LSAT Test Anxiety Study Findings: What the Research Actually Shows

Lsat test anxiety study findings are consistent and smaller than most applicants expect: Caviola and colleagues put the correlation with achievement at r = -.23, and Schillinger and colleagues put the variance it explains in standardized test performance at 2 to 5 percent. Verified July 2026.

What does the research on test anxiety actually find?

Test anxiety and achievement correlate negatively and modestly. The largest meta-analysis of it, Caviola and colleagues in 2022, puts the pooled correlation at r = -.23 across 177 studies and more than 900,000 participants, which is a real effect and a small one. Verified July 2026.

FindingFigureStudySourceTest anxiety and achievement, pooledr = -.23, 95% CI [-.26, -.19]Caviola et al. 2022, 177 studies, 906,311 participantshttps://link.springer.com/article/10.1007/s10648-021-09618-5Variance explained in standardized tests2 to 5 percentSchillinger et al. 2021https://d-nb.info/1231607017/34Anxiety interventions on anxiety itselfg = -0.76Huntley et al. 2019https://pubmed.ncbi.nlm.nih.gov/30826687/The same interventions on academic performanceg = 0.37Huntley et al. 2019, authors flag publication biashttps://pubmed.ncbi.nlm.nih.gov/30826687/

Caviola and colleagues report a pooled correlation of r = -.23 with a 95 percent confidence interval of -.26 to -.19, across 177 studies and 906,311 participants (https://link.springer.com/article/10.1007/s10648-021-09618-5), retrieved July 31, 2026. Verified July 2026.

A correlation of that size is real and modest. Squared, it accounts for roughly 5 percent of the variance between people, our arithmetic on the Caviola and colleagues figure. Verified July 2026.

That is the honest headline. Test anxiety matters and it is not the main thing standing between you and a higher LSAT score. Verified July 2026.

How much of a score does test anxiety actually cost?

Test anxiety explains only 2 to 5 percent of variance in standardized test performance, per Schillinger and colleagues, which is the most useful number in this literature and the least quoted one. That is a ceiling on how much composure could ever buy you. Verified July 2026.

Schillinger and colleagues report that test anxiety explained only 2 to 5 percent of variance in standardized test performance (https://d-nb.info/1231607017/34), retrieved July 31, 2026. Verified July 2026.

Set that against the LSAT's own measurement error. LSAC publishes a standard error of about 2.6 scaled score points (https://www.lsac.org/lsat-interpretive-guide/2019-2020), retrieved July 31, 2026, which is noise present in every score regardless of anxiety. Verified July 2026.

The practical implication is uncomfortable and worth stating. If your practice scores sit ten points below your target, anxiety is not the explanation and treating it as one costs you the study time that is. Verified July 2026.

Do anxiety interventions raise test scores?

Anxiety interventions reduce anxiety substantially and raise academic performance about half as much, and the authors of the largest review flag publication bias in their own results. Huntley and colleagues report an effect of g = -0.76 on anxiety itself but only g = 0.37 on academic performance, and flag publication bias, retrieved July 31, 2026. Verified July 2026.

That gap is the finding. Interventions reliably make people feel better and move scores considerably less, from that same Huntley and colleagues review. Verified July 2026.

Feeling better is a legitimate goal on its own terms and this page is not dismissing it. It is a different goal from scoring higher, and conflating them wastes both. Verified July 2026.

Why does test anxiety feel more decisive than it measures?

Anxiety is vivid and studying is not, so the memory of a bad test day is far more available than the memory of the weeks that produced the score. The learning literature documents the same pattern elsewhere. Verified July 2026.

Kornell and Bjork found 78 percent of learners judged massed practice as good as or better than spacing while 78 percent actually performed better spaced (https://gwern.net/doc/psychology/spaced-repetition/2008-kornell.pdf), retrieved July 31, 2026. Verified July 2026.

Deslauriers and colleagues found active-learning students rated their own learning 0.56 standard deviations lower while scoring 0.46 higher (https://www.pnas.org/doi/10.1073/pnas.1821936116), retrieved July 31, 2026. Self-report about learning is systematically unreliable. Verified July 2026.

Applied here, that means your sense that anxiety cost you the points is exactly the kind of judgment the literature shows people get wrong. Verified July 2026.

What does the research say actually moves an LSAT score?

Retrieval practice with feedback has the largest documented effect in this literature, and it is a study method rather than a state of mind. Rowland reports that testing beats restudy at g = 0.50 across 61 studies and 159 effect sizes, retrieved July 31, 2026. Verified July 2026.

Feedback roughly doubles that effect in the same Rowland analysis, at g = 0.73 with feedback against g = 0.39 without (https://pubmed.ncbi.nlm.nih.gov/25150680/), retrieved July 31, 2026. Verified July 2026.

Against those figures, the Schillinger and colleagues 2 to 5 percent for anxiety is a second-order problem (https://d-nb.info/1231607017/34), retrieved July 31, 2026. Fix the study method first. Verified July 2026.

Is test anxiety ever the main problem?

Yes, in one identifiable pattern: when your timed performance is materially below your untimed performance on the same material. That gap is measurable rather than felt; The diagnostic is arithmetic rather than introspection; Score a section untimed, score a comparable section timed, and compare. Verified July 2026.

The diagnostic is arithmetic rather than introspection. A large stable gap is evidence; a bad test day is not. Verified July 2026.

Where that gap exists, the Huntley and colleagues finding is encouraging on anxiety itself at g = -0.76 (https://pubmed.ncbi.nlm.nih.gov/30826687/), retrieved July 31, 2026, even though its academic effect is smaller. Verified July 2026.

Where the gap does not exist, the untimed score is the ceiling and the work is content rather than composure. That is the distinction this page exists to draw. Verified July 2026.

What this research does not tell you

None of these studies is about the LSAT specifically, none measures scaled-score points, and one of them flags publication bias in its own results. Caviola and colleagues, Schillinger and colleagues and Huntley and colleagues all study standardized testing and academic performance broadly rather than the LSAT, retrieved July 31, 2026. Verified July 2026.

None reports an effect in LSAT scaled points, so anyone quoting you a points figure for anxiety has converted between scales without a published conversion. Verified July 2026.

Huntley and colleagues flag publication bias in their own intervention results (https://pubmed.ncbi.nlm.nih.gov/30826687/), retrieved July 31, 2026, which means the true intervention effect is probably smaller than g = 0.37 rather than larger. Verified July 2026.

What should you actually do with this?

Run the timed against untimed comparison, and if the gap is small, spend your remaining weeks on retrieval practice with feedback rather than on managing your nerves. That is a testable recommendation rather than reassurance; The comparison takes two sections and an afternoon. Verified July 2026.

If the gap is large, treat anxiety as a real target and expect a modest score effect, per Huntley and colleagues at g = 0.37 on performance (https://pubmed.ncbi.nlm.nih.gov/30826687/), retrieved July 31, 2026. Verified July 2026.

Either way, the Rowland finding stands: testing with feedback at g = 0.73 is the largest documented lever available to you (https://pubmed.ncbi.nlm.nih.gov/25150680/), retrieved July 31, 2026. Verified July 2026.

Where every figure on this page came from

Every figure on this test anxiety research page carries a Verified July 2026 stamp because a dated stamp is the only honest way to publish a number that moves. Each source below was retrieved in July 2026 and is linked in full rather than named vaguely, so nothing here is modelled, averaged or inferred.

What it supportsSourceCaviola et al. 2022, test anxiety and achievementhttps://link.springer.com/article/10.1007/s10648-021-09618-5Schillinger et al. 2021, variance explainedhttps://d-nb.info/1231607017/34Huntley et al. 2019, anxiety interventionshttps://pubmed.ncbi.nlm.nih.gov/30826687/Rowland 2014, retrieval practicehttps://pubmed.ncbi.nlm.nih.gov/25150680/LSAC on the standard error of measurementhttps://www.lsac.org/lsat-interpretive-guide/2019-2020

FAQ

Does test anxiety lower LSAT scores?

Modestly. Caviola and colleagues put the pooled correlation between test anxiety and achievement at r = -.23 across 177 studies, and Schillinger and colleagues put the variance it explains at 2 to 5 percent. Verified July 2026.

Do anxiety interventions raise test scores?

They reduce anxiety substantially at g = -0.76 and raise academic performance about half as much at g = 0.37, and the authors flag publication bias in their own results. Verified July 2026.

How do I know if anxiety is my real problem?

Compare a timed section against an untimed section on comparable material. A large stable gap is evidence.

A single bad test day is not. Verified July 2026.

How many LSAT points does anxiety cost?

No study reports an effect in LSAT scaled points. Anyone quoting you a points figure has converted between scales without a published conversion. Verified July 2026.

What moves an LSAT score more than managing anxiety?

Retrieval practice with feedback. Rowland reports testing beats restudy at g = 0.50, rising to g = 0.73 when feedback is added. Verified July 2026.

What to do next

The measurable version of this question is a timed against untimed comparison, and it takes an afternoon. Where you actually stand against school bands is the next one. The Lovare Diagnostic takes about fifteen minutes and returns your own numbers against published school data rather than a single probability, and it is free.

Written by Ali, Georgetown Law, founder of Lovare Institut.

Book a call