to fully engage with test content, leading to disengaged behaviors such as rapid guessing or skipping

Overview

This source page is a mechanical bulk-ingest record for a PDF in the methylmercury infant-formula research pull. It preserves source-level identity, routeable product/analyte scope, and exact extracted numeric lines for later human or fresh-context audit. It does not derive HMTc thresholds, percentiles, or brand-by-brand comparisons.

Key numbers

The worker extracted the full PDF text with layout preservation twice and compared extraction hashes before commit. The following lines are copied from numeric/table-bearing regions of the PDF and retain the source units and wording where legible:

• ISSN: 1309 – 6575
• Low-stakes assessments, commonly used in higher education and large-scale K-12 testing, face unique challenges
• Low-stakes assessments are widely used in higher education and large-scale K-12 testing to monitor
• items altogether (Wise, 2017). For example, data from the 2018 Programme for International Student
• Assessment (PISA) indicate that over 50% of examinees were disengaged on about one in every ten
• threat to the validity of score-based inferences (Rios, 2021). When examinees do not exert sufficient
• Association, et al., 2014).
• motivational dynamic is well explained by expectancy-value theory (EVT; Wigfield & Eccles, 1992),
• • Assoc. Prof. Dr., James Madison University, Harrisonburg-US, e-mail: leventbc@jmu.edu, ORCID ID: 0000-0002-6480-
• ** PhD Student., James Madison University, Harrisonburg-US, e-mail: hunsbeja@jmu.edu, ORCID ID: 0000-0003-1680-2545
• Leventhal, B. C. & Hunsberger, J. (2025). Branching out: How the IRTree model can root out disengagement in a variety
• of low-stakes contexts. Journal of Measurement and Evaluation in Education and Psychology, 16(3), 179-202.
• beneficial (low value), leading to reduced effort and engagement (Flake et al., 2015).
• Molnár, 2024; Dibek, 2020; Goldhammer et al., 2017). Its effects are far-reaching as disengagement can
• distort individual scores (Wise et al., 2020), affect aggregated score interpretations (Wise, 2020), bias
• parameter estimates (Rios et al., 2017), and compromise the reliability of test scores (Kuang & Sahin,
• 2023). Given these implications, identifying and accounting for disengagement is essential for
• have since shifted focus to the item level (Alahmadi & DeMars, 2024). Item-level disengagement has
• patterns in test-taking persistence (Nagy et al., 2023), behavioral indicators such as confidence in
• responses (Maqsood et al., 2019), and advanced technologies such as blink tracking (Hollander &
• Huette, 2022), and the number of clicks during an assessment (Wang et al., 2019).
• 2023). Among these, the Effort-Moderated (EM) IRT model (Wise & DeMars, 2006) has been widely
• cognitively engage with the item; Wise & Kong, 2005) as missing at random (MAR) conditional on the
• ability estimates of non-rapid guessed responses (Alahmadi & DeMars, 2024; Wise & DeMars, 2006).
• biased estimates when such a relationship exists (Rios & Soland, 2021).
• (Ulitzsch et al., 2020). While these models offer valuable insights, many rely on strong assumptions,
• between disengagement and the latent ability of interest (Deribo et al., 2021; Liu et al., 2019). However,
• et al., 2021), which may not capture all disengaged behaviors (e.g., Schaefer & Finney, 2025).
• smaller sample sizes (Rios & Deng, 2023).
• by Leventhal and Pastor (2024). Although originally developed using a rapid guessing
• Leventhal and Pastor (2024) demonstrated the model’s potential through a simulation study under
• ISSN: 1309 – 6575 Eğitimde ve Psikolojide Ölçme ve Değerlendirme Dergisi
• Journal of Measurement and Evaluation in Education and Psychology 180
• 1. Under what situational and statistical conditions of low-stakes testing does the model
• 2. Under what situational and statistical conditions of low-stakes testing does the model
• parameters are jointly estimated using the methods presented by Leventhal and Pastor (2024), and that
• 2005). This lack of motivation can lead to disengaged behaviors that compromise the interpretability of
• scores (e.g., Rios et al., 2017). While disengagement was once conceptualized primarily at the test level,
• can be more precisely observed and modeled (Alahmadi & DeMars, 2024).
• despite known issues with misclassification (e.g., Wise, 2017). However, disengagement can also
• (Maqsood et al., 2019). In digital assessments, process data, such as the frequency and nature of item
• interactions, can provide additional insight into examinee engagement (Kara, 2025). Recently, biometric

Methods (brief)

• including variations of sample size, test length, prior distributions, and the correlation between latent traits. Using
• with factors like test length and prior distributions moderately influencing accuracy. While larger sample sizes and
• smaller sample sizes (Rios & Deng, 2023).
• items), 2) relationship between traits (𝜌𝜃𝑗,𝑇𝑂𝐼 ,𝜃𝑗,𝐷𝐼𝑆𝐸𝑁𝐺 = 0, -0.4, -0.8), 3) sample size (N = 500, 2000),
• Sample Size. To account for the variability in administration sizes for low-stakes testing, we
• selected sample sizes that would generalize the model across both large and small contexts. Specifically,
• we included a small sample size to address concerns about parameter estimate stability and to
• administer low-stakes assessments to a sample of students for accreditation purposes (Pastor et al.,
• Furthermore, it is critical to evaluate the model’s performance in conditions of low sample sizes because
• sample sizes, such as 500 (Rios & Deng, 2023).
• information from the collected data, which is influenced by sample size, while the prior distribution is
• essential to investigate prior distributions in Bayesian simulation studies whenever sample size is
• when sample sizes are small, but this influence diminishes as sample sizes increase.
• Trait Parameters. True person parameters, 𝜃𝑗,𝑇𝑂𝐼 and 𝜃𝑗,𝐷𝐼𝑆𝐸𝑁𝐺 , were randomly sampled from a
• Item Parameters. Item parameters were randomly sampled from the following distributions:
• randomly sampled a single 𝑎1 parameter for each replication, representing the common relationship
• 𝑟 Sample size Test length Mean Percentage of Standard deviation Percentage
• Figures 3 and 4. The bias in the 𝑎1 item parameter is influenced by the interaction between sample size

Implications

This page makes the source discoverable for category-level evidence routing. Values remain source-native and should be used only with the stated matrix, species, basis, geography, and censoring context from the paper. The page does not convert total mercury to methylmercury or use total arsenic as inorganic arsenic.

Wiki pages this source may touch

• Fish — marine, non-predatory (sardines, anchovies, salmon, cod)
• Lead
• Aluminum

Verification notes

• Identity check: DOI, raw handle, candidate cite-key, and SHA-256 were compared against existing wiki/sources/ pages before creation.
• Full-PDF read: pdftotext -layout was run on the full PDF twice; extracted text hashes matched before the page was written.
• Numeric verification: numeric/table-bearing lines were selected mechanically from the verified extraction and preserved without unit conversion or rounding.
• Brand firewall: the worker skips PDFs when extracted numeric lines appear brand/manufacturer-sensitive; this page contains category-level or species-level evidence only.
• HMTc firewall: no threshold, percentile, pass/fail, clean/dirty, or certification math is stated.

Update history

The five most recent substantive edits to this page, classified major (evidence or structure moved), correction (a published value or statement was wrong and has been fixed), or minor (narrative rewritten without changing the underlying evidence). Each description is derived from what the edit did to this page; the linked commit is the authoritative record, routine regeneration passes are excluded, and the full version history lives in git. When DOI minting comes online (see schema docs), each entry below will also link to a version-pinned DataCite DOI.