Psychometric Characteristics and Cultural Interpretation of LOTCA Visual Subtests Among Adults in Taiwan Authors: Chia-Hui Hung, Ling-Jie Hung, Wan-Yi Hung, Shih-Chen Fan, Hsiao-Ping Chiu Published: 2026 ABSTRACT AbstractObjectivesPerformance-based cognitive assessments may function differently across sociocultural contexts because language, experience, and culturally acquired representational systems influence task performance. This study examined the psychometric characteristics and culturally informed interpretation of selected Loewenstein Occupational Therapy Cognitive Assessment (LOTCA) visual subtests among adults in Taiwan.MethodsA cross-sectional design was used. A total of 101 community-dwelling Taiwanese adults with intact global cognition completed selected LOTCA visual subtests. Internal consistency, content validity, exploratory factor analysis, ordinal logistic regression, and qualitative content analysis of video-recorded assessment sessions were conducted.ResultsInternal consistency was low at the domain level. Exploratory factor analysis identified a two-factor solution that did not reproduce the original visual perception and visuomotor construction structure. Factor 1 comprised Shape Identification (SI), Object Identification (OI), Reproduction of a Two-Dimensional Model (TM), and Puzzle Reproduction (RP), whereas Factor 2 was represented primarily by Drawing a Clock (DC). Age was associated with SI and TM, and gender was associated with TM after Bonferroni correction. Demographic influences were concentrated in SI and TM (Nagelkerke R 2 = .630 and .582, respectively). Qualitative findings identified linguistically mediated categorization, experiential familiarity, stimulus-based misidentification, and strategy-based task regulation as influences on score interpretation.ConclusionsLOTCA visual subtest performance reflected interactions among perceptual demands, language, education, experience, and culturally acquired representational systems. Score variation should not be interpreted solely as evidence of cognitive limitation. Culturally informed interpretation and response-process levels may inform more accurate interpretation for visual perception performance in occupational therapy cognitive assessments.Keywords: cognitive assessment, cultural interpretation, visual perception, visuomotor construction, loewenstein occupational therapy cognitive assessment, occupational therapy Introduction The Loewenstein Occupational Therapy Cognitive Assessment (LOTCA) was developed as a performance-based instrument to support occupational therapy evaluation of cognitive functioning in individuals with neurological conditions (Katz et al., 1989; Natar et al., 2015). From an occupational perspective of cognition, LOTCA performance is understood as emerging through task engagement within environmental contexts, rather than as a decontextualized indicator of neuropsychological ability alone (Baum et al., 2023; Giles et al., 2025; Toglia & Foster, 2025). Subsequent versions, including LOTCA-II and the Dynamic LOTCA (DLOTCA), incorporated mediated learning principles to extend assessment from static task outcomes to cognitive learning mechanisms and potential for change (Itzkovich et al., 2000; Josman et al., 2011; Katz et al., 2011; Lund et al., 2014). Despite these conceptual advances, limited evidence explains how LOTCA scores should be interpreted when the assessment is used across sociocultural contexts (Almomani et al., 2018; Lund et al., 2014). Cross-cultural neuropsychology has demonstrated that cognitive task performance, including visual and nonverbal tasks, is shaped by educational exposure, linguistic conventions, and culturally learned representational systems, raising concerns about the assumed cultural neutrality and validity of standardized assessments (Rosselli & Ardila, 2003). Variability in LOTCA performance may therefore reflect culturally mediated response tendencies rather than differences in cognitive processing alone. Widely accepted validity frameworks further emphasize that meaningful score interpretation requires attention to response processes, including how individuals understand, interpret, and generate responses to task demands (American Educational Research Association [AERA] et al., 2014). Within the LOTCA framework, the Visual Perception (VP) and Visuomotor Construction (VC) domains have been consistently identified as sensitive to sociocultural and educational factors (Almomani et al., 2018; Katz et al., 1997; Lund et al., 2014). VP centers on perceptual categorization and discrimination, whereas VC requires the integration of visual input, spatial organization, and motor planning during constructional task performance (Natar et al., 2015). Previous research has reported distinct psychometric characteristics across these domains. The VP domain showed psychometric vulnerability in the DLOTCA, with relatively low internal consistency attributable primarily to the Shape Identification subtest, which was subsequently omitted (Katz et al., 2011). Ceiling effects and reduced discriminative capacity have also been reported in neurologically typical and culturally diverse samples, suggesting that visually mediated tasks may draw on culturally learned representational conventions rather than perceptual processing alone (Almomani et al., 2018; Lund et al., 2014; Rosselli & Ardila, 2003). The VC domain, by contrast, has shown consistent associations with age, educational exposure, and environmental experience across pediatric and adult samples, reflecting reliance on acquired constructional strategies rather than neurological status alone (Josman et al., 2011; Katz et al., 1997; Parush et al., 2000). Cross-cultural adaptations of visual perception assessments in Asian populations have similarly documented age-related score decline, underscoring the need to interpret visually mediated task performance considering cultural and demographic context (Hong et al., 2021). Sociocultural variation in VP and VC performance indicates that visually mediated task performance may be shaped by culturally situated experience (Josman et al., 2011; Parush et al., 2000). In cultural neuropsychology, cognitive testing is understood as a socially situated interaction influenced by implicit testing norms, examiner instructions, examinee interpretation of task demands, and examiner interpretation of responses (AERA et al., 2014; Ardila, 2005). Assessment performance captures not only individual cognitive processing but also the fit between test demands, examinee meaning-making, and scoring expectations. Evidence from international adaptations indicates that direct translation of test stimuli does not necessarily establish conceptual equivalence, or the extent to which test stimuli carry comparable meanings across sociocultural groups (Chithiramohan et al., 2024; Harachi et al., 2006; Neumann-Collyer et al., 2023). The Danish LOTCA-II adaptation reported psychometric limitations in several domains, suggesting limited discriminative capacity outside the original normative context (Lund et al., 2014), whereas the Malaysian LOTCA-G adaptation similarly highlighted the importance of linguistic accessibility during administration (Natar et al., 2015). Therefore, subtest-level structural analysis and qualitative examination of response processes may provide complementary evidence regarding whether LOTCA visual tasks function as intended across populations (AERA et al., 2014; Zumbo, 2007). Age and educational attainment have been consistently associated with visually mediated task performance, operating through differences in strategy selection and task familiarity (Ebaid & Crewther, 2019; Katz et al., 1997). In prior LOTCA research, older adults performed more slowly and benefited from simplified versions of constructional tasks (Katz et al., 1997). Higher educational attainment has similarly been associated with greater exposure to symbolic reasoning and structured problem-solving, which may attenuate age-related performance variation in visually mediated tasks (Ebaid & Crewther, 2019). Whether these demographic influences operate similarly among adults in Taiwan, where educational experiences and sociocultural contexts differ from those of previously studied populations, remains to be examined. LOTCA performance in this context may therefore reflect not only cognitive capacity but also the accumulated influence of culturally specific experiential learning and educational exposure. The absence of integrated psychometric and response process evidence carries meaningful consequences for occupational therapy practice. When clinicians interpret LOTCA performance without understanding how cultural, linguistic, and educational factors shape task responses, there is a risk of attributing score variation to cognitive impairment rather than to differences in how tasks are understood and approached (Ardila, 2005; Rosselli & Ardila, 2003). This study addresses these limitations by integrating structural analysis with systematic qualitative examination of response patterns among adults in Taiwan. Although LOTCA-related evidence has been reported in selected Taiwanese clinical populations (Su et al., 2007), validated normative data and culturally informed interpretive guidance for community-dwelling adults remain insufficient. This study examined the psychometric characteristics and culturally informed interpretation of selected LOTCA visual subtests among adults in Taiwan, with a focus on subtest-level patterns and the role of age, educational attainment, and sociocultural experience in assessment interpretation. Seven subtests were selected based on prior evidence of sensitivity to age, education, and sociocultural experience (Katz et al., 1997): the VC domain comprised Reproduction of a Two-Dimensional Model, Reproduction of a Puzzle, and Drawing a Clock, whereas the VP domain comprised Object Identification, Shape Identification, Figure-Ground, and Object Constancy. Although some VP subtests were subsequently removed from the DLOTCA because of psychometric concerns, they were retained here to examine linguistically and educationally mediated response patterns that remain insufficiently explored in Asian adult samples. Specifically, this study investigated how language, object familiarity, and educational experience may influence performance and contribute evidence for culturally informed occupational therapy cognitive assessment. Methodology Study Design This was a cross-sectional study aimed to examine the psychometric performance and cultural applicability of selected VP and VC subtests of the LOTCA in a Taiwanese context. The analysis focused on subtest-level performance patterns and their associations with age, education, and sociocultural experience. Seven subtests were selected for single-session assessment because prior LOTCA studies, including research with adult and pediatric samples, have suggested that visual and constructional task performance may be sensitive to age, educational exposure, and sociocultural experience (Josman et al., 2011; Katz et al., 1997; Parush et al., 2000). Participants and Recruitment A convenience sample of 101 local community-dwelling adults was recruited in Taiwan between December 2024 and December 2025. The study received ethical approval from the Institutional Review Board of Chung Shan Medical University Hospital (approval number CS1 23218), and all participants provided written informed consent before enrollment. Participants were required to be 18 years or older and have intact global cognition, as measured by a Mini-Mental State Examination (MMSE) score of 24 or higher (Creavin et al., 2016; Tombaugh & McIntyre, 1992). Limiting the sample to cognitively intact adults allowed us to test subtest variability independently of overt cognitive impairment. Individuals with conditions that could interfere with task performance were excluded, including diagnosed neurological disorders, severe psychiatric illness, uncorrected visual impairment, or motor impairment preventing manipulation of test materials. Sensory or motor mild changes associated with age were not considered exclusionary unless they interfered with task completion. Educational status was determined by the completion of total years of formal schooling as self-reported by the participants. Panel of Experts in Content Validity Eight Taiwanese occupational therapy experts evaluated the cultural recognizability and contextual appropriateness of the selected LOTCA visual subtests. The panel included academic and clinical occupational therapists with experience in cognitive assessment and LOTCA use, representing pediatric, mental health, and physical rehabilitation practice settings, with an average of 14.6 years (range 8–25 years) of professional experience. All experts completed the appraisal in Mandarin to preserve linguistic and cultural nuance. Experts rated each subtest in relation to stimulus recognition, cultural familiarity, clarity of task interpretation, and appropriateness of scoring criteria in the Taiwan context. A five-point Likert scale (1 = strongly disagree to 5 = strongly agree) was used, allowing ratings to be dichotomized for Content Validity Index calculation. Experts also provided written comments regarding culturally embedded assumptions, contextual uncertainty, and scoring ambiguity. Instrument The official English editions of the LOTCA-II (Itzkovich et al., 2000) and DLOTCA (Katz et al., 2011) were used because no officially standardized Taiwanese version is currently available. Test instructions were administered in Mandarin Chinese by trained research team members and occupational therapists, and participants were allowed to respond in Mandarin or Taiwanese Southern Min expressions. Mandarin Chinese is the predominant language used in education and formal communication in Taiwan and is widely understood across the population. Taiwanese Southern Min is also commonly used in everyday communication, particularly among middle-aged and older adults. Accordingly, LOTCA instructions were administered in Mandarin to maintain consistency across participants, while responses in either Mandarin or Taiwanese Southern Min were accepted to preserve natural language use during assessment. A formal forward-backward translation process was not conducted because the study aimed to examine response patterns and cultural interpretation rather than to develop a Taiwanese version of the LOTCA. Although LOTCA-II psychometric evidence has been reported in persons with schizophrenia in Taiwan (Su et al., 2007), validation in community-dwelling Taiwanese adults remains insufficient. Selected subtests from the VP and VC domains of LOTCA were administered. Subtests were drawn from the LOTCA-II and the DLOTCA based on their theoretical correspondence with the original VP and VC domains, prior empirical use, and reported sensitivity to sociocultural or educational factors. Previous DLOTCA research has reported acceptable internal consistency for these domains, with Cronbach’s α ranging from 0.60 to 0.81 (Katz et al., 2011). Shape Identification was retained from LOTCA-II to examine possible linguistic and educational effects. The final subtests included Object Identification (OI), Shape Identification (SI), Figure-Ground (FG), Object Constancy (OC), Reproduction of a Two-Dimensional Model (TM), Reproduction of a Puzzle (RP), and Drawing a Clock (DC). Administration procedures and scoring criteria are provided in Supplemental Table S1. Procedure Assessments were conducted individually in a quiet testing environment under standardized conditions. After MMSE screening, participants completed the seven selected LOTCA subtests in a fixed administration order. Administration of the seven LOTCA subtests required approximately 10 minutes per participant. All sessions were audio- and video-recorded. Observational behavioral data were extracted for subsequent qualitative coding and reliability verification. Data Analysis Quantitative and qualitative analyses were conducted to examine subtest-level performance patterns and culturally situated response processes. Demographic characteristics were summarized descriptively. Associations between subtest performance and sociodemographic variables were examined using ordinal logistic regression, with age and educational attainment entered as continuous variables and gender entered as a categorical variable. Separate models were constructed for each subtest, and a Bonferroni-corrected significance threshold of p < .007 (.05/7) was applied to control for multiple comparisons. Internal consistency was evaluated using Cronbach’s α for the VP and VC domains, with item-total correlations examined to assess each subtest’s contribution to its domain score. Content validity was assessed using the Content Validity Index (CVI) following Lynn (1986). Expert ratings of 4 or 5 on the five-point scale were classified as agreement. I-CVI ≥.78 and S-CVI/Ave ≥.90 were interpreted as adequate (Polit & Beck, 2006). Exploratory factor analysis (EFA) was conducted using principal axis factoring with Direct Oblimin rotation. Sampling adequacy and factorability were evaluated using the Kaiser-Meyer-Olkin measure and Bartlett’s test of sphericity. Given the small number of variables (seven subtests), a sample of 101 participants was considered sufficient for EFA (MacCallum et al., 1999). Factor retention was determined using eigenvalues greater than 1.0, scree plot inspection, and theoretical interpretability. Pattern matrix loadings ≥.40 were considered salient, loadings between .20 and .39 were examined as weak or secondary loadings, and coefficients below .20 were suppressed for clarity (Hair et al., 2019). Statistical analyses were conducted using SPSS Version 23 (IBM Corp., Armonk, NY, USA). Qualitative behavioral error analysis was applied using an inductive qualitative content analysis approach (Hsieh & Shannon, 2005). Video recordings were reviewed to identify recurrent response patterns within each subtest. Coding categories were developed inductively through open coding and refined through repeated comparison across cases. Two trained coders reviewed all responses, resolved discrepancies through discussion, and documented final coding decisions. Expert panel comments were analyzed using the same inductive approach to identify recurring concerns related to cultural relevance and scoring clarity. Methodological rigor was supported through consensus-based coding, iterative verification, and documentation of analytic decisions (Nowell et al., 2017). Results Demographic Analysis A total of 101 Taiwanese adults (aged 18 to 83 years; mean = 44.6, SD = 19.8) participated in the study. There were 37 male (36.6%) and 64 female (63.4%) participants. All participants demonstrated intact global cognition, with MMSE scores ≥ 24. Details of demographic characteristics are presented in Table 1. Table 1.Demographic Characteristics of Participants (N = 101)VariableParticipants (n)Percentage (%)Gender Male3736.6 Female6463.4Age group, years 18–30 years3332.7 31–50 years3332.7 >50 years3534.7Years of education ≤9 years2322.8 10-12 years4241.6 >12 years3635.6Cognitive screening MMSE score ≥ 24101100Note. MMSE = Mini-Mental State Examination. Demographic Characteristics of Participants (N = 101) Note. MMSE = Mini-Mental State Examination. Internal Consistency As summarized in Table 2, domain-level internal consistency was poor, with Cronbach’s α coefficients of .35 for VP and .49 for VC. Corrected item-total correlations varied across the seven subtests, ranging from .06 to .41, suggesting uneven subtest contributions within their respective domain classifications. Within the VP domain, OC showed the lowest corrected item-total correlation, whereas DC showed the lowest corrected item-total correlation within the VC domain. Table 2.Internal Consistency and Content Validity of LOTCA Visual Subtests​​ReliabilityValidityDomainSubtestsMean (SD)Domain Cronbach’s αCorrected ITC I-CVI Visual perceptionOI3.99 (0.10).35.151.00SI3.63 (0.53).30.25FG3.25 (0.48).31.37OC3.62 (0.51).06.63Visuomotor constructionTM4.12 (0.61).49.41.75RP4.05 (0.67).40.63DC3.78 (0.89).16.75Note. OI = Object Identification; SI = Shape Identification; FG = Figure-Ground; OC = Object Constancy; TM = Reproduction of a Two-Dimensional Model; RP = Reproduction of a Puzzle; DC = Drawing a Clock; ITC = corrected item-total correlation; I-CVI = item-level Content Validity Index. Corrected item-total correlations were calculated for all subtests within their respective domain classifications. Internal Consistency and Content Validity of LOTCA Visual Subtests Note. OI = Object Identification; SI = Shape Identification; FG = Figure-Ground; OC = Object Constancy; TM = Reproduction of a Two-Dimensional Model; RP = Reproduction of a Puzzle; DC = Drawing a Clock; ITC = corrected item-total correlation; I-CVI = item-level Content Validity Index. Corrected item-total correlations were calculated for all subtests within their respective domain classifications. Content Validity Expert content validity indices are presented in Table 2 with detailed distribution of expert ratings in Supplemental Table S2. Overall agreement regarding the cultural appropriateness of the LOTCA visual subtests was modest at S-CVI/Ave values of .56 for the VP domain and .71 for the VC domain. At the subtest level, OI provided complete expert consensus (I-CVI = 1.00). Conversely, the other VP subtests had lower and variable levels of agreement, with SI demonstrating the lowest agreement (I-CVI = .25), followed by FG (I-CVI = .37) and OC (I-CVI = .63). Within the VC domain, I-CVI values ranged from .63 to .75, with RP demonstrating lower agreement (I-CVI = .63) compared to TM and DC (both I-CVI = .75). Qualitative Expert Feedback Qualitative feedback from the expert panel (N = 8) identified four recurring concerns across the selected subtests: linguistic variation, stimulus familiarity, generational differences in object exposure, and scoring ambiguity. First, certain shape names may differ across Taiwanese languages, especially the use of Taiwanese Southern Min dialect, and regional differences in terminology may play a role in verbal responses. Shapes considered to be less frequently used in everyday language and therefore may cause examinees to list descriptive instead of conventional shapes. Second, concerns were raised regarding the familiarity of certain visual stimuli. For example, agricultural tools such as hoes were considered less familiar in the context of contemporary urban life in Taiwan, which may contribute to variability in naming responses. In addition, characteristics of the visual presentation, such as overlapping figures, were noted to allow multiple plausible naming interpretations. Third, several experts focused on the ambiguity of scoring criteria used, specifically relating to the operational interpretation of trial-and-error performance in construction tasks and the scoring differentiation in the clock-drawing subtest. Lastly, experts identified minor generational differences in stimulus familiarity, such as the use of traditional telephone imagery rather than mobile phones. These elements were not considered to alter construct representation but may influence familiarity with the stimuli. Structural Validity EFA was conducted using principal axis factoring with direct oblimin rotation. The Kaiser-Meyer-Olkin measure indicated acceptable sampling adequacy (KMO = .626), and Bartlett’s test of sphericity was significant (χ 2 (21) = 102.653, p < .001), supporting factorability of the correlation matrix. Two factors had initial eigenvalues greater than 1.0 (2.284 and 1.214), and scree plot inspection supported retention of a two-factor solution. Based on extraction sums of squared loadings, the two-factor solution accounted for 31.39% of the common variance. This structure suggested an alternative organization of the selected visual subtests within this Taiwanese sample. The subtest clustering did not fully correspond to the theoretically proposed VP and VC domains (Table 3). The factor correlation matrix showed a low correlation between the two factors (r = .198). Factor 1 was interpreted as visuospatial organization and included salient loadings from TM, SI, OI, and RP. The factor comprised two VP subtests (OI and SI) and two VC subtests (TM and RP). Factor 2 was represented primarily by DC. FG and OC did not achieve salient pattern matrix loadings (≥.40) on either factor and exhibited low communalities, suggesting limited alignment with the extracted factor structure. Table 3.Exploratory Factor Analysis of LOTCA Visual Subtests Using Principal Axis Factoring With Direct Oblimin Rotation (N = 101)Subtests (domain)Factor 1Factor 2Communality (h 2 )TM (from VC).797​.642SI (from VP).643.331.607OI (from VP).488−.212.242RP (from VC).437​.202DC (from VC).224.507.277OC (from VP)​​.070FG (from VP)−.232.269.156Extraction eigenvalue1.754.443​Variance, %25.066.33​Cumulative variance, %25.0631.39​Note. Pattern matrix coefficients from principal axis factoring with direct oblimin rotation are presented. Coefficients below .20 were suppressed for clarity. Extraction eigenvalues and variance percentages are based on extraction sums of squared loadings. Because oblique factors are correlated, rotated sums of squared loadings were not used to calculate total explained variance. Pattern matrix loadings ≥.40 were considered salient for factor interpretation (Hair et al., 2019). Loadings below .40 were examined descriptively to aid interpretation of subtests demonstrating limited convergence with the extracted factor structure. VP = Visual Perception; VC = Visuomotor Construction; TM = Reproduction of a Two-Dimensional Model; SI = Shape Identification; OI = Object Identification; RP = Reproduction of a Puzzle; DC = Drawing a Clock; OC = Object Constancy; FG = Figure-Ground. Exploratory Factor Analysis of LOTCA Visual Subtests Using Principal Axis Factoring With Direct Oblimin Rotation (N = 101) Note. Pattern matrix coefficients from principal axis factoring with direct oblimin rotation are presented. Coefficients below .20 were suppressed for clarity. Extraction eigenvalues and variance percentages are based on extraction sums of squared loadings. Because oblique factors are correlated, rotated sums of squared loadings were not used to calculate total explained variance. Pattern matrix loadings ≥.40 were considered salient for factor interpretation (Hair et al., 2019). Loadings below .40 were examined descriptively to aid interpretation of subtests demonstrating limited convergence with the extracted factor structure. VP = Visual Perception; VC = Visuomotor Construction; TM = Reproduction of a Two-Dimensional Model; SI = Shape Identification; OI = Object Identification; RP = Reproduction of a Puzzle; DC = Drawing a Clock; OC = Object Constancy; FG = Figure-Ground. Associations With Participant Characteristics Ordinal logistic regression analyses were conducted to examine associations between LOTCA visual subtest performance and participant characteristics. Due to near-complete separation in the score distribution of OI, the regression model for this subtest produced unstable parameter estimates and was not interpreted. Results are summarized in Table 4. Table 4.Ordinal Logistic Regression Models for Associations Between Participant Characteristics and LOTCA Visual Subtest PerformanceSubtestModel χ 2 (df)Model pNagelkerke R2PredictorBSEWald p OINot interpreted​​Not interpreted​​​​SI66.955 (3)<.001.630Age−.076.02211.419.001*Education.373.1486.315.012Gender−.605.5911.049.306FG15.575 (3).001.189Age−.015.0151.060.303Education.242.1005.859.015Gender−.505.4851.084.298OC2.980 (3).395.038Age−.022.0142.500.114Education−.111.0742.225.136Gender.106.444.058.810TM58.632 (3)<.001.582Age−.106.02517.476<.001*Education−.025.097.067.796Gender2.703.74313.227<.001*RP17.855 (3)<.001.186Age−.034.0146.178.013Education.060.072.682.409Gender.580.4291.825.177DC11.410 (3).010.188Age−.023.026.785.376Education.147.1181.569.210Gender−.271.676.161.688Note. OI = Object Identification; SI = Shape Identification; FG = Figure-Ground; OC = Object Constancy; TM = Reproduction of a Two-Dimensional Model; RP = Reproduction of a Puzzle; DC = Drawing a Clock. Ordinal logistic regression models included age and educational attainment as continuous predictors and gender as a categorical predictor. Education indicates years of formal schooling. Gender was coded with female participants as the reference category. The OI model was not interpreted because near-complete separation resulted in unstable parameter estimation. Statistical significance was determined using a Bonferroni-adjusted threshold of p < .007. *p < .007. Ordinal Logistic Regression Models for Associations Between Participant Characteristics and LOTCA Visual Subtest Performance Note. OI = Object Identification; SI = Shape Identification; FG = Figure-Ground; OC = Object Constancy; TM = Reproduction of a Two-Dimensional Model; RP = Reproduction of a Puzzle; DC = Drawing a Clock. Ordinal logistic regression models included age and educational attainment as continuous predictors and gender as a categorical predictor. Education indicates years of formal schooling. Gender was coded with female participants as the reference category. The OI model was not interpreted because near-complete separation resulted in unstable parameter estimation. Statistical significance was determined using a Bonferroni-adjusted threshold of p < .007. *p < .007. After Bonferroni correction, age was significantly associated with SI (B = −.076, Wald = 11.419, p = .001) and TM (B = −.106, Wald = 17.476, p < .001), indicating that higher age was associated with lower performance on these subtests. Age was not significantly associated with FG, OC, RP, or DC after correction. Educational attainment was not significantly associated with any LOTCA visual subtest after Bonferroni correction. Although education attainment showed uncorrected associations with SI (p = .012) and FG (p = .015), these associations did not meet the corrected significance threshold. Gender was significantly associated with TM (B = 2.703, Wald = 13.227, p < .001), with male participants demonstrating higher performance than female participants. No significant gender associations were observed for the remaining subtests after Bonferroni correction. Examination of TM score distributions indicated that no participant achieved the maximum score of 5. Most participants (67.3%) received a score of 4, indicating successful construction through trial and error. Male participants achieved a score of 4 at a higher rate than female participants (86.1% vs. 56.9%), with female participants more frequently receiving a score of 3 (40.0% vs. 13.9%). Nagelkerke R2 values ranged from .038 for OC to .630 for SI, indicating that the explanatory strength of the demographic models varied substantially across subtests. Patterns of Behavioral Response Systematic analysis of video-recorded sessions identified four categories of culturally situated response patterns across LOTCA visual subtests: language-related factors, educational influences, daily-life experience effects, and response habit tendencies. Language-associated errors were observed in 56% of participants during VP tasks, with cultural familiarity, dialect variation, and educational background influencing naming accuracy, geometric classification, and object recognition across subtests. Strategic compensatory behaviors, including material rotation and model-dependent referencing, were more frequently observed among older participants in VC tasks. Detailed coding categories, representative response examples, and narrative descriptions of each pattern are provided in Supplemental Table S3. Discussion The present findings demonstrate that LOTCA visual subtest performance among Taiwanese adults was shaped by factors beyond perceptual demand alone. Low internal consistency, modest content validity, an alternative factor structure, and culturally situated response patterns collectively indicate that subtest scores reflected interactions among task demands, language use, object familiarity, educational exposure, and response strategies. Interpretation of Factor Structure Internal consistency estimates for the VP and VC domains were low in the present sample (VP α = .35; VC α = .49), suggesting that the selected subtests did not operate as homogeneous indicators of the original LOTCA domains among Taiwanese adults. Similar findings have been reported in culturally adapted LOTCA applications, where psychometric instability and reduced discriminative capacity have been observed outside the original normative context (Almomani et al., 2018; Lund et al., 2014). Content validity findings converged with internal consistency and factor analytic results. S-CVI/Ave values for VP (.56) and VC (.71) fell below the recommended threshold of .90, and most subtests did not achieve the I-CVI criterion of .78. Exploratory factor analysis further identified a two-factor structure that did not fully reproduce the original VP and VC domains. Collectively, low content validity, poor internal consistency, and a factor structure that differed from the original model indicate that the selected visual subtests were not interpreted through a single shared framework by participants or experts. Rather than indicating a lack of construct relevance, the convergence suggests that the selected visual stimuli, task demands, and scoring expectations may not align fully with the linguistic, educational, and sociocultural experiences of Taiwanese adults (AERA et al., 2014; Ardila, 2005; Rosselli & Ardila, 2003). Factor 1 was characterized by salient loadings from TM, SI, OI, and RP and was interpreted as visuospatial organization. Although the factor retained a visuospatial profile, its composition suggests a broader process involving the organization of visual information through learned representational systems. OI required participants to link visual forms to stored object concepts, SI required assignment of shapes to learned categorical systems, and TM and RP required organization of visual elements according to structural and spatial conventions. The convergence of subtests from both VP and VC domains therefore suggests that performance depended on culturally and educationally acquired systems for making visual information meaningful and actionable, rather than visual perception or visuomotor construction alone. Qualitative findings further support this interpretation, as alternative categorization strategies in SI and strategy-based approaches in constructional tasks indicated reliance on culturally familiar frameworks for interpreting and organizing visual information. Functional cognition perspectives similarly emphasize that task performance emerges through the interaction between cognitive processes and prior experience with comparable task demands (Giles et al., 2025). Factor 2 was more difficult to interpret because only DC demonstrated a salient loading. Although DC is classified within the original LOTCA VC domain, its performance encompasses substantially broader demands than visuoconstructive ability alone, requiring integration of semantic memory, executive planning, visuomotor organization, and familiarity with culturally learned symbolic conventions such as Arabic numerals, clock layouts, and temporal representation (Ardila, 2005; Rosselli & Ardila, 2003; Shulman, 2000). Familiarity with numerals, clock layouts, and representations of time is acquired through education and everyday participation, making clock drawing fundamentally dependent on a culturally shaped representational system (Ardila, 2005; Rosselli & Ardila, 2003). Independent loading of DC therefore suggests that successful performance may rely not only on visuoconstructive ability but also on access to culturally shared symbolic knowledge. OC and FG did not achieve salient loading on either factor. Findings from the qualitative analysis suggest that performance on these subtests may depend less on the application of learned organizational rules and more on the interaction between stimulus characteristics and prior experiential knowledge. Successful performance requires that visual stimuli evoke sufficiently familiar object representations and meanings. Variability in responses to OC and FG therefore may reflect differences in object familiarity, cultural experience, and interpretation of visual stimuli rather than a common latent cognitive process. Similar concerns have been raised in cross-cultural assessment research, which emphasizes that visually identical stimuli do not necessarily carry equivalent meanings across sociocultural contexts (Harachi et al., 2006; Rosselli & Ardila, 2003). The absence of salient factor loadings for OC and FG, therefore, suggests that score variability may reflect the interaction between stimulus characteristics and experiential knowledge, rather than the targeted cognitive processes alone. Linguistic Mediation in Visual Task Performance Qualitative findings suggest that language played a substantial role in performance on LOTCA visual subtests. Visual recognition is shaped by representational systems through which individuals learn to categorize, name, and interpret the world (Ardila, 2005; Rosselli & Ardila, 2003). In the present study, language appeared to organize visual interpretation by shaping which stimulus features were treated as meaningful and which categorical distinctions were recognized as relevant. For example, participants frequently used the Taiwanese Southern Min expression sì-kak-hîng (四角形, four-corner shape), grouping squares and rectangles under a shared category based on corner topology rather than side-length distinctions. Participants also occasionally named objects according to their function rather than their conventional label (e.g., referring to a hammer as a tool for pounding), suggesting that some responses deviated from scoring expectations because the assessment assumed a different system of categorical conventions. Response process validity is central to interpreting such findings (AERA et al., 2014). When examinees’ language organizes visual categories differently from the classificatory logic embedded in the assessment, score variation may reflect representational mismatch rather than cognitive limitation. The mismatch is not only between examiner and examinee, but also between the examinee’s culturally shaped language system and the linguistic assumptions built into the instrument. Standardized cognitive assessments implicitly assume that object names, categorical distinctions, and visual representations carry shared meanings (Harachi et al., 2006; Rosselli & Ardila, 2003). When these assumptions are not fully shared, LOTCA visual subtest scores should be interpreted with attention to the linguistic and cultural contexts, through which examinees make sense of task demands. Experiential Familiarity and Stimulus Interpretation If language shapes how individuals categorize what they perceive, lived experience influences whether a visual stimulus activates a recognizable object representation. Lower expert agreement for FG and OC, together with weak factor alignment for these subtests, suggests that several visual stimuli did not function as a shared interpretive foundation across participants (Harachi et al., 2006; Lund et al., 2014). Direct visual translation of test stimuli does not guarantee equivalent cultural meaning. A perceptually clear image may still be experientially unfamiliar, and object recognition may depend as much on everyday life experience as on perceptual acuity (Harachi et al., 2006; Rosselli & Ardila, 2003). Qualitative findings illustrated this pattern across subtests. Several responses reflected substitution with more familiar object representations, function-based naming, or reliance on isolated visual features rather than perceptual inaccuracy alone. Expert feedback similarly indicated that some stimuli, including agricultural tools and traditional object images, may not provide equally familiar interpretive anchors across generations and living contexts in Taiwan. These findings suggest that object recognition scores may be shaped by the encounter between stimulus properties and the examinee’s personally and culturally grounded knowledge of the world (Harachi et al., 2006; Rosselli & Ardila, 2003). Demographic Variation in Classification and Spatial Organization Tasks Ordinal logistic regression analyses identified significant demographic associations with selected LOTCA visual subtests after Bonferroni correction. Age was negatively associated with SI (B = −.076, Wald = 11.419, p = .001) and TM (B = −.106, Wald = 17.476, p < .001), and gender was significantly associated with TM (B = 2.703, Wald = 13.227, p < .001). Educational attainment was not significant after correction, although uncorrected associations were observed for SI (p = .012) and FG (p = .015). Demographic associations were concentrated in SI and TM, the two subtests with the highest model explanatory power (Nagelkerke R 2 = .630 and .582, respectively). Both subtests rely on abstract geometric forms rather than representations of everyday objects. Performance therefore depends on formal systems of geometric classification and spatial organization that are typically acquired through education. SI requires identification of shapes according to standardized geometric categories, whereas TM requires organization of geometric elements according to a structured spatial model. By contrast, OC and FG demonstrated substantially weaker demographic associations (Nagelkerke R 2 = .038 and .189, respectively), suggesting that performance on these tasks may depend more on experiential object familiarity and stimulus interpretation than on formally acquired classification systems. Qualitative findings further supported this interpretation. In SI, participants with lower educational attainment more frequently grouped squares and rectangles as four-corner shapes, reflecting a coherent but educationally divergent classification strategy. Age-related variation in SI may therefore partly reflect generational differences in educational exposure to standardized geometric categories rather than perceptual decline alone (Rosselli & Ardila, 2003). TM performance was associated with both age and gender. No participant achieved the maximum score of 5, and most participants (67.3%) received a score of 4, indicating that successful construction through trial and error was the predominant outcome across the participants. The observed gender difference therefore appeared to reflect successful completion through trial and error rather than the use of a trial-and-error strategy itself, with male participants receiving a score of 4 more frequently than female participants (86.1% vs. 56.9%). The source of this difference cannot be determined from the present data and should be interpreted cautiously pending replication. Age-related variation showed a similar pattern, with older participants more often receiving incomplete construction scores. Qualitative findings indicated that participants over 50 years of age more frequently engaged in repeated model comparison, cautious positional matching, and active material manipulation. Such behaviors may reflect a careful, deliberative approach involving repeated checking before committing to a response, rather than difficulty organizing visual information (Katz et al., 1997). Significant associations in demographics were found in subtests requiring formal geometric classification and spatial organization, highlighting the value of subtest-level interpretation when assessing individuals from diverse cultural, educational, and generational backgrounds. Limitations and Future Directions Several limitations should be acknowledged. Convenience sampling and modest sample size may limit generalizability. The sample consisted of cognitively intact adults, restricting the range of performance and limiting direct application to clinical populations. Typical age-related conditions were not systematically examined, and the analysis focused on selected subtests rather than the full LOTCA battery. Future research should include larger and more diverse samples, clinical comparison groups, and item-level analyses to examine whether LOTCA visual stimuli function equivalently across cultural, educational, and generational groups. Clinical and Cross-Cultural Implications for Occupational Therapy Assessment LOTCA visual subtest performance reflected the interaction of cognitive processes with language, experiential familiarity, educational exposure, and culturally acquired representational systems. Apparent errors may therefore carry different meanings depending on the examinee’s linguistic repertoire, object familiarity, educational history, and response strategies. Functional descriptions, alternative categorization strategies, cautious constructional approaches, and stimulus-based misidentifications may reveal how individuals organize and assign meaning to visual information within their sociocultural context, providing clinically meaningful information beyond what scores alone convey. Cultural adaptation of cognitive assessments should extend beyond direct translation to encompass conceptual equivalence, experiential relevance, and the representational systems embedded within test stimuli and scoring criteria (Toglia & Foster, 2025). Culturally informed cognitive assessment therefore requires attention not only to obtained scores but also to how individuals understand, approach, and assign meaning to task demands. Cognitive assessment, practiced well, is not the application of a culturally neutral instrument; it is a clinical encounter that requires the examiner to hold the instrument’s assumptions alongside the examinee’s lived experience and to interpret observed performance in light of both. Conclusion LOTCA visual subtest scores among adults in Taiwan reflected socioculturally situated response processes as well as visual task demands. Score variability should not be interpreted solely as evidence of perceptual or cognitive limitation. Culturally informed interpretation may assist occupational therapists in interpreting cognitive limitations from differences in language, experience, and educationally shaped approaches to visual perception performance. Supplemental Material Supplemental Material - Psychometric Characteristics and Cultural Interpretation of LOTCA Visual Subtests Among Adults in Taiwan Supplemental Material for Psychometric Characteristics and Cultural Interpretation of LOTCA Visual Subtests Among Adults in Taiwan by Chia-Hui Hung, Ling-Jie Hung, Wan-Yi Hung, Hsiao-Ping Chiu, and Shih-Chen Fan in Hong Kong Journal of Occupational Therapy Supplemental Material - Psychometric Characteristics and Cultural Interpretation of LOTCA Visual Subtests Among Adults in Taiwan Supplemental Material for Psychometric Characteristics and Cultural Interpretation of LOTCA Visual Subtests Among Adults in Taiwan by Chia-Hui Hung, Ling-Jie Hung, Wan-Yi Hung, Hsiao-Ping Chiu, and Shih-Chen Fan in Hong Kong Journal of Occupational Therapy Supplemental Material - Psychometric Characteristics and Cultural Interpretation of LOTCA Visual Subtests Among Adults in Taiwan Supplemental Material for Psychometric Characteristics and Cultural Interpretation of LOTCA Visual Subtests Among Adults in Taiwan by Chia-Hui Hung, Ling-Jie Hung, Wan-Yi Hung, Hsiao-Ping Chiu, and Shih-Chen Fan in Hong Kong Journal of Occupational Therapy Acknowledgements We thank all participants for their time and involvement in this study. Funding: The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was partly funded by National Science and Technology Council, Taiwan (114-2813-C-040-076-H and 113-2410-H-040-003-MY3). The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article. Supplemental Material: Supplemental material for this article is available online. ORCID iDs Chia-Hui Hung https://orcid.org/0000-0002-8339-8565 Ling-Jie Hung https://orcid.org/0009-0008-0888-2497 Wan-Yi Hung https://orcid.org/0009-0000-1834-9452 Shih-Chen Fan https://orcid.org/0000-0002-8875-6157 Hsiao-Ping Chiu https://orcid.org/0000-0002-4810-6440 Ethical Considerations The study was approved by the Institutional Review Board of Chung Shan Medical University Hospital. References Associated Data Supplementary Materials Supplemental Material - Psychometric Characteristics and Cultural Interpretation of LOTCA Visual Subtests Among Adults in Taiwan Supplemental Material for Psychometric Characteristics and Cultural Interpretation of LOTCA Visual Subtests Among Adults in Taiwan by Chia-Hui Hung, Ling-Jie Hung, Wan-Yi Hung, Hsiao-Ping Chiu, and Shih-Chen Fan in Hong Kong Journal of Occupational Therapy Supplemental Material - Psychometric Characteristics and Cultural Interpretation of LOTCA Visual Subtests Among Adults in Taiwan Supplemental Material for Psychometric Characteristics and Cultural Interpretation of LOTCA Visual Subtests Among Adults in Taiwan by Chia-Hui Hung, Ling-Jie Hung, Wan-Yi Hung, Hsiao-Ping Chiu, and Shih-Chen Fan in Hong Kong Journal of Occupational Therapy Supplemental Material - Psychometric Characteristics and Cultural Interpretation of LOTCA Visual Subtests Among Adults in Taiwan Supplemental Material for Psychometric Characteristics and Cultural Interpretation of LOTCA Visual Subtests Among Adults in Taiwan by Chia-Hui Hung, Ling-Jie Hung, Wan-Yi Hung, Hsiao-Ping Chiu, and Shih-Chen Fan in Hong Kong Journal of Occupational Therapy