Smaragda Κazi

Prof. Smaragda Κazi

orcid

Research Focus & Metrics

Academic Career Length

4Years (2023-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
31 December 2023 in Psychology: the Journal of the Hellenic Psychological Society
Verified
Journal Article
Open access
PDF file
ZBZoe Bablekou
Zoe Bablekou
Elisavet ChrysochoouElisavet Chrysochoou
Elisavet Chrysochoou
SΚSmaragda Κazi
Smaragda Κazi
Despite the established interplay between executive functioning and metacognition, evidence remains limited regarding children’s metacognitive monitoring in the form of offline performance estimates of their executive efficiency. Moreover, we know little about the relations among listening comprehension, executive functions, and metalinguistic awareness in childhood. The present study constitutes a first exploration of developmental profiles related to executive functioning, listening comprehension, and corresponding metacognitive-metacomprehension monitoring in periods marking their emergence and interplay. Participants were 95 preschoolers (Mage 5.5), and 109 second graders (Mage = 7.5). Executive function (i.e., working memory, inhibition, and cognitive flexibility) and a listening comprehension task were administered, along with offline performance estimates following completion of each task. Fluid and crystalized intelligence tests were also administered. Cluster analyses were applied. Across ages, one cluster was characterized by lower cognitive performance, but overestimated metacognitive judgements, while a second cluster demonstrated the exact opposite pattern. Older children performed at a higher cognitive level. Yet, metacognitive monitoring did not improve as a function of age, in the sense that a third, metacognitively accurate cluster did not emerge among second graders. Working memory and listening comprehension, along with their metacognitive-metacomprehension estimates, demonstrated the greatest contribution to the cluster solution in both age groups. The inhibition-metainhibition measures proved stronger profile determinants in the older group. The fluid intelligence measure showed a reduced contribution with age and vocabulary proved to be a weak and insignificant distinctive characteristic of the profiles emerging among preschoolers and second graders, respectively. Findings are discussed in relation to relevant research data and theoretical suggestions regarding cognitive and metacognitive development.

1 October 2026 in Journal of Intelligence
Verified
Journal Article
Open access
PDF file
Andreas DemetriouAndreas Demetriou
Andreas Demetriou
GSGeorge Spanoudis
George Spanoudis
EKElena Kazali
Elena Kazali
NMNikolaos Makris
Nikolaos Makris
SΚSmaragda Κazi
Smaragda Κazi
AMAntigoni Mouyi
Antigoni Mouyi
In developmental research, it is assumed that development is ergodic, i.e., differences between children reveal how individuals change over time. Across four longitudinal studies, we examined ergodicity across cognitive systems—including attention, working memory, reasoning, and cognizance—timescales and developmental phases. We found that development was not globally ergodic. In Study 1, spanning ages 4–17, cross-sectional relations partially resembled patterns of change. Individual development involved overlapping, process-specific periods of acceleration, consolidation, and differentiation rather than one common developmental timetable. Study 2, spanning ages 4–12, showed short-term convergence and partial profile correspondence, with relational integration as the main carrier of change. Study 3, spanning ages 6–13, revealed that psychometric organization was highly stable across three waves, but children changing broadly in the first annual interval were not the same as children changing in the second. In Study 4, 5–7-year-old children were intensively trained on attention control and examined three times on other processes. The trained processes formed a recognizable common organization, but within-person coupling was substantially weaker than between-person coupling, rank correspondence declined with temporal distance between waves, and group dynamics represented individual children only approximately. Across phases, the carriers of change shifted from attention-related control, through relational integration, toward complex reasoning. Psychometric g, although strong across time, was very weakly related to developmental g (i.e., phase-sensitive mechanisms transforming states of psychometric g over time), indicating that population-level cognitive organization does not determine individual development. Implications for psychometric and developmental theories are discussed here.

1 July 2025 in Psychology: the Journal of the Hellenic Psychological Society
Verified
Journal Article
Open access
PDF file
SΚSmaragda Κazi
Smaragda Κazi
Alexandra KarousouAlexandra Karousou
Alexandra Karousou
AMAngelos Markos
Angelos Markos
DEDimitra Economacou
Dimitra Economacou
TBTryfonas Bikos
Tryfonas Bikos
NMNIKOLAOS MAKRIS
NIKOLAOS MAKRIS
A large body of research highlights the important role of both executive functions (EF) and metacognitive monitoring (MM) in children's cognitive functioning in general, and in their performance on specific cognitive tasks in particular. Recent approaches suggest that EF and MM constitute different expressions of cognitive self-regulation. Despite the theoretical interest, there has been limited research on the complex interactions between EF, MM, and cognitive performance. In this paper, we study the relationship between EF and MM and their association with cognitive performance in mathematical and spatial problem-solving while taking into account children's age and performance in a fluid intelligence task. Specifically, based on a sample of 277 children aged 4 to 11 years, two alternative hypotheses are explored: (1) Cognitive performance mediates the relationship between EF and MM, and (2) MM mediates the effect of EF on cognitive performance. The results highlight the relationship between some aspects of these two theoretical constructs and their effect on children's cognitive performance. MM emerged as a significant mediator in the relationship between EF and cognitive performance. The alternative hypothesis was not supported. Results are discussed on the basis of previous studies on the development of metacognitive monitoring and executive functions.

2 July 2026 in Preprints.org
Verified
Preprint
Open access
PDF file
Andreas DemetriouAndreas Demetriou
Andreas Demetriou
GSGeorge Spanoudis
George Spanoudis
EKElena Kazali
Elena Kazali
ASAndreas Savva
Andreas Savva
NMNIKOLAOS MAKRIS
NIKOLAOS MAKRIS
SΚSmaragda Κazi
Smaragda Κazi
We compared four Large Language Models (LLMs; ChatGPT, Grok, Gemini, DeepSeek) with humans on tests of cognitive development, assessing relational integration, linguistic awareness, general and domain-specific reasoning, and cognitive self-awareness to specify how LLMs compare with humans along cognitive development hierarchies. LLMs also discussed how Descartes’s Cogito applies to them and rated themselves on aspects of Artificial General Intelligence (AGI). Hence, we propose a novel interdisciplinary comparison between human and LLM capabilities integrating developmental, cognitive, and psychometric psychology. Overall, the structure of processes in humans and LLMs was highly similar. All LLMs attained perfect linguistic and metalinguistic performance. ChatGPT and Gemini outperformed university students in mathematics and causal reasoning. Grok performed slightly and DeepSeek considerably lower. All LLMs underperformed in visual–spatial tasks. Self-evaluation profiles broadly mirrored performance profiles: ChatGPT and Grok rated themselves high in reasoning and low in visualization, Gemini inflated visualization by reframing it as linguistic creativity, and DeepSeek consistently underrated itself. Each LLM restated Descartes’s Cogito differently, reflecting its own priorities, and denied having high AGI. Therefore, LLMs display human-like “subjective” task scaling implying algorithmic or functional metacognition, capturing their architectural gap between symbolic reasoning and visuospatial cognition, but they were modest in claiming top human intelligence. Implications for an integrated natural-artificial intelligence theory are discussed. Also, a developmental engineering model is sketched that might allow removing limitations of each LLM.

17 September 2026 in AI
Verified
Journal Article
Open access
PDF file
Andreas DemetriouAndreas Demetriou
Andreas Demetriou
GSGeorge Spanoudis
George Spanoudis
EKElena Kazali
Elena Kazali
ASAndreas Savva
Andreas Savva
NMNIKOLAOS MAKRIS
NIKOLAOS MAKRIS
SΚSmaragda Κazi
Smaragda Κazi
We compared four large language models (LLMs; ChatGPT, Grok, Gemini, and DeepSeek) with humans on tests of cognitive development, assessing relational integration, linguistic awareness, general and domain-specific reasoning, and cognitive self-awareness to specify how LLMs compare with humans along cognitive development hierarchies. LLMs also discussed how Descartes’s Cogito applies to them and rated themselves on aspects of Artificial General Intelligence (AGI). Accordingly, we propose a novel interdisciplinary comparison of human and LLM capabilities that integrates developmental, cognitive, and psychometric psychology. Overall, the processes in humans and LLMs were highly similar. All LLMs attained perfect linguistic and metalinguistic performance. ChatGPT and Gemini outperformed university students in mathematics and causal reasoning. Grok performed slightly better and DeepSeek considerably worse. All LLMs underperformed in visual–spatial tasks. Self-evaluation profiles broadly mirrored performance profiles: ChatGPT and Grok rated themselves highly in reasoning and low in visualization, Gemini inflated visualization by reframing it as linguistic creativity, and DeepSeek consistently underrated itself. Each LLM restated Descartes’s Cogito differently, reflecting its own priorities, and denied having high AGI; these self-characterizations were generally stable about a year later. Therefore, LLMs displayed “subjective” task scaling, implying algorithmic or functional self-monitoring, capturing their architectural profile of performance, but they were modest in claiming above-human intelligence. We discuss implications for an integrated natural–artificial intelligence theory. We also sketch a developmental engineering model that might remove the limitations of each LLM.

4 November 2025 in Preprints.org
Verified
Preprint
Open access
PDF file
Andreas DemetriouAndreas Demetriou
Andreas Demetriou
GSGeorge Spanoudis
George Spanoudis
EKElena Kazali
Elena Kazali
ASAndreas Savva
Andreas Savva
NMNIKOLAOS MAKRIS
NIKOLAOS MAKRIS
SΚSmaragda Κazi
Smaragda Κazi
We compared four large language models (ChatGPT, Grok, Gemini, DeepSeek) with humans in reference to tests of cognitive development addressed to relational integration, linguistic awareness, general and domain-specific reasoning, and cognitive self-awareness. We aimed to specify how LLMs compare with humans along several cognitive development hierarchies. Given their theoretical importance for intelligence, LLMs were also asked to indicate how Descartes’s Cogito applies to them and self-rate on aspects of Artificial General Intelligence (AGI). There was a huge divide between verbal and logico-mathematical tasks, on the one hand, and visuo-spatial tasks, on the other hand. All LLMs attained perfect linguistic and metalinguistic performance. ChatGPT and Gemini matched or exceeded university-level human performance in mathematics and causal reasoning, Grok performed slightly lower, and DeepSeek weakest overall. All LLMs underperformed in visual–spatial tasks or reasoning tasks when shown visually as presented to children. Performance recovered when these tasks were presented in a fashion allowing LLMs to employ an analytical approach to visual patterns, signifying their unique architecture. Self-concept ratings broadly mirrored performance profiles: ChatGPT and Grok rated themselves high in reasoning and low in imagination, Gemini inflated imagination by reframing it as linguistic creativity, and DeepSeek consistently underrated itself. Each LLM restated Descartes’s Cogito differently as a description of itself and denied having much AGI. Hence, LLMs display human-like “subjective” task scaling implying algorithmic or functional metacognition, which captures the architectural gap between symbolic reasoning and imaginative cognition, but they are modest in claiming top human intelligence. Overall, LLMs display "savant-like intelligence" rather than top expert intelligence. Implications for an integrated natural-artificial intelligence theory are discussed. Also, a developmental engineering model is sketched that would allow removing limitations of each LLM.