Picture an admissions reader opening your file on a clear day, a partly cloudy day or a fully overcast day. In one university’s records, which sky went with academic ratings counting for the most?
The cloudier the day, the more reviewers’ recommendations leaned on academic ratings, and the sunnier the day, the more they leaned on social and special ratings. That held across 682 undergraduate applications at one unnamed university (Simonsohn, 2007). It is an association in old records, not proof that clouds changed anyone’s mind.
One twist: the overall rate did not move. In the author’s analysis, cloud cover showed no main link to how often reviewers recommended applicants. What differed with the sky was which ratings carried weight, not how many applicants were recommended.
Same file, two readers, two skies
The evidence comes from a 2007 study of 682 undergraduate applications at one unnamed university. Each file was read by two reviewers, 1,364 readings in all from at least 15 different reviewers, and the two disagreed on 119 files. The paper does not name the school, the review years or the country; its NOAA weather data and a US author suggest the school was probably in the US (our inference).
For each review day, the author took the city’s cloud cover from the National Oceanic and Atmospheric Administration, on a scale from 0 for clear skies to 10 for complete overcast; the average was 7.91. About 80% of files were read by the two reviewers under different cloud cover. The review offices, the paper notes, all have windows, giving reviewers, in the author’s words, “ample opportunity for cloudcover to be perceived”.
The reader who had the cloudier day put more weight on academic ratings, and the reader with the sunnier day put more on social and special ratings. The pattern held after the author added rain, wind and temperature, and after adding the average cloud cover for the same calendar date over the four preceding years.
The outcome here is a reviewer’s recommendation, not an admission decision; the paper says data on final decisions are not available. Cloud cover on the first review day did not predict the ratings themselves, and the number of files reviewed each day was not linked to it either. Applicants recommended on days with cloud cover above 5 had slightly higher academic ratings relative to social ones than those recommended below 5, 1.239 against 1.195 on the author’s ratio (p = .041), though the full pools reviewed did not differ.
How big is the link? For each applicant, the author’s model compared the predicted chance of a positive recommendation under the clearest and the cloudiest skies, and the gap averaged 11.9%. The paper counts that gap whichever sky came out ahead; since academic ratings counted for more on cloudier days and social ones on sunnier days, which sky came out ahead would depend on the applicant (our reading of the model). That is a model estimate, not a count of applicants, and the paper does not say whether the 11.9% is in percentage points.
A rainy morning in Toronto
A separate 2009 study looked at 2,926 Toronto medical school candidates interviewed from 2004 to 2009, almost all in early spring, with no exclusions. Interviewers scored candidates from 0 to 20, where 16 means “good” and 18 means “excellent”. A day counted as rainy if rain, freezing rain, snow or hail was recorded in the morning or afternoon; otherwise it was dry, which is not the same as clear.
Candidates interviewed on rainy days scored 16.31 on average, against 16.49 on dry days (p = 0.042). That is a gap of 0.18 points, about 1% (our arithmetic, matching the authors’ “about 1%”). With junior interviewers, the scores were 16.23 on rainy days and 16.42 on dry days (p = 0.041). With senior interviewers they were 16.39 and 16.55, but at p = 0.08 that gap is not statistically significant by the usual 0.05 standard.
The authors report that candidates’ demographic characteristics were unrelated to the weather. They frame the gap as about the size of a 10% lower total mark on the Medical College Admission Test. The paper does not report what happened to final offers.
Data limits
Both studies are observational records, so they show associations, not causes. The Toronto study covers one Canadian medical school from 2004 to 2009 and measures interview scores, not final offers. The undergraduate study covers one unnamed university, with review years not stated, and measures reviewers’ recommendations, not final decisions. One reading, labeled as the authors’ and not a finding: the authors suggest mood may play a role; the studies did not measure it. Neither study tested admissions as they run today, so whether these patterns hold now is unknown.
The short version
Two older datasets show small, measurable links between weather and admissions judgments. In one university’s files, cloudier review days went with more weight on academic ratings, and in one Toronto medical school, rainy interview days went with scores about 1% lower. Neither says anything about any applicant’s chances today, and neither is a reason to read the forecast before you apply.
Last checked: 2026-10-19, against Redelmeier and Baxter (CMAJ 2009) and Simonsohn (JBDM 2007). This article was machine-checked against the sources above and has not been expert-reviewed. Spotted an error or a newer public figure? Please tell us on our contact page with the article link and we will review it.
Sources
- Redelmeier DA, Baxter SD. “Rainy weather and medical school admission interviews.” CMAJ 2009 Dec 8;181(12):933 (Holiday Review section, Research). doi:10.1503/cmaj.091546. PMID 19969588. Accessed Oct 19, 2026, via the Wayback Machine copy of the CMAJ full text: https://web.archive.org/web/20151107201058id_/http://www.cmaj.ca:80/content/181/12/933.full
- Simonsohn U. “Clouds Make Nerds Look Good: Field Evidence of the Impact of Incidental Factors on Decision Making.” Journal of Behavioral Decision Making 2007;20(2):143 to 152. doi:10.1002/bdm.545. Accessed Oct 19, 2026, in the author’s repository copy (ScholarlyCommons, dated April 2007), via https://web.archive.org/web/20240307212405id_/https://core.ac.uk/download/pdf/219378519.pdf . The publisher’s typeset version was not opened; figures could differ from it.
- Method: the 0.18 point gap and its “about 1%” share are Book of Odds arithmetic from the printed means (16.49 minus 16.31; 0.18 divided by 16.49 is 1.09%). “Probably in the US” is our inference from the paper’s NOAA data. Any mood reading is the authors’ suggestion; neither study measured mood.