Julio-Diciembre 2015
Vol. 5 No. 2
Octubre-diciembre 2024
Vol. 14 No. 3
Interacción y Perspectiva. Revista de Trabajo Social Vol. 14 No3 / octubre-diciembre, 2024
620
Interacción y Perspectiva Dep. Legal pp 201002Z43506
Revista de Trabajo Social ISSN 2244-808X
Vol. 14 No3 620-646 pp. Copyright © 2024
Octubre-diciembre
ARTÍCULO DE INVESTIGACIÓN
Aprendizaje basado en las preferencias humanas: un estudio piloto sobre la
percepción de los estudiantes sobre el uso de la IA y ChatGPT
DOI: https://doi.org/10.5281/zenodo.11154746
Dalina Dumitrescu *
Resumen
El estudio examinó la conciencia, comprensión y actitudes de los estudiantes hacia la
inteligencia artificial (IA) y su impacto en la gestión del proceso educativo. Se utilizó una
encuesta con preguntas abiertas. El objetivo era analizar la percepción de los
encuestados sobre la importancia, uso e impacto de la IA, incluyendo ChatGPT, en el
ámbito laboral y educativo en el que participan. Se aplicó a 25 cursantes de maestría en
la Facultad de Finanzas y Banca de la Universidad de Estudios Económicos de Bucarest.
Las conclusiones se basaron en el procesamiento de las respuestas abiertas utilizando
una herramienta de IA. Los hallazgos revelaron que todos los encuestados eran
conscientes de la presencia de la IA. Las principales fuentes de información sobre la IA
fueron el lugar de trabajo, las redes sociales y los amigos. El análisis del sentimiento de
las respuestas mostró un coeficiente positivo más alto con respecto al impacto de la IA
en la educación. Aunque se reconoce que este estudio piloto tiene limitaciones, las
principales conclusiones indican que los estudiantes están receptivos a la sociedad
digitalizada. El uso de plataformas como ChatGPT y software de IA en actividades
prácticas es limitado. Valoran con cautela la experiencia de trabajar con generadores de
texto de IA. El predominio de sentimientos positivos relacionados con el uso de ChatGPT
tanto en el lugar de trabajo como en las actividades educativas proporciona una base
sólida para la integración formal, coherente, eficiente y participativa de la IA en enfoques
y procesos educativos responsables.
Palabras clave: educación superior, IA; impacto de la IA; ChatGPT; Digitalización y
gestión educativa.
Abstract
Learning based on human preferences: A pilot study regarding the student's
perception of the AI and the use of ChatGPT
The study examined the awareness, understanding, and attitudes of students towards
artificial intelligence (AI) and its impact on educational management. A survey with
open-ended questions was used for data collection. The objective was to analyze the
respondents' perception of the importance, use, and impact of AI, including ChatGPT, in
their workplace and educational settings. It was administered to 25 master's students
Dalina Dumitrescu / Aprendizaje basado en las preferencias humanas: un estudio piloto sobre la percepción
de los estudiantes sobre el uso de la IA y ChatGPT
621
at the Faculty of Finance and Banking, University of Economic Studies in Bucharest. The
conclusions were based on processing the open-ended responses using an AI tool. The
findings revealed that all respondents were aware of the presence of AI. The main
sources of information about AI were the workplace, social media, and friends. The
sentiment analysis of the responses showed a higher positive coefficient regarding the
impact of AI on education. While acknowledging the limitations of this pilot study, the
main conclusions indicate that students are receptive to the digitalized society. The use
of platforms like ChatGPT and AI software in practical activities is limited. They cautiously
value the experience of working with AI text generators. The predominance of positive
sentiments related to the use of ChatGPT in both the workplace and educational activities
provides a solid foundation for the formal, coherent, efficient, and participatory
integration of AI into responsible educational approaches and processes.
Keywords: higher education, AI; impact of AI; ChatGPT; Digitization and education
management.
Recibido: 18/03/2024 Aceptado: 29/04/2024
* Universidad de Estudios Económicos de Bucarest, Rumania, ORCID ID: https://orcid.org/0000-0002-4633-
1791 E-mail: dalina.dumitrescu@fin.ase.ro
1. Introduction
In the last 10 years, artificial intelligence (AI) has developed a potentially disruptive
impact on all human society. Artificial Intelligence (AI) refers to reproducing human
ability to get and use knowledge and skills by machines, particularly computer systems.
These include learning (acquiring knowledge and skills through experience), reasoning
(concluding available information), problem-solving (finding solutions to complex
challenges), and decision-making (selecting the best course of action based on analysis).
AI aims to create systems that mimic human cognitive abilities, enabling them to
perform activities that typically require human intelligence. Here are some key areas
where AI has made and continues to make a significant impact:
TABLE Nro. 1
KEY AREAS WHERE AIHAS A SIGNIFICANT
Application Area
References
Key Points
Industries
Revolutionization via
AI Automation
Jarret et al.
(2021), Sarmah &
Shekhar (2019)
Streamlining processes, reducing human
intervention, and enhancing efficiency
(Automation and Efficiency)
AI in Medicine
Saraswat et al.
(2022), Yu et al.
(2018)
Assisting in medical diagnosis, drug
discovery, personalized treatment plans,
and predictive analytics (Healthcare)
AI in Finance
Milana et al.
(2021), Shamima
et al. (2022)
Analyzing financial data, informing trading
decisions, detecting fraud, and predicting
market trends (Finance)
Interacción y Perspectiva. Revista de Trabajo Social Vol. 14 No3 / octubre-diciembre, 2024
622
AI in Manufacturing
Nti et al. (2022),
Yao et al. (2017)
Optimizing production processes,
enhancing quality control, and enabling
predictive maintenance of machinery
(Manufacturing)
AI in Transportation
Prashant et al.
(2022), Bharadiya
(2023)
Revolutionizing transportation, reducing
accidents and congestion, and improving
efficiency (Transportation)
AI in Entertainment
Fu et al. (2002),
Nader et al.
(2022)
Content recommendation, computer-
generated graphics, animations, and
enhanced video game experiences
(Entertainment)
AI in Natural
Language Processing
Chowdhary
(2020), Raina et
al. (2022)
Enabling chatbots, virtual assistants,
accurate language translation, and
revolutionizing interactions with
technology (Natural Language Processing)
AI in Global
Connectivity
Leyva-Mayorga et
al. (2022),
Akhavan et al.
(2020)
Facilitating communication across
language barriers, fostering global
collaboration and understanding (Global
Connectivity)
AI in Environmental
Sustainability
Tsolakis et al.
(2022), Feroz et
al. (2021)
Climate modeling, analyzing environmental
data, optimizing energy consumption, and
managing natural resources efficiently
(Environmental Sustainability)
AI in Personalization
in Marketing
Chandra et al.
(2022), Gao et al.
(2022)
Analyzing consumer behavior, offering
personalized product recommendations,
and executing targeted marketing
campaigns (Personalization in Marketing)
AI in Cybersecurity
Ansari et al.
(2023), Zhang et
al. (2022)
Detecting cyber threats, identifying
anomalies, and strengthening security
measures against data breaches and
attacks (Cybersecurity)
AI in Agriculture
Liu (2020),
Sharma (2021)
Precision farming, optimizing irrigation,
monitoring crop health, and predicting
yield outcomes (Agriculture)
AI in Social Impact
Shangyao et al
(2007), Linardos
et al. (2022),
Lee et al. (2023),
Chiou et al. (2022)
Significant contributions in disaster
response, analyzing large-scale data
during emergencies, and assisting in
resource allocation for relief efforts (Social
Impact)
AI in Education
Alam (2022)
Liu et al. (2023)
Sung-Hee et al.
(2023)
Transforming education through
personalized learning experiences,
intelligent tutoring systems, automated
grading, and adaptive learning platforms
(Education)
Source: constructed by the author in dialog with ChatBPT (2023)
In essence, AI's significance lies in its ability to process and analyze massive volumes
of data, make predictions, and perform tasks that were previously considered beyond
Dalina Dumitrescu / Aprendizaje basado en las preferencias humanas: un estudio piloto sobre la percepción
de los estudiantes sobre el uso de la IA y ChatGPT
623
the capabilities of machines. It is reshaping industries, enhancing human capabilities,
and driving innovation in countless ways.
While the transformative impact of AI has been largely positive, there are also
concerns about ethical considerations, job displacement, privacy issues, and biases in AI
algorithms. Balancing these challenges while harnessing AI's potential for positive
change is a critical ongoing endeavor.
2. Specific background information
In the last 15 years, the main stakeholders and quality accreditation agencies for
higher education had a specific objective to assess the introduction of technology in the
education process, in general, and at the level of individual courses in special Bakir
(2016), Hodges et al. (2022)
Incorporating technology into the classroom has been a positive step forward for
education. Using laptops and dedicated sites to store documents, searching the web for
resources, and utilizing statistical software such as Excel, Stata, R, and Python have all
been beneficial in the social sciences. Additionally, simulations and games can help
transfer knowledge in an engaging way. As we move towards in-person learning again,
it's important to consider how we can continue to integrate technology effectively and
thoughtfully while preserving the value of face-to-face interaction. The adoption and the
extent of implementation of education technology were highly related to the resources
of the university (approved budget to buy the technological infrastructure in the social
fields where the teaching process could have been realized in a classical way without
technology). The ability and openness of teachers to identify, design, and adopt new
elements of educational technology in the classroom is a great challenge, especially
where teachers' area of expertise is not modern technology.
The historical development of AI in education reflects a journey from early theoretical
concepts to practical applications that are transforming how learning is delivered,
monitored, and personalized. As AI technologies become more sophisticated, the
potential for enhancing educational experiences and outcomes continues to expand.
Table 2. Summarizes the key developments in AI's role in education across different
decades.:
TABLE Nro. 2
AI'S ROLE IN EDUCATION ACROSS DIFFERENT DECADES
Key Developments
Examples and Applications
1950s-1960s
Emergence of AI as a field of
study, Turing Test proposed by
Alan Turing
Early programs: "Logic Theorist,"
"Geometry Theorem Prover"
1970s-1980s
Development of knowledge-
based systems and expert
systems, MYCIN aids in
diagnosing infectious diseases
Cognitive science and AI research
influence intelligent tutoring systems
(ITS), e.g., Soar, ACT-R
Interacción y Perspectiva. Revista de Trabajo Social Vol. 14 No3 / octubre-diciembre, 2024
624
1990s
Rise of Intelligent Tutoring
Systems (ITS), examples like the
Geometry Tutor and AutoTutor
World Wide Web emergence facilitates
web-based learning, AI used for adaptive
e-learning environments
2000s
Advances in machine learning
and data analytics enhance ITS
capabilities, educational games
and simulations with AI
integration
Games like "Zoombinis" and "Math
Blaster" incorporate AI for engaging
learning experiences
2010s
Rise of Massive Open Online
Courses (MOOCs) with AI scaling
instruction, automated
assessments, and personalized
recommendations
Introduction of chatbots and virtual
assistants like IBM's "Jill Watson" for
student assistance
2020s and
Beyond
Continued role of AI in education,
supporting adaptive learning,
learning analytics, and real-time
feedback systems
Natural language processing and
sentiment analysis gauge student
emotions and engagement levels,
addressing ethical concerns
2022
November 30, 2022 Chat GPT
was launched
ChatGPT employs deep learning, a
branch of machine learning, to generate
text that closely resembles human
language using transformer neural
networks. The transformer model
anticipates text outcomes, such as the
following word, sentence, or paragraph,
by leveraging patterns observed in its
training data.
The training process initiates with
general data and progressively
transitions to more specialized data
aligned with a particular task. ChatGPT
initially undergoes training with diverse
online text to grasp the nuances of
human language. Subsequently, it
refines its conversational skills by
learning from transcripts, honing its
ability to engage in meaningful
dialogues.
Source: Constructed based on literature review by the author in dialog with
ChatBPT (2023)
The limitations and the new rules imposed by the COVID crisis forced the adoption
of technology in the education process to a new level of adoption and implementation.
Given the forced and sudden lockdown of the entire social life, to maintain the
educational process uninterrupted during the pandemic, the universities looked to the
experience of the total online programs as a rescue solution to accomplish their mission.
Dalina Dumitrescu / Aprendizaje basado en las preferencias humanas: un estudio piloto sobre la percepción
de los estudiantes sobre el uso de la IA y ChatGPT
625
The aim was to ensure for the graduates the exposure and experience regarding the
whole curriculum, the core professional set of knowledge, skills, and attitude as close as
possible to normal time. In other words, to attain learning goals and objectives as in the
face-to-face educational process. Starting with 2020 a lot of debates were in the main
flow of literature related to the pros and cons related to in-person and virtual learning,
As a result, Universities struggled to build strong virtual learning communities as new
online education models are going beyond the replication of face-to-face activities, see
Webinar: Unlocking the Power of AI: How Tools Like ChatGPT Can Make Teaching Easier
and More Effective-HBP, May 2, 2023.
However, the COVID pandemic did not mean only the adoption of online education
models with already calibrated technology that coexisted previously with face-to-face
ones. Technology has advanced and opened generously to the public with tools, models,
and products, especially in the area of artificial intelligence (AI).
The tools and products offered on a large scale and often free of charge by AI, and
recently developed Artificial Intelligence Text Generators (AITG), with application for
research work and implicitly also in educational activity (reports, essays, finalization
works), they represent completely new facilities with profound implications, Bahroun
(2023).
Using AI, the online research of information and documentation sources is much
simplified the part of conception, structuring and elaboration of research works (essays,
final papers doctoral thesis) including the innovation and personalization components is
done automatically. Finding automatic answers to the questions to quizzes or field exams
including science exams is more and more accurate. Thus, the human factor's
contribution and work based on his/her inherent knowledge, skills, and abilities is no
longer relevant.
The tasks to be realized by the students such as admission essays, research papers,
field essays and reports, innovative code lines writing, and final papers (such as
undergraduate, dissertation, and Ph.D. thesis) supposed to the personal contributions
based on the accumulated knowledge, skills, and attitudes. The individual contributions
and the personal innovative approach were built by educational experiences at courses
and trans-curricular levels and were the base for assessing the learning and grading,
Holland (2023).
The code of ethics in all the universities states strong rules for the ethics in using
online resources and plagiarism (similitude check).
In this aim, in many universities introduced for checking the effective enforcement
of integrity rules in the academic intellectual contributions and assessing the knowledge,
different systems (such as shields in place to limit the access of the students to the web
pages for the in-campus exams) or antiplagiarism softs- similitude check (for the final
papers).
In the current period, an extremely important question is to what extent AI should
be allowed, adopted, or limited/reduced/prohibited in the process of building and
Interacción y Perspectiva. Revista de Trabajo Social Vol. 14 No3 / octubre-diciembre, 2024
626
consolidating the fund of knowledge and skills for young people, in the process of forming
the professional bases of independent work.
ChatGPT was released in 2019 by OpenAI, followed by an updated version of this
conversational chatbot in late November 2022 based on Reinforcement Learning with
Human Feedback (RLHF).
ChatGPT is algorithm-driven and aims to generate human-like text that can be used
in conversation. As a functionality, it can answer questions, provide information, and
participate in natural language discussions with users. ChatGPT is used as a customer
service chatbot, virtual assistant, and conversational interface for websites or mobile
applications. It can also be used to generate content for social media or create chatbot
scripts for marketing or entertainment purposes. It is a powerful tool for creating
chatbots that can have intelligent and engaging conversations with users, - see Webinar-
Will AI Replace the Educator- HBP August 10, 2023,
As for educational applications, it is mentioned by (Hwang et al., 2020) that there
are several roles of AI in education, such as serving as an intelligent tutor, guidance,
learning tool/policy partner, or advisor.
Chen et al. (2020) found that research in the field of Artificial Intelligence in
Education AIEd rapidly developed around 2012, and the number of articles, grants, and
citations in the field is still increasing. The significant increase in citations received by
AIEd studies demonstrates the wider impact and influence of AIEd research in academia.
The findings are consistent with those from the main flow of research results (e.g., Chen
et al., 2022, Chu et al., 2022, Chiu et al., 2023)
Recent research studies have significantly contributed to the understanding of how
AIEd and technology can be effectively integrated into educational settings, fostering
personalized learning experiences, improving learning outcomes, and addressing
important ethical and practical considerations.
As we transition back to face-to-face learning, it's important for us to consider how
we can incorporate the benefits of virtual learning communities into our in-person
education. We need to ask ourselves what worked well in the online format and how we
can integrate those elements into the traditional classroom experience. At the same
time, we should also reflect on what we value about face-to-face learning and make sure
we preserve those elements. It's a balancing act, but by taking a thoughtful and
intentional approach, we can create a truly dynamic and effective learning environment.
Important works in educational data mining and learning analytics contributed to
research on using data to understand and improve learning processes (Ryan et al.,
2016).
Another important research line is intelligent tutoring systems, affective computing,
and human-computer interaction in education, (Woolf et al., 2009), the development of
cognitive tutors and intelligent tutoring systems mainly applies cognitive science
principles to education (Aleven et al., 2006).
Dalina Dumitrescu / Aprendizaje basado en las preferencias humanas: un estudio piloto sobre la percepción
de los estudiantes sobre el uso de la IA y ChatGPT
627
3. Description of the gap in our knowledge that the study was designed to fill
Mijwil et al. (2020), studied the impact of artificial intelligence applications, their
implication in learning, and how they can be employed in the future development of the
teaching/learning process. They concluded that artificial intelligence is a key factor of
growth in future digital societies and must be properly exploited to build a new world
where future graduates must be competitive.
Ömer Osmanoǧlu et al. (2020), collected student feedback from a university was
utilized to evaluate course materials. The machines were employed to categorize the
materials based on positive, negative, or neutral sentiments regarding each teaching
resource. Subsequently, the course materials were enhanced by addressing the negative
feedback.
Nikolić et al. (2020), started from student course assessment conducted at sentence-
level analysis to identify one or more aspects in the sentences and classify their polarity
into positive or negative sentiment.
Misuraca et al. (2021), Gkontzis et al. (2020) concluded that in an educational
context, Opinion Mining allows processing students’ comments and creating powerful
analytics.
Shaik et al. (2023), concluded that through advancements in sentiment techniques
and AI methodologies, student comments can depict their sentiment orientation with
minimal human intervention. The significance of emotional analysis in education was
examined across four levels: document level, sentence level, entity level, and aspect
level. In conjunction with sentiment annotation techniques, the study delved into the
role of AI in sentiment analysis using methodologies such as machine learning, deep
learning, and transformers. The findings highlight the influence of sentiment analysis on
educational processes, aiming to improve pedagogy, decision-making, and evaluation.
As we transition back to face-to-face learning, it's important for us to consider how
we can incorporate the benefits of virtual learning communities into our in-person
education. We need to ask ourselves what worked well in the online format and how we
can integrate those elements into the traditional classroom experience. At the same
time, we should also reflect on what we value about face-to-face learning and make sure
we preserve those elements. It's a balancing act, but by taking a thoughtful and
intentional approach, we can create a truly dynamic and effective learning environment.
How do we educate using responsible the new technology mainly the one generated by
AI and AITG specifically? The question is going beyond the simple one what the
implications of the use of laptops and smartphones during class will be? It is about
accepting or resisting using ChatGPT to do assignments and examinations. If accepted,
what to do as responsible educators is not to diminish the quality of the learning process
and attain properly the objectives assumed by the missions of our institutions.
In a recent webinar held by HBP were presented the results of a survey from June
2023 applied in multiple countries (UK, US, Germany, Spain, Australia, Nederland,
Interacción y Perspectiva. Revista de Trabajo Social Vol. 14 No3 / octubre-diciembre, 2024
628
France) with the question “How is your school using the AI?” The 12 schools that
responded formulated the answers at 3 levels of responsibility and implication:
discussion, at the faculty level, and the school level. The main conclusion was that most
educational institutions are at the early stages of using AI in their activity. Currently,
they are in a period of exploration. The initiatives are mainly tactical and incremental,
and the main uses of AI are material creation and AI chatbots, see Webinar-
Experimenting with AI in the Classroom-Harvard School of Education- September 6,
2023
4. The study objective
The work, a pilot study, investigates to what extent the students in the master
programs are aware of artificial intelligence, from which sources they got information
about AI, and specifically about ChatGPT. We also investigated their opinion about the
impact of AI on jobs and education, if they used ChatGPT, and for what purpose. Those
questions are also the research questions of the study and the procession of the answers
to them are presented in detail. According to our knowledge, there is no study
investigating the perception and the sentiments of students regarding the AI influence
at work and in education.
The positive effects of investigating students’ opinions and sentiments are supported
by research studies. For example, Chakraborty et al. (2020) used as an instrument for
research a survey in which undergraduate students in an Indian university expressed
their opinions on online education during the pandemic. The conclusions were that the
students learn better in physical classrooms (65.9%) than through online education, and
they felt that the teachers improved their online teaching. The students gave favorable
evaluations to the software and online study materials employed as support. However,
they expressed concerns about the stress associated with online education, noting its
impact on their health and social life.
The paper is structured as follows: The methodology part will describe the approach
and the tools used for the analysis of the answers, the next two parts will present the
results and the discussion of the result. The conclusion will close the study.
5. Methodology
As presented in the introduction, most educational institutions are currently in the
early stage of incorporating AI into their instructional activities.
From this perspective, the interest arose to investigate, in the form of a pilot study,
whether at the level of the faculty where I work, the master's program students -
predominantly employees - are familiar with the concept of AI. The study aims to explore
from which sources they gather information about AI, and how they perceive the impact
of AI on their professional activities, as well as on educational endeavors. Specifically,
regarding ChatGPT, the students were questioned about their usage of it, their
Dalina Dumitrescu / Aprendizaje basado en las preferencias humanas: un estudio piloto sobre la percepción
de los estudiantes sobre el uso de la IA y ChatGPT
629
experience with working hours involving it, and the specific activities for which they
employed it. To the best of our knowledge, there is currently no such study conducted
that delves into students' opinions regarding the impact of Artificial Intelligence Text
Generators (AITG) on education, research, and the workplace.
At the end of May 2023, a questionnaire was distributed to the master's program
students at the Faculty of Finance and Banking. The questionnaire was administered in
paper format to the 25 students who were in attendance. Responses to the
questionnaire, in English, were anonymous, and the collection process was conducted
randomly. The purpose of gathering this information was to communicate to the
students, and the response rate was 100%.
The questionnaire comprises 5 questions, out of which 4 are open-ended questions
and one is quantitative. Two out of the 5 questions of the survey were inspired by the
questions addresses online to the educated participants during the Webinar: Unlocking
the Power of AI: How Tools Like ChatGPT Can Make Teaching Easier and More Effective-
HBP, May 2, 2023.
The preparation of the survey is based on literature review to ensure that the
questions are grounded in existing research and theories as well as the logic in the
investigation of the awareness and use of new concepts/products services by individuals,
In (2017).
Given the relatively short time after the release of the CHATGDP – only 6 months-
the current study was developed as a pilot study, with a small sample of respondents to
to identify potential issues with question wording, clarity, ambiguity, or issues with the
questions addressed. Conducting the pilot study was intended to save time and to obtain
valuable insights in an area where no other information was available. Some validation
conditions were observed: (a) Ensure that questions are clear, concise, and easily
understood by the target audience; (b) Ensure that questions are neutral and do not
lead respondents to a particular answer; (c) Avoid wording that might introduce bias or
influence respondents; (d) Gather feedback from respondents after the questionnaire
has been administered, Mellinger et al. (2020).
The research questions of the study are aligned with the questions of the survey as
follows: If the students are aware of artificial intelligence, and from which sources do
they get information about AI (Q1); What is their opinion about the impact of AI on
current jobs (Q2); What is their opinion about the impact of AI on education, (Q3); If
they used ChatGPT- as working hours (Q4); For what purpose they used ChatGPT (Q5)
Considering the investigated area of interest, the decision was made to utilize open-
ended questions.
The option for open-ended questions was determined by the arguments that the free-
form gives space to respondents to answer in an open-text format and the freedom to
express himself/herself in as much (or as little) detail as they prefer. Open-ended
questions help to see the respondent’s perspective, as the feedback is in their own words
instead of preformatted answers. The free format allows getting more meaningful
Interacción y Perspectiva. Revista de Trabajo Social Vol. 14 No3 / octubre-diciembre, 2024
630
answers, accurate responses, and a better way to analyze the sentiments of
respondents, Siedlecki, (2022).
The processing of responses for the open-ended questions was carried out using the
AI Monkey Learn tool, while the quantitative question was processed by the author using
Excel.
AI text analysis employs natural language processing (NLP) to automate the
classification and extraction of data from texts. This approach is highly effective,
particularly in automatically analyzing open-ended survey responses. From the site
Monkey Learn, it was extracted the relevant characteristics of the program. The machine
learning models it uses can apply several pre-trained text analysis models that can
process the survey responses very rapidly on the next dimensions (from the MonleyLearn
site):
“Sentiment Analyzer: automatically analyzes text for opinion polarity
(positive, negative, neutral).
NPS Feedback Analyzer: automatically sorts NPS responses by Ease of Use,
Features, Pricing, and Support.
Keyword Extractor: extracts the most used and most important words from your
survey responses.
Company Extractor: automatically extracts the names of businesses and
organizations from surveys or any text.”- Excerpt from Monkey Learn site.
https://monkeylearn.com/blog/survey-analysis/
Given the purpose of the pilot study and the reduced complexity of the data collected,
the survey answers were analyzed with a Sentiment Analyzer and keyword extractor.
For each open question processed with Monkey Learn, the paper presents the overall
picture of the processing of the answers as well as the graphic representation of the
focused subsections. For the quantitative question- question 4-, the graph of the results
processed by the author manually is also presented.
6. Results
The main results of the analysis of survey answers are presented in line with the
research question of the paper investigated by the survey questions. For question 1 will
be presented all the elements obtained from the AI processing: The synthesis image
(dashboard), the result from the keywords extractor (aspect count and Keywords cloud),
the result for sentiment analyzer (positive-green, neutral orange, and negative–red) and
the overall sentiment. The results of questions 2, 3 and 5 will be presented as the results
of processing for the two dimensions- keywords extractor and sentiment analyzer. The
answers for question 4, quantitative, manually processed, are presented in Fig. 10.