Pre-testing is an educational strategy that involves giving students ungraded tests before they’ve had a chance to learn material. Though it can sometimes feel counterintuitive, it has a continually...

Search, browse and learn about the Federal Register. Federal Register 2.0 is the unofficial daily publication for rules, proposed rules, and notices of Federal agencies and organizations, as well as e...

The Department of Consumer Affairs’ Office of Professional Examination Services described its exam-development processes for BAR programs: occupational analyses, SME involvement, pretesting, criterion-referenced pass-score methods (modified Angoff), and adoption of item-response theory for inspector and repair exams.

Python 3.15's new built-in types, a new way to delay imports, new comprehension syntax, new tools for finding slow code, and smaller standard library changes. 14 questions, answers revealed as you go…

Predicts Rasch item difficulty from item features (text embeddings from any model, template family, content metadata) by ridge regression (Hoerl and Kennard, 1970, &lt;<a href="https://doi.org/10.1080%2F00401706.1970.10488634" target="_top">doi:10.1080/00401706.1970.10488634</a>&gt;) with honest out-of-sample uncertainty, uses the predictions as robust informative priors to calibrate new items from small pretest samples, in the spirit of using collateral information in calibration (Mislevy, Sheehan and Wingersky, 1993, &lt;<a href="https://doi.org/10.1111%2Fj.1745-3984.1993.tb00422.x" target="_top">doi:10.1111/j.1745-3984.1993.tb00422.x</a>&gt;), plans the pretest sample size needed for a target precision, and diagnoses item families whose predictions cannot be trusted.

Patrick Sturgis (London School of Economics) presented a survey methods webinar about using generative AI in surveys on 15 September 2026. Large language models have the potential to transform many…

Provides quantitative tools for substantive and content-oriented scale pretesting. Implements item-sort indices from Anderson and Gerbing (1991) &lt;<a href="https://doi.org/10.1037%2F0021-9010.76.5.732" target="_top">doi:10.1037/0021-9010.76.5.732</a>&gt;, exact item-sort inference following Howard and Melloy (2016) &lt;<a href="https://doi.org/10.1007%2Fs10869-015-9404-y" target="_top">doi:10.1007/s10869-015-9404-y</a>&gt;, empirical interpretation benchmarks from Colquitt et al. (2019) &lt;<a href="https://doi.org/10.1037%2Fapl0000406" target="_top">doi:10.1037/apl0000406</a>&gt;, and the construct-rating procedure of Hinkin and Tracey (1999) &lt;<a href="https://doi.org/10.1177%2F109442819922004" target="_top">doi:10.1177/109442819922004</a>&gt; with HTC/HTD indices and repeated-measures item screening. The expert-panel workflow combines Aiken's V with score confidence intervals, Lawshe content validity ratios with exact inference, content validity indices with modified kappa and score intervals, item-objective congruence, and panel-level agreement using Krippendorff's alpha as described by Hayes and Krippendorff (2007) &lt;<a href="https://doi.org/10.1080%2F19312450709336664" target="_top">doi:10.1080/19312450709336664</a>&gt;. Also provides judge and rater heterogeneity analysis following the generalizability-theory treatment of content-validity ratings in Crocker, Llabre and Miller (1988) &lt;<a href="https://doi.org/10.1111%2Fj.1745-3984.1988.tb00309.x" target="_top">doi:10.1111/j.1745-3984.1988.tb00309.x</a>&gt;, content-domain coverage and expert-perceived content structure following Sireci and Geisinger (1992) &lt;<a href="https://doi.org/10.1177%2F014662169201600102" target="_top">doi:10.1177/014662169201600102</a>&gt;, comparison across successive pretest rounds, and exact expert-panel planning. Where published methods compete, users choose among them through arguments with evidence-based defaults. User-facing workflows emphasize interpretable summaries and transparent review recommendations rather than isolated coefficients.

Provides quantitative tools for substantive and content-oriented scale pretesting. Implements item-sort indices from Anderson and Gerbing (1991) &lt;<a href="https://doi.org/10.1037%2F0021-9010.76.5.732" target="_top">doi:10.1037/0021-9010.76.5.732</a>&gt;, exact item-sort inference following Howard and Melloy (2016) &lt;<a href="https://doi.org/10.1007%2Fs10869-015-9404-y" target="_top">doi:10.1007/s10869-015-9404-y</a>&gt;, empirical interpretation benchmarks from Colquitt et al. (2019) &lt;<a href="https://doi.org/10.1037%2Fapl0000406" target="_top">doi:10.1037/apl0000406</a>&gt;, and the construct-rating procedure of Hinkin and Tracey (1999) &lt;<a href="https://doi.org/10.1177%2F109442819922004" target="_top">doi:10.1177/109442819922004</a>&gt; with HTC/HTD indices and repeated-measures item screening. The expert-panel workflow combines Aiken's V with score confidence intervals, Lawshe content validity ratios with exact inference, content validity indices with modified kappa and score intervals, item-objective congruence, and panel-level agreement using Krippendorff's alpha as described by Hayes and Krippendorff (2007) &lt;<a href="https://doi.org/10.1080%2F19312450709336664" target="_top">doi:10.1080/19312450709336664</a>&gt;. Also provides judge and rater heterogeneity analysis following the generalizability-theory treatment of content-validity ratings in Crocker, Llabre and Miller (1988) &lt;<a href="https://doi.org/10.1111%2Fj.1745-3984.1988.tb00309.x" target="_top">doi:10.1111/j.1745-3984.1988.tb00309.x</a>&gt;, content-domain coverage and expert-perceived content structure following Sireci and Geisinger (1992) &lt;<a href="https://doi.org/10.1177%2F014662169201600102" target="_top">doi:10.1177/014662169201600102</a>&gt;, comparison across successive pretest rounds, and exact expert-panel planning. Where published methods compete, users choose among them through arguments with evidence-based defaults. User-facing workflows emphasize interpretable summaries and transparent review recommendations rather than isolated coefficients.

Patrick Sturgis (London School of Economics) presented a survey methods webinar about using generative AI in surveys on 15 September 2026.

PretestPro, software that guides patients through four eye exam pretests in less than 2 minutes, has launched in the U.S., according to a press release from Heru. The software, available on virtual reality-powered Heru Prime headsets, guides patients through: PretestPro is intended to streamline workflows and be more consistent than traditional tests that rely on a clinician’s subjective assessment. “The difference in practice is that PretestPro produces the same assessments as quantitative, recorded measurements performed the same way on every patient, and those values stay in the record for

Patrick Sturgis (London School of Economics) presented a survey methods webinar about using generative AI in surveys on 15 September 2026.

An implementation of the DML-Bounds procedure of Villena (2026) &lt;<a href="https://doi.org/10.2139%2Fssrn.6472826" target="_top">doi:10.2139/ssrn.6472826</a>&gt; for testing cointegration in data-rich time-series settings. The Autoregressive Distributed Lag (ARDL) bounds test of Pesaran, Shin and Smith (2001) &lt;<a href="https://doi.org/10.1002%2Fjae.616" target="_top">doi:10.1002/jae.616</a>&gt; avoids pretesting the integration order of the regressors but is not designed for a high-dimensional conditioning set. Residualising the lagged levels against persistent controls can absorb stochastic trends and thereby change the finite-sample null distribution, so what governs the null is the effective number of stochastic trends surviving residualisation rather than the integration order of the original regressors. The procedure combines h-block cross-fitting, a balanced nuisance projection in the Double Machine Learning (DML) style of Chernozhukov and others (2018) &lt;<a href="https://doi.org/10.1111%2Fectj.12097" target="_top">doi:10.1111/ectj.12097</a>&gt;, adaptive weighting after Zou (2006) &lt;<a href="https://doi.org/10.1198%2F016214506000000735" target="_top">doi:10.1198/016214506000000735</a>&gt;, and a restricted system wild bootstrap that regenerates the dependent variable and the focal regressor jointly. No critical-value table is shipped: the classical bracket is regenerated by simulation and the operational critical value is bootstrapped. A trend-absorption diagnostic and a penalty-sensitivity sweep report whether a verdict survives a change of conditioning set. Monthly United States macroeconomic series from the 'FRED-MD' database of McCracken and Ng (2016) &lt;<a href="https://doi.org/10.1080%2F07350015.2015.1086655" target="_top">doi:10.1080/07350015.2015.1086655</a>&gt; are bundled so every example runs offline.

An implementation of the DML-Bounds procedure of Villena (2026) &lt;<a href="https://doi.org/10.2139%2Fssrn.6472826" target="_top">doi:10.2139/ssrn.6472826</a>&gt; for testing cointegration in data-rich time-series settings. The Autoregressive Distributed Lag (ARDL) bounds test of Pesaran, Shin and Smith (2001) &lt;<a href="https://doi.org/10.1002%2Fjae.616" target="_top">doi:10.1002/jae.616</a>&gt; avoids pretesting the integration order of the regressors but is not designed for a high-dimensional conditioning set. Residualising the lagged levels against persistent controls can absorb stochastic trends and thereby change the finite-sample null distribution, so what governs the null is the effective number of stochastic trends surviving residualisation rather than the integration order of the original regressors. The procedure combines h-block cross-fitting, a balanced nuisance projection in the Double Machine Learning (DML) style of Chernozhukov and others (2018) &lt;<a href="https://doi.org/10.1111%2Fectj.12097" target="_top">doi:10.1111/ectj.12097</a>&gt;, adaptive weighting after Zou (2006) &lt;<a href="https://doi.org/10.1198%2F016214506000000735" target="_top">doi:10.1198/016214506000000735</a>&gt;, and a restricted system wild bootstrap that regenerates the dependent variable and the focal regressor jointly. No critical-value table is shipped: the classical bracket is regenerated by simulation and the operational critical value is bootstrapped. A trend-absorption diagnostic and a penalty-sensitivity sweep report whether a verdict survives a change of conditioning set. Monthly United States macroeconomic series from the 'FRED-MD' database of McCracken and Ng (2016) &lt;<a href="https://doi.org/10.1080%2F07350015.2015.1086655" target="_top">doi:10.1080/07350015.2015.1086655</a>&gt; are bundled so every example runs offline.