Psychometric Characteristics and Cultural Interpretation of LOTCA Visual Subtests Among Adults in Taiwan.

Cross-sectional study of 101 community-dwelling adults in Taiwan (aged 18–83, mean 44.6), all cognitively intact (MMSE ≥ 24), recruited by convenience sampling between December 2024 and December 2025. Each participant completed seven LOTCA visual subtests (four from the Visual Perception domain, three from the Visuomotor Construction domain) in a single session of about 10 minutes, administered in Mandarin with responses accepted in Mandarin or Southern Min. No formal forward-backward translation was done. Quantitative analyses included Cronbach's α for domain consistency, a Content Validity Index rated by eight occupational therapy experts, exploratory factor analysis (principal axis factoring, oblimin rotation), and ordinal logistic regression linking subtest scores to age, education, and gender (Bonferroni-corrected threshold p < .007). Qualitative analysis used inductive coding of video-recorded sessions to identify culturally situated response patterns.

Internal consistency was poor for both domains (VP α = .35, VC α = .49), with corrected item-total correlations ranging from .06 to .41. Expert content validity was below the authors' adequacy threshold: S-CVI/Ave was .56 for VP and .71 for VC (threshold .90); only Object Identification reached the item-level criterion (I-CVI = 1.00). Exploratory factor analysis yielded a two-factor solution accounting for 31.39% of common variance; Factor 1 mixed subtests from both original domains (TM, SI, OI, RP) and Factor 2 was essentially just Drawing a Clock, so the structure did not reproduce the VP/VC split. After Bonferroni correction, age was negatively associated with Shape Identification (B = −.076, p = .001) and Reproduction of a Two-Dimensional Model (B = −.106, p < .001); gender was positively associated with TM (B = 2.703, p < .001, males higher). Education was not significant after correction. No participant achieved the maximum TM score of 5; 67.3% scored 4. Qualitative coding identified four recurring response-pattern categories: language-related factors (56% of participants showed language-associated errors on VP tasks), educational influences, daily-life experience effects, and response habit tendencies.

If you are using LOTCA visual subtests with patients from a Taiwanese or similar East-Asian background, a low score on Object Identification, Shape Identification, Figure-Ground, or Object Constancy may reflect unfamiliar objects (e.g., agricultural tools, traditional telephone imagery), dialect-based shape naming (e.g., grouping squares and rectangles as 'four-corner shapes'), or a different construction strategy rather than a perceptual or cognitive deficit. Before attributing a low visual-subtest score to impairment, consider whether the stimulus is culturally familiar to the patient, whether their language organises visual categories differently from the scoring key, and whether their construction approach (trial-and-error, repeated model comparison) is a deliberate strategy rather than a deficit. The factor structure also suggests that the VP and VC domain scores may not be meaningful as separate indices in this population.

Convenience sample of 101 adults from a single region in Taiwan; generalisability to other populations is limited. All participants were cognitively intact (MMSE ≥ 24), so the findings say nothing about how these subtests perform in the clinical populations where LOTCA is most often used. Only seven subtests were examined, not the full LOTCA battery; results cannot be extrapolated to other domains. No formal forward-backward translation was conducted; the study used the English edition with Mandarin administration, so it does not establish a validated Taiwanese version. Cross-sectional design with no longitudinal or clinical comparison group. The gender difference on TM (males scoring higher) is based on a single subtest and the authors note the source of the difference cannot be determined from the data.

Funded by the National Science and Technology Council, Taiwan (grants 114-2813-C-040-076-H and 113-2410-H-040-003-MY3). The authors declared no conflicts of interest.

Do not read the gender difference on Reproduction of a Two-Dimensional Model (males 86.1% vs. females 56.9% scoring 4) as evidence that women have a visuoconstructive deficit. No participant achieved the maximum score of 5, the difference reflects successful completion through trial-and-error rather than a different strategy, and the authors themselves state the source of the difference cannot be determined from the data and should be interpreted cautiously pending replication.