accuracy using action sequences extracted from the 2012 PIAAC assessment. We selected three countries (i.e., Finland, Slovakia, and the United States) representing different performance levels in the overall PIAAC assessment. We found promising results for predicting response accuracy using the fine‐tuned BERT model. Additionally, we examined algorithmic bias in the prediction models trained with different countries. We found differences in model performance, suggesting that some trained models are not free from bias, and thus the models are less generalizable across countries. Our results highlighted the importance of investigating algorithmic fairness in prediction models utilizing algorithmic systems to ensure models are bias‐free.

Application of the Contemporary Psychometrics for Assessing Economic Literacy

Currently, new skills and various types of "new literacies" relevant to the modern world are becoming issues of growing importance. One of them is economic literacy; however, there are only few assessment instruments that fulfil the academic requirements for its assessment among university students. One of such internationally established instruments is the Test of Understanding in College Economics (TUCE), which is a popular tool in empirical studies of economic literacy in many countries around the world. Despite its advantages, the currently available version of the TUCE designed for American colleges back in 2006, is prone to cheating and provides limited opportunities for formative feedback. The purpose of this paper is to present the Updated Test of Understanding in College Economics (U-TUCE). In developing the U-TUCE, we utilized the capabilities of contemporary psychometrics, which offer sufficient advances in overcoming all limitations of the original TUCE mentioned before. First, we present a revised theoretical framework of the U-TUCE, highlighting that the test measures different types of mastery of economic literacy. Second, we describe the approaches used for modifying the TUCE items and developing new items. A half of the original test items have been replaced or redesigned to reflect the economic context that has changed since 2006. Third, we utilize the logic of automatic item generation algorithms to increaseg the level of test protection against cheating. We made all changes in such a way as to maintain comparability with the previous versions of the TUCE test if necessary. Finally, the use of the Item Response Theory (IRT) is paired up with that of Cognitive Diagnostic Modeling (CDM) to ensure the quality of the U-TUCE and enhance its formative value. We show that IRT can be used to estimate the construct as a whole (which is of interest to researchers, administrators, and policy makers), while CDM provides information relating to each of the construct components, which are of interest to educational practitioners and students themselves. The results of the data analyses show that the test can be used for both purposes simultaneously.

Automatic generation of physics items with Large Language Models (LLMs)

High-quality items are essential for producing reliable and valid assessments, offering valuable insights for decision-making processes. As the demand for items with strong psychometric properties increases for both summative and formative assessments, automatic item generation (AIG) has gained prominence. Research highlights the potential of large language models (LLMs) in the AIG process, noting the positive impact of generative AI tools like ChatGPT on educational assessments, recognized for their ability to generate various item types across different languages and subjects. This study fills a research gap by exploring how AI-generated items in secondary/high school physics aligned with educational taxonomy. It utilizes Bloom's taxonomy, a well-known framework for designing and categorizing assessment items across various cognitive levels, from low to high. It focuses on a preliminary assessment of LLMs ability to generate physics items that match the Bloom’s taxonomy application level. Two leading LLMs, ChatGPT (GPT-4) and Gemini, were chosen for their strong performance in creating high-quality educational content. The research utilized various prompts to generate items at different cognitive levels based on Bloom's taxonomy. These items were assessed using multiple criteria: clarity, accuracy, absence of misleading content, appropriate complexity, correct language use, alignment with the intended level of Bloom's taxonomy, solvability, and assurance of a single correct answer. The findings indicated that both ChatGPT and Gemini were skilled at generating physics assessment items, though their effectiveness varied based on the prompting methods used. Instructional prompts, particularly, resulted in excellent outputs from both models, producing items that were clear, precise, and consistently aligned with the Application level of Bloom's taxonomy.

Let the algorithm speak: How to use neural networks for automatic item generation in psychological scale development.

Measurement is at the heart of scientific research. As many-perhaps most-psychological constructs cannot be directly observed, there is a steady demand for reliable self-report scales to assess latent constructs. However, scale development is a tedious process that requires researchers to produce good items in large quantities. In this tutorial, we introduce, explain, and apply the Psychometric Item Generator (PIG), an open-source, free-to-use, self-sufficient natural language processing algorithm that produces large-scale, human-like, customized text output within a few mouse clicks. The PIG is based on the GPT-2, a powerful generative language model, and runs on Google Colaboratory-an interactive virtual notebook environment that executes code on state-of-the-art virtual machines at no cost. Across two demonstrations and a preregistered five-pronged empirical validation with two Canadian samples (NSample 1 = 501, NSample 2 = 773), we show that the PIG is equally well-suited to generate large pools of face-valid items for novel constructs (i.e., wanderlust) and create parsimonious short scales of existing constructs (i.e., Big Five personality traits) that yield strong performances when tested in the wild and benchmarked against current gold standards for assessment. The PIG does not require any prior coding skills or access to computational resources and can easily be tailored to any desired context by simply switching out short linguistic prompts in a single line of code. In short, we present an effective, novel machine learning solution to an old psychological challenge. As such, the PIG will not require you to learn a new language-but instead, speak yours. (PsycInfo Database Record (c) 2023 APA, all rights reserved).

ChatGPT for generating multiple-choice questions: Evidence on the use of artificial intelligence in automatic item generation for a rational pharmacotherapy exam.

Artificial intelligence, specifically large language models such as ChatGPT, offers valuable potential benefits in question (item) writing. This study aimed to determine the feasibility of generating case-based multiple-choice questions using ChatGPT in terms of item difficulty and discrimination levels. This study involved 99 fourth-year medical students who participated in a rational pharmacotherapy clerkship carried out based-on the WHO 6-Step Model. In response to a prompt that we provided, ChatGPT generated ten case-based multiple-choice questions on hypertension. Following an expert panel, two of these multiple-choice questions were incorporated into a medical school exam without making any changes in the questions. Based on the administration of the test, we evaluated their psychometric properties, including item difficulty, item discrimination (point-biserial correlation), and functionality of the options. Both questions exhibited acceptable levels of point-biserial correlation, which is higher than the threshold of 0.30 (0.41 and 0.39). However, one question had three non-functional options (options chosen by fewer than 5% of the exam participants) while the other question had none. The findings showed that the questions can effectively differentiate between students who perform at high and low levels, which also point out the potential of ChatGPT as an artificial intelligence tool in test development. Future studies may use the prompt to generate items in order for enhancing the external validity of the results by gathering data from diverse institutions and settings.

Case-based MCQ generator: A custom ChatGPT based on published prompts in the literature for automatic item generation

What is the Educational Challenge? A fundamental challenge in medical education is creating high-quality, clinically relevant multiple-choice questions (MCQs). ChatGPT-based automatic item generation (AIG) methods need well-designed prompts. However, the use of these prompts is hindered by the time-consuming process of copying and pasting, a lack of know-how among medical teachers, and the generalist nature of standard ChatGPT, which often lacks the medical context. What are the Proposed Solutions? The Case-based MCQ Generator, a custom GPT, addresses these challenges. It has been trained by using GPT Builder, which is a platform designed by OpenAI for customizing ChatGPT to meet specific needs, in order to allow users to generate case-based MCQs. By using this free tool for those who have ChatGPT Plus subscription, health professions educators can easily select a prompt, input a learning objective or item-specific test point, and generate clinically relevant questions. What are the Potential Benefits to a Wider Global Audience? It enhances the efficiency of MCQ generation and ensures the generation of contextually relevant questions, surpassing the capabilities of standard ChatGPT. It streamlines the MCQ creation process by integrating prompts published in medical education literature, eliminating the need for manual prompt input. What are the Next Steps? Future development aims at sustainability and addressing ethical and accessibility issues. It requires regular updates, integration of new prompts from emerging health professions education literature, and a supportive digital ecosystem around the tool. Accessibility, especially for educators in low-resource countries, is vital, demanding alternative access models to overcome financial barriers.

Using OpenAI GPT to Generate Reading Comprehension Items

AbstractThe purpose of this study is to introduce and evaluate a method for generating reading comprehension items using template‐based automatic item generation. To begin, we describe a new model for generating reading comprehension items called the text analysis cognitive model assessing inferential skills across different reading passages. Next, the text analysis cognitive model is used to generate reading comprehension items where examinees are required to read a passage and identify the irrelevant sentence. The sentences for the generated passages were created using OpenAI GPT‐3.5. Finally, the quality of the generated items was evaluated. The generated items were reviewed by three subject‐matter experts. The generated items were also administered to a sample of 1,607 Grade‐8 students. The correct options for the generated items produced a similar level of difficulty and yielded strong discrimination power while the incorrect options served as effective distractors. Implications of augmented intelligence for item development are discussed.

Tıp Eğitiminde Otomatik Soru Üretme Yöntemi Kullanılarak Oluşturulan İlk Türkçe Çoktan Seçmeli Soruların Psikometrik Analizi

Aim: Automatic item generation is "a process of using models to generate items using computer technology". The use of automatic item generation typically involves one of three primary methods: syntax-based, semantic-based, and template-based. Non-template automatic item generation approaches leverage natural language processing techniques. A study showed the potential of using template-based automatic item generation to create high-quality multiple-choice questions for assessing clinical reasoning in Turkish, marking a first in the field. However, the findings of the study were based only on expert opinions, necessitating further research to examine the psychometric qualities of Turkish items. The aim of this study was to reveal psychometric characteristics of the first Turkish case-based multiple-choice questions generated by using automatic item generation in medical education. Methods: This was a psychometric study. Three Turkish case-based multiple-choice questions generated using template-based automatic item generation on essential hypertension were included in an exam that 281 fourth-year medical students participate in. This examination was carried out in-person in classroom settings under proctor supervision. Item difficulty and item discrimination (point-biserial correlation) were calculated, and non-functioning distractors were determined. Results: All three items had acceptable levels (higher than 0.20) of point-biserial correlation (p<0.001). The item difficulty levels indicated the presence of one easy, one moderate, and one difficult question. Each item had 2-3 non-functioning options among five options. All three items had acceptable levels (higher than 0.20) of point-biserial correlation (p<0.001). The item difficulty levels indicated the presence of one easy, one moderate, and one difficult question. Each item had 2-3 non-functioning options among five options. Conclusions: The results indicated that the items successfully discriminate between high and low performers, providing validity evidence on the quality of the questions in evaluating students' comprehension of the subject. Additionally, the findings suggest that it is feasible to create multiple-choice questions with different difficulty levels in Turkish using a single automatic item generation model. This study demonstrated for the first time that automatic generation of case-based multiple-choice questions in Turkish produces acceptable psychometric characteristics in an authentic assessment setting in medical education. The ability to automatically generate effective multiple-choice questions in Turkish holds promise for enhancing the efficiency of written assessment in Turkish medical education.

Automatic item generation for non-verbal reasoning items

The purpose of this study is to generate non-verbal items for a visual reasoning test using templated-based automatic item generation (AIG). The fundamental research method involved following the three stages of template-based AIG. An item from the 2016 4th-grade entrance exam of the Science and Art Center (known as BİLSEM) was chosen as the parent item. A cognitive model and an item model were developed for non-verbal reasoning. Then, the items were generated using computer algorithms. For the first item model, 112 items were generated, and for the second item model, 1728 items were produced. The items were evaluated based on subject matter experts (SMEs). The SMEs indicated that the items met the criteria of one right answer, single content and behavior, not trivial content, and homogeneous choices. Additionally, SMEs' opinions determined that the items have varying item difficulty. The results obtained demonstrate the feasibility of AIG for creating an extensive item repository consisting of non-verbal visual reasoning items.

A ChatGPT Prompt for Writing Case-Based Multiple-Choice Questions

The significant challenge faced by medical schools is the effortful process of writing a high quantity of high-quality case-based multiple-choice questions (MCQs) to assess the higher-order skills of medical students. The demand for a high volume of MCQs in education has led to the development of Automatic Item Generation (AIG), specifically template-based AIG, which involves creating cognitive and item models by subject matter experts to generate hundreds of MCQs at once using software. It demonstrated significant success in various languages and even being incorporated into national medical licensure exams. However, this method still heavily depends on the efforts of subject matter experts. This paper introduces a detailed ChatGPT prompt for quickly generating case-based MCQs and provides important research questions for future exploration into ChatGPT's potential in generating items, signaling the beginning of the artificial intelligence era in medical education, encouraging health professions education researchers to delve deeper into its potential.

Progress is impossible without change: implementing automatic item generation in medical knowledge progress testing

Progress tests (PT) are a popular type of longitudinal assessment used for evaluating clinical knowledge retention and long-life learning in health professions education. Most PTs consist of multiple-choice questions (MCQs) whose development is costly and time-consuming. Automatic Item Generation (AIG) generates test items through algorithms, promising to ease this burden. However, it remains unclear how AIG-items behave in formative assessment (FA) modalities such as PTs compared to manually written items. The purpose of this study was to compare the quality and validity of AIG-items versus manually written items. Responses to 126 (23 automatically generated) dichotomously scored single best-answer five-option MCQs retrieved from the 2021 University of Minho PT of medicine were analyzed. Procedures based on item response theory (IRT), dimensionality testing, item fit, reliability, differential item functioning (DIF) and distractor analysis were used. Qualitative assessment was conducted through expert review. Validity evidence of AIG-items was assessed by using hierarchical linear modeling (HLM). The PT proved to be a viable tool for assessing medical students cognitive competencies. AIG-items were parallel to manually written-items, presenting similar indices of difficulty and information. The proportion of functional distractors for both AIG and manually written items was similar. Evidence of validity for AIG-items was found while showing higher levels of item quality. AIG-items functioned as intended and were appropriate for evaluating medical students at various levels of the knowledge spectrum.

Automatic item generation for online measurement and evaluation: Turkish literature items

Developments in the field of education have significantly affected test development processes, and computer-based test applications have been started in many institutions. In our country, research on the application of measurement and evaluation tools in the computer environment for use with distance education is gaining momentum. A large pool of items is required for computer-based testing applications that provide significant advantages to practitioners and test takers. Preparing a large pool of items also requires more effort in terms of time, effort, and cost. To overcome this problem, automatic item generation has been widely used by bringing together item development subject matter experts and computer technology. In the present research, the steps for implementing automatic item generation are explained through an example. In the research, which was based on the fundamental research method, first a total of 2560 items were generated using computer technology and SMEs in field of Turkish literature. In the second stage, 60 randomly selected items were examined. As a result of the research, it was determined that a large item pool could be created to be used in online measurement and evaluation applications using automatic item generation.

Development of the automatic item generation system for the diagnosis of misconceptions about force and laws of motion

The understanding of force and laws of motion is a fundamental foundation for learning mechanics and understanding other complex physics-related subjects. Automatic item generation (AIG) is also suitable for generating items and able to reduce the chance of item exposure. We, thus, developed an AIG system for the diagnosis of misconceptions about force and laws of motion in order to create a large number of quality items that would be used to diagnose students’ misconceptions. AIG system that has been developed contains 18 item models; it can generate 320-3,200 test items. The system contains six menus, i.e., (1) users’ data, (2) item models, (3) item generation, (4) test generation, (5) the users’ guide, and (6) the system’s developer. Based on the examination of AIG system’s quality by experts on educational assessment and experts on information technology, AIG’s quality in terms of utility, feasibility, propriety, and accuracy is at the highest level. The system was improved using the two dimensions of users’ experiences with physics instructors, i.e., (1) pragmatic dimension and (2) hedonic dimension. This research offers an approach to developing AIG system that responds to users’ needs.

Automatic item generation: foundations and machine learning-based approaches for assessments

This mini review summarizes the current state of knowledge about automatic item generation in the context of educational assessment and discusses key points in the item generation pipeline. Assessment is critical in all learning systems and digitalized assessments have shown significant growth over the last decade. This leads to an urgent need to generate more items in a fast and efficient manner. Continuous improvements in computational power and advancements in methodological approaches, specifically in the field of natural language processing, provide new opportunities as well as new challenges in automatic generation of items for educational assessment. This mini review asserts the need for more work across a wide variety of areas for the scaled implementation of AIG.

Using Automatic Item Generation to Create Multiple-Choice Questions for Pharmacy Assessment

ObjectiveAutomatic item generation (AIG) is a new area of assessment research where a set of multiple-choice questions (MCQs) are created using models and computer technology. Although successfully demonstrated in medicine and dentistry, AIG has not been implemented in pharmacy. The objective was to implement AIG to create a set of MCQs appropriate for inclusion in a summative, high-stakes, pharmacy examination. MethodsA 3-step process, well evidenced in AIG research, was employed to create the pharmacy MCQs. The first step was developing a cognitive model based on content within the examination blueprint. Second, an item model was developed based on the cognitive model. A process of systematic distractor generation was also incorporated to optimize distractor plausibility. Third, we used computer technology to assemble a set of test items based on the cognitive and item models. A sample of generated items was assessed for quality against Gierl and Lai’s 8 guidelines of item quality. ResultsMore than 15,000 MCQs were generated to measure knowledge and skill of patient assessment and treatment of nausea and/or vomiting within the scope of clinical pharmacy. A sample of generated items satisfies the requirements of content-related validity and quality after substantive review. ConclusionThis research demonstrates the AIG process is a viable strategy for creating a test item bank to provide MCQs appropriate for inclusion in a pharmacy licensing examination.

Tıp Eğitiminde Klinik Akıl Yürütme Becerisinin Ölçümü için Türkçede İlk Otomatik Soru Üretimi

Aim: Writing high-quality items (questions) is a resource-intensive task. Particularly, the development of one context-rich multiple-choice question (MCQ) for assessing higher-order cognitive skills may cost hours of medical teachers. The aim of this study was to find out whether it is possible the use of Automatic Item Generation (AIG) in Turkish to generate case-based MCQs that assess clinical reasoning skills. 
 Methods: By following the template-based AIG method developed by Gierl et al., MCQs on hypertension were generated with the help of software after the development of a cognitive model and an item model. The cognitive model and the item model was developed by a medical doctor and a cardiologist by considering Turkish Hypertension Consensus Report. The software was built as a Python-based code intended for single use without a user interface. The items were recorded in a MySQL database. Of these questions, 10 questions were randomly chosen to be reviewed by three subject matter experts (cardiologists). The evaluation was based on the quality of the questions and whether the questions assess higher-order skills such as clinical reasoning rather than factual recall. 
 Results: In 1.73 seconds, 1600 MCQs on hypertension were generated. Although there were some minor revision suggestions in a few questions, each question was stated by all cardiologists as an acceptable item. The cardiologists also stated that the questions assess clinical reasoning skills rather than factual recall.
 Conclusions: This study demonstrated for the first time that AIG for assessing clinical reasoning skills in the context of medical education in Turkish is possible. This method of augmented intelligence to generate items can be used in Turkish as it has been used in other five languages. The use of this method could bring about more questions to assess clinical reasoning skills. It may also lead medical teachers to spend less amount of time and effort compared to traditional item writing.

A suggestive approach for assessing item quality, usability and validity of Automatic Item Generation

Automatic Item Generation (AIG) refers to the process of using cognitive models to generate test items using computer modules. It is a new but rapidly evolving research area where cognitive and psychometric theory are combined into digital framework. However, assessment of the item quality, usability and validity of AIG relative to traditional item development methods lacks clarification. This paper takes a top-down strong theory approach to evaluate AIG in medical education. Two studies were conducted: Study I—participants with different levels of clinical knowledge and item writing experience developed medical test items both manually and through AIG. Both item types were compared in terms of quality and usability (efficiency and learnability); Study II—Automatically generated items were included in a summative exam in the content area of surgery. A psychometric analysis based on Item Response Theory inspected the validity and quality of the AIG-items. Items generated by AIG presented quality, evidences of validity and were adequate for testing student’s knowledge. The time spent developing the contents for item generation (cognitive models) and the number of items generated did not vary considering the participants' item writing experience or clinical knowledge. AIG produces numerous high-quality items in a fast, economical and easy to learn process, even for inexperienced and without clinical training item writers. Medical schools may benefit from a substantial improvement in cost-efficiency in developing test items by using AIG. Item writing flaws can be significantly reduced thanks to the application of AIG's models, thus generating test items capable of accurately gauging students' knowledge.

The Impact of Item Model Parameter Variations on Person Parameter Estimation in Computerized Adaptive Testing With Automatically Generated Items.

Sibling items developed through automatic item generation share similar but not identical psychometric properties. However, considering sibling item variations may bring huge computation difficulties and little improvement on scoring. Assuming identical characteristics among siblings, this study explores the impact of item model parameter variations (i.e., within-family variation between siblings) on person parameter estimation in linear tests and Computerized Adaptive Testing (CAT). Specifically, we explore (1) what if small/medium/large within-family variance is ignored, (2) if the effect of larger within-model variance can be compensated by greater test length, (3) if the item model pool properties affect the impact of within-family variance on scoring, and (4) if the issues in (1) and (2) are different in linear vs. adaptive testing. Related sibling model is used for data generation and identical sibling model is assumed for scoring. Manipulated factors include test length, the size of within-model variation, and item model pool characteristics. Results show that as within-family variance increases, the standard error of scores remains at similar levels. For correlations between true and estimated score and RMSE, the effect of the larger within-model variance was compensated by test length. For bias, scores are biased towards the center, and bias was not compensated by test length. Despite the within-family variation is random in current simulations, to yield less biased ability estimates, the item model pool should provide balanced opportunities such that "fake-easy" and "fake-difficult" item instances cancel their effects. The results of CAT are similar to that of linear tests, except for higher efficiency.

Supplemental Material for Let the Algorithm Speak: How to Use Neural Networks for Automatic Item Generation in Psychological Scale Development

Supplemental Material for Let the Algorithm Speak: How to Use Neural Networks for Automatic Item Generation in Psychological Scale Development

An Evaluation of Automatic Item Generation: A Case Study of Weak Theory Approach

AbstractThis case study applied the weak theory of Automatic Item Generation (AIG) to generate isomorphic item instances (i.e., unique but psychometrically equivalent items) for a large‐scale assessment. Three representative instances were selected from each item template (i.e., model) and pilot‐tested. In addition, a new analytical framework, differential child item functioning (DCIF) analysis, based on the existing differential item functioning statistics, was applied to evaluate the psychometric equivalency of item instances within each template. The results showed that, out of 23 templates, nine successfully generated isomorphic instances, five required minor revisions to make them isomorphic, and the remaining templates required major modifications. The results and insights obtained from the AIG template development procedure may help item writers and psychometricians effectively develop and manage the templates that generate isomorphic instances.