When the data changes, has reality changed?

Column

Data Analysis

September 30, 2026

When the data changes, has reality changed?

Analysis of nearly 20 million faculty-years between 2020 and 2025 shows how changes in population composition and record production can alter the interpretation of a historical series

Between 2020 and 2025, the number of formal teaching positions identified in RAIS (Annual Relation of Social Information), from the Ministry of Labor, increased from about 2.58 million to 4.03 million. In the same period, there was an expressive transformation in the composition of these positions: statutory ones, which were 63.1% of the total in 2020, fell to 40.2% in 2025, while non-permanent or temporary public positions rose from 16.8% to 43.8%.

The data on sick leave also draws attention. The proportion of employment relationships with registered leave increased from 4.07% in 2020 to 10.54% in 2022, and then decreased, practically returning to the starting point: 4.08% in 2025.

At first glance, it would be tempting to conclude that, after a strong increase until 2022, there was a significant improvement in the following years. When the data are examined by type of employment, however, a result that is difficult to ignore emerges. Among non-permanent or temporary public employment, the registered rate goes from 8.47% in 2022 to only 0.41% in 2025.

What happened? Was there really a reduction of this magnitude in dismissals, or did the way in which this phenomenon came to be recorded change? This question illustrates a central problem in data analysis: not every change observed in a time series necessarily represents an equivalent change in the phenomenon we intend to measure. Before interpreting a result, it is necessary to understand how the data were produced.

Behind the numbers

The results were obtained from RAIS microdata, considering five occupational families related to teaching. The analysis covers all Brazilian states and tracks the characteristics of formal employment relationships, such as occupation, age, sex, type of employment, and records of leave due to illness.

There is an important distinction: the unit analyzed is not necessarily a person, but a work relationship. The same professor can have more than one relationship and, therefore, appear more than once in the database. For this reason, 4.03 million records in 2025 do not mean 4.03 million professors, but approximately 4.03 million teaching positions.

This difference seems merely technical, but it is essential. Administrative databases are originally produced for registration and management purposes. When they are used for data analysis, it is necessary to understand the meaning of each variable and, mainly, what each record represents.

Structure Transformation

Between 2020 and 2025, the number of teaching positions considered in the analysis increased by approximately 56%, from 2.58 million to 4.03 million. More important than the absolute growth, however, is observing how the composition of these positions has changed. In 2020, statutory public positions accounted for 63.1% of the analyzed records, while non-tenured or temporary public positions represented 16.8%. In 2025, the percentages changed to 40.2% and 43.8%, respectively.

If these numbers were observed in isolation, it would be possible to interpret the period as evidence of a rapid transformation in the structure of teaching work. But precisely in this interval, an important change occurred in the production of RAIS itself.

Data source

Starting from the base year of 2023, RAIS compliance is now measured solely by extracting events reported in eSocial for all declarant groups, including public agencies. The RAIS 2023 Technical Note records a significant break in the historical series and recommends that the results from that year not be directly compared with previous years, due to the transition process in data capture methods.

This does not mean that subsequent data are invalid or that all observed changes stem from eSocial. It means that the comparison requires caution. Instead of asking only why the numbers changed, it is necessary to ask how much of this change belongs to the phenomenon studied and how much may be related to the way it began to be recorded.

Attrition Data

In 2020, 4.07% of the analyzed links showed a record of leave. The rate rises to 6.83% in 2021 and reaches 10.54% in 2022, the highest value in the series. After that, it recedes to 8.25% in 2023, 5.50% in 2024, and 4.08% in 2025. There are two temptations in this reading. The first would be to automatically associate the 2022 peak with the effects of the pandemic. The second would be to interpret the subsequent drop as evidence of improvement in teachers’ health conditions. The data used here, in isolation, do not allow us to support either of these causal conclusions. What can be affirmed is more restricted: the leave indicator recorded in the database increased sharply until 2022 and fell in the following three years.

Among statutory public positions, the registered rate of leave increased from 4.63% in 2020 to 10.68% in 2022. It then decreased to 8.19% in 2023, 7.55% in 2024, and 6.59% in 2025. Among non-effective or temporary public positions, the trend is much more pronounced: 2.44% in 2020, 8.47% in 2022, 6.47% in 2023, 1.53% in 2024, and only 0.41% in 2025.

A reduction from 8.47% to 0.41% could produce an attractive headline. But we do not know, from these numbers, if teachers with non-effective appointments became less ill. What we do know is that the frequency with which absence is recorded in these appointments has dropped drastically. The difference is fundamental.

To investigate whether the general rate drop could be explained simply by the change in the composition of the links, the variation between 2022 and 2025 was decomposed into two parts: one associated with the change in the participation of different types of links and another related to the changes in the rates recorded within these groups.

The aggregate rate fell from 10.54% to 4.08%, a reduction of 6.46 percentage points. Of this total, approximately 0.92 percentage points are associated with the change in the composition of employment relationships, while 5.54 percentage points result from changes in the rates recorded within groups. In proportional terms, about 14% of the reduction is associated with composition, and approximately 86% with within-group changes.

Mathematically, the decomposition shows that the aggregate drop did not occur solely because the share of a certain type of link increased. But a statistical technique can correctly decompose what is recorded in the database and still not solve a measurement problem. If the way data is produced changed during the period, the decomposition will also reflect this change.

In 2020, the aggregate dropout rate was 4.07%. In 2025, it was 4.08%. Looking only at the two extremes, practically nothing seems to have changed. The decomposition tells another story: the change in the composition of employment relationships contributed approximately -1.15 percentage points to the rate, while changes in internal rates contributed in the opposite direction, with approximately +1.16 percentage points. The two movements practically canceled each other out.

Thus, 4.07% and 4.08% are almost identical numbers produced by quite different structures. Aggregate indicators are useful, but they can hide important transformations in the analyzed population.

What this case teaches

The RAIS case shows that a data analysis does not start with the algorithm. It starts with understanding the base. We could have observed the drop between 2022 and 2025, applied statistical techniques, and ended the analysis with an apparently consistent conclusion. The calculations would be correct. The problem would lie in the interpretation.

This problem is not exclusive to RAIS. A new filling rule, system migration, a classification change, the inclusion of new users, or a change in the mandatory nature of a certain field can create a break that, in a graph, looks exactly like a change in behavior.

Therefore, some questions should precede any more sophisticated model: did the observed population remain comparable? Did the definitions of the variables remain the same? Did the coverage change? Was there any alteration in the system that produces the records?

More data also do not necessarily mean more information. With millions of links, small differences can be calculated with enormous precision. But numerical precision and validity of interpretation are different things. If a variable starts to be recorded in a different way, millions of records can reproduce this difference with enormous consistency. Before asking “which model should I use?”, often the most important question is “what exactly is this data measuring?”.

The value of an unexpected result

The result of 0.41% could have been treated as a discovery. Instead, it served as a warning. Results that are very different from what was expected do not need to be discarded, nor should they be immediately turned into conclusions. They may indicate a relevant phenomenon, a processing error, a population change, a methodological alteration, or a characteristic of the source that we do not yet understand.

In this case, the analysis began as an exploration of millions of faculty links and ended up revealing something broader about the work with data itself: sometimes, the main finding is not in the number found, but in the reason why it should be examined with caution.

RAIS allows us to see important transformations in the Brazilian formal labor market and offers a wealth of information difficult to reproduce from other sources. The analysis identified expressive growth in teaching positions, relevant changes in their composition, and large fluctuations in records of sick leave between 2020 and 2025.

These results remain informative. What changes is the way to interpret them. The transition to eSocial shows why historical series should not be treated merely as sequences of numbers. Each point in a series is the result of rules, systems, definitions, and administrative processes that also have a history.

For those who work with Data Science, one of the main conclusions of this exercise is that, before explaining why an indicator has changed, one should verify if the same phenomenon continues to be measured in the same way.

An algorithm can find patterns in millions of records. Knowing whether these patterns represent the phenomenon we want to understand remains a human task.

To access the references of this text click here

Who wrote this column

José Erasmo Silva

José Erasmo Silva é professor, formado em Matemática e Administração, com mais de 25 anos de experiência em gestão empresarial e de pessoas. É mestre e doutor em Administração, com foco em Finanças, e especialista em Data Science e Analytics e em Finanças e Controladoria. Realizou pós-doutorado na Universidade Federal da Bahia (UFBA). Atualmente, atua como professor orientador no MBA em Data Science, Inteligência Artificial e Analytics da USP/Esalq e leciona na EEP/FUMEP e na rede estadual de ensino de São Paulo.

You may also like

November 17, 2026

Detection and Diagnosis of Diseases in Corn and Rice in Mozambique with Deep Learning

Mozambican agriculture depends on corn and rice, fundamental crops for food security, but their productivity is strongly affected by phytosanitary diseases, whose early and accurate identification remains a challenge in resource-limited contexts. The study aimed to develop and compare Deep Learning-based models for the automatic detection and diagnosis of diseases in corn and rice leaves in Mozambique. Three approaches were used: a classic Convolutional Neural Network (CNN) developed from scratch, and two transfer learning models, ResNet50 and MobileNetV2. The dataset consisted of images from public databases and collected in the field, which underwent cleaning, balancing, and artificial augmentation steps. The models were trained and evaluated with standard performance metrics, including accuracy, precision, recall, F1-score, Log Loss, and Cohen’s Kappa. The classic CNN presented the best overall performance, with 94% precision, recall, and F1-score, and 93% accuracy and Cohen’s Kappa. MobileNetV2 achieved 91% F1-score and 90% accuracy, standing out for its lower Log Loss (0.30) and training time (6h32min59s). ResNet50 reached 80% F1-score and 79% accuracy. It was concluded that Deep Learning techniques have high potential to support phytopathological diagnosis in real agricultural contexts, contributing to faster and more accurate decisions in the management of corn and rice diseases.

Keywords: Deep learning; Rice; Plant diseases; Corn; Computer vision.

November 17, 2026

Artificial intelligence in accounting operational routines and financial efficiency

The incorporation of Artificial Intelligence (AI) and automation tools has reconfigured the routines of Brazilian accounting offices, impacting productivity, error mitigation, and financial efficiency. These impacts were analyzed in a small accounting firm, specialized in serving service providers in the medical field in São Caetano do Sul, São Paulo. The research adopted a qualitative approach, with a single case study design and incorporated units of analysis. Data were collected through a structured questionnaire, applied via Google Forms between January and February 2026, to five professionals in the operational structure of the firm in a census survey. The treatment followed content analysis with a priori and emergent categories. All respondents declared the use of automated tools or tools with AI resources. The perception of reduced execution time, decreased operational errors, rework, and better use of the team’s time converged in 80% of affirmative responses and 20% partial responses. The most benefited activities were bank reconciliation, import of fiscal documents, and tax assessment, with visible gains on typing errors, duplicate entries, and incorrect classification. Human intervention remained indispensable in tax planning, scenario analysis, and verification of ancillary obligations. The overall evaluation of AI’s contribution was predominantly favorable, with moderation linked to the complexity of fiscal regulations.

Keywords: Content analysis; Process automation; Accounting 4.0; Small business; Digital transformation.

November 17, 2026

Factors influencing schedule management in sewage treatment plant “retrofit” projects

Projects involving “retrofit” in Sewage Treatment Plants (STP) are characterized by high complexity, operational restrictions, and multiple technical interfaces. The study analyzed the perception of the professionals involved regarding the management and adherence to deadlines in “retrofit” projects for sewage treatment plants, considering technical, operational, and organizational factors. The research was characterized as an applied study, with a quali-quantitative approach, designed as a field survey. Data analysis was performed using descriptive statistics and qualitative categorization, employing SWOT, risk matrix (Probability x Impact), and 5W2H action plan tools. The results indicated a high perception of clarity (77.8%) and realism of deadlines (100% agreement), evidencing consistent initial planning. However, 77.7% of respondents pointed out that rework and scope changes directly impacted deadline adherence. The integration between engineering, manufacturing, and Production Planning and Control (PPC) showed an intermediate evaluation (77.7% partial/total agreement), indicating fragility in interfunctional coordination. The risk matrix highlighted as critical: scope changes, operational limitations of the STP, and technical complexity, all with a high level. As a central improvement, the standardization of scope change control, analysis, and approval was proposed, aiming to reduce rework and increase schedule predictability.

Keywords: Performance; Integration; Planning; Rework; Risk.

November 17, 2026

B2C Acquisition in diagnostic medicine: strategic partnerships with specialized medical offices

The diagnostic medicine sector in Brazil operates with high dependence on health plan providers, which limits the financial autonomy of service providers. This study proposed and analyzed a direct patient acquisition model (B2C) for a highly complex diagnostic medicine institution, structured through partnerships with Oncology and Cardiology practices and operationalized by a personalized customer service called concierge. The investigation adopted a mixed-methods approach in three phases: a field survey with a structured questionnaire applied to 87 patients, mapping of the operational flow using “Service Blueprint”, and analysis of the conversion rate from 312 medical referrals registered in the pilot project. The results indicated that trust in medical referral achieved 92.0% agreement, establishing itself as the main factor for adherence. The convenience of performing the exam on the same day reached 78.2%, and the willingness to pay directly was 55.2%, with polarization associated with the predominance of health plan users. The “Service Blueprint” identified three critical points in the transitions between stages of the patient journey. The overall conversion rate was 51.0%, with superior performance in Oncology (58.2%) compared to Cardiology (45.0%), and the greatest loss was concentrated in the acceptance and scheduling stage (28.3%). It was concluded that the model demonstrated possible operational viability and indications of patient acceptance, conditioned by price transparency and a reduction in the interval between consultation and approach by the concierge.

Keywords: Direct payment; Patient journey; Supplementary health; Concierge service; Conversion rate.

November 17, 2026

The role of the school manager in promoting an inclusive culture: leadership and educational policies

The research analyzed the school manager’s performance in promoting an inclusive culture, considering the need to transform legal guidelines and pedagogical foundations into effective institutional practices. It was assumed that school inclusion requires planning, organization, and strategic leadership in the school’s daily operations. A qualitative approach of an applied nature was adopted, developed in three complementary stages: documentary analysis of Brazilian legislation on inclusive education, bibliographic review on school leadership and management, and data systematization using management tools such as SWOT analysis and the PDCA cycle. The results indicated that the school management’s performance directly influenced the consolidation of inclusive environments, especially in the organization of processes, continuous team training, and strengthening of a collaborative institutional culture respectful of diversity. Structural, pedagogical, and relational challenges that demanded planned interventions were also identified. Finally, a Practical Guide for School Managers was developed, with guidelines, protocols, and operational orientations for decision-making and the organization of inclusive processes. It was concluded that the construction of an inclusive culture depended on the management’s ability to articulate public policies, pedagogical practices, and institutional processes, consolidating inclusion as a permanent commitment in the school context.

Keywords: educational accessibility; educational equity; participatory management; strategic planning; institutional processes.

November 17, 2026

An AI artifact based on Design Thinking to support the elicitation of software requirements from user needs

The development of software oriented to user needs frequently faces challenges in requirements elicitation, stemming from ambiguities in natural language and communication failures. Given this, an intelligent artifact integrating Design Thinking and Artificial Intelligence was proposed and validated, focusing on the Empathy and Definition stages, to support the collection, interpretation, and structuring of user needs. The research, of an applied, qualitative, and exploratory nature, conducted a Proof of Concept in a controlled environment, employing hypothetical scenarios. The artifact operated through guided natural language interviews, organizing information into elements such as pain points, context, and impact, and generating traceable problem syntheses. The results indicated that the system organized unstructured reports, reduced ambiguities, and preserved the meaning of the original information in most cases. The ability to reorganize logically in complex scenarios was also observed, although with limitations in separating multiple problems and dependence on instructions. It was concluded that the artifact has potential as a support for requirements elicitation, improving the clarity and structuring of information, but requires supervision and contextual control.

Keywords: Design Thinking; Requirements Elicitation; Requirements Engineering; Artificial Intelligence; Natural Language Processing.

November 17, 2026

The Extinction of the Monophasic Tax System in the Pharmaceutical Industry in Light of the Tax Reform

The consumption Tax Reform, established by Constitutional Amendment No. 132/2023, introduced structural changes to the Brazilian tax system, with the replacement of the Social Integration Program (PIS) and the Social Security Financing Contribution (COFINS) by the Contribution on Goods and Services (CBS), in a dual Value Added Tax (VAT) model. The objective was to analyze the impacts of the extinction of the single-stage system on the pharmaceutical industry, in light of the new regulatory design and its economic and legal effects. A qualitative, exploratory research was conducted through a case study and documentary analysis of legislation and specialized doctrine. It was examined whether the end of the single-stage system would imply a relevant increase in the tax burden for the sector. The results indicated that, although the formal overcoming of the single-stage system has occurred, the material exemption of medicines has been preserved, mainly due to the application of reduced rates of 60% and, in certain cases, a zero rate of the Tax on Goods and Services (IBS) and CBS. The change in model may cause occasional variations in revenue collection, without characterizing a structural increase in the tax burden, tending towards neutrality over time. It was preliminarily concluded that the Tax Reform reorganized the taxation technique in the pharmaceutical industry, replacing a concentrated regime with more transparent exemption mechanisms. This approach proved compatible with the principles of neutrality, full non-cumulativeness, and fiscality, maintaining tax protection for essential medicines and aligning fiscal policy with the fundamental right to health.

Keywords: CBS; COFINS; IBS; Dual VAT; PIS.