Correlation Between Construction Typology and Accident Rate—Case Study: Balearic Islands (Spain)

Abstract

This study examines occupational accident rates in the construction sector through a case study conducted in the Balearic Islands (Spain) between 2009 and 2018. The analysis is structured around three key dimensions: macroeconomic trends, changes in occupational accident reporting systems, and legislative reforms enacted during the study period. The research evaluates the influence of business, economic, and regulatory factors on two distinct construction typologies: private residential buildings (single-family and multi-family dwellings) and public-use residential buildings (hotels, aparthotels, and tourist apartments). The objective is to determine how business characteristics and construction environments influence occupational accident prevalence, establishing a causal framework informed by economic indicators and regulatory shifts. Using local data, the study analyzes variations in accident rates by construction type, accident form, and company size. The results indicate that accident rates are structurally shaped by economic and regulatory factors, with notable differences across construction typologies and company sizes. These findings highlight the need for evidence-based, targeted prevention strategies. This study enhances understanding of how structural conditions influence occupational risk in construction and offers a foundation for developing more effective safety policies.

1. Introduction

Persistently high rates of occupational fatalities and injuries, particularly in high-risk industries, are exacerbated by an increasingly diverse workforce driven by globalization. This diversity introduces linguistic barriers, high staff turnover, and mobile work environments, underscoring the pressing need for evidence-based approaches to safety training [1,2]. The construction sector stands out for its high accident rates, constituting a global concern [3]. Within the European Union, construction has consistently been identified as one of the economic activities with the highest rate of occupational accidents, both in absolute and relative terms, especially regarding fatal incidents [4,5]. In 2022, construction reported the highest incidence of non-fatal occupational accidents in the EU, with 2961 cases per 100,000 workers and the second-highest rate of fatal accidents (6.1 per 100,000), surpassed only by mining and quarrying (10.2) [4]. According to Eurostat [6], between 2008 and 2016, construction accounted for a significant share of fatal occupational accidents across the EU.

Building on this European panorama, it is essential to focus on the Spanish context, where construction consistently constitutes one of the most hazardous economic activities. In Spain, during the 2009–2018 period, the construction sector consistently recorded the highest incidence rate of occupational accidents among all economic activities (Figure 1), partly due to the dynamic nature of construction sites and the simultaneous presence of workers from multiple companies and trades [3].

Within Spain, particular attention should be paid to the Balearic Islands, a region with a strong dependence on tourism and construction, where accident rates exceed the national average and thus represent a relevant case study.

The construction sector faces multiple global challenges, including on-site safety issues, labor shortages, low productivity, and rising material costs [8]. Various studies have identified key risks, enabling the planning of preventive strategies [9,10,11]; nevertheless, an appropriate combination of these strategies is essential to effectively reduce accident rates [12]. Understanding the causes of construction-related accidents is essential to preventing their increase [13]. The high accident rate in the sector is attributed both to its inherent characteristics, such as the temporary nature of worksites, outdoor labor, task-specific conditions, and extensive subcontracting, and to organizational factors related to work rhythms and methodologies. These conditions heighten risks for construction workers, leading to a high number of accidents of varying severity and elevated mortality rates. Therefore, comprehensive and effective management of occupational health and safety within companies is essential for accident prevention [14], going beyond regulatory compliance and requiring coordinated efforts among stakeholders and the use of technological tools [15,16,17].

Latent errors, arising from management decisions, planning processes, or organizational culture, create conditions conducive to operational failures (active errors) and represent one of the main risks to occupational safety. Their identification and management are key to accident prevention. In the specific context of construction, characterized by temporary work environments, chain subcontracting, and fragmented production, these latent errors tend to accumulate and manifest during critical phases of the construction process, significantly increasing the likelihood of adverse events. Furthermore, continuous and specialized training, innovation in preventive methodologies, and the integration of safety measures throughout all project phases are essential factors for reducing accident rates and strengthening the sector’s safety culture [18].

Regarding business management and information gathering, employers have a legal obligation to prevent occupational accidents, which entails the establishment of an organizational culture based on continuous learning and ongoing training [19]. According to Law 31/1995 on Occupational Risk Prevention [20], the primary responsibility for workplace health and safety lies with the employer, who must guarantee it in all aspects related to work (Art. 14), while employees are required to cooperate by complying with the preventive measures adopted (Art. 29). Within this framework, it is essential to implement safety procedures and/or regulations that structure and align these responsibilities [21]. The investigation of occupational accidents is a fundamental pillar of industrial safety management systems, as collecting information about incidents enables learning from them and receiving continuous feedback that enhances prevention [22]. Systematic data collection in specialized databases allows organizations to identify specific risks, facilitating the implementation of short-term corrective actions and long-term preventive strategies to avoid the repetition of similar incidents [23]. In line with this approach, the present study is based on data extracted from the Delt@ System (Electronic Declaration of Injured Workers) [24], a database used in Spain that enables the classification and analysis of various variables related to the occurrence of workplace accidents.

1.1. Study Period for the Case Analysis

The 2009–2018 period was selected based on methodological criteria and data availability. It coincides with the full operational deployment of the Delt@ System in Spain, which has ensured consistent and traceable records since 2003; it covers a complete economic cycle, including phases of crisis and recovery in the sector; and from 2019 onwards, accidents involving self-employed workers were included, introducing a methodological discontinuity in the statistical series, which is why the study is limited to 2018. Additionally, this period aligns with the adoption of the ESAW methodology promoted by Eurostat, established in Regulation (EU) No. 349/2011, already implemented in Spain in previous years [25].

The analysis of the incidence rate of occupational accidents resulting in sick leave during the working day in autonomous communities in Spain between 2009 and 2018 reveals notable regional disparities. The Balearic Islands and Castile-La Mancha reported the highest values, well above the national average, while regions such as Madrid and Catalonia, despite registering a high absolute number of accidents, presented the lowest incidence rates. The national average during the period was 3432 accidents with sick leave per 100,000 workers [7] (

Table 1. Incidence rate of occupational accidents with sick leave in Spain, per 100,000 workers (2009–2018) [7].

	2009	2010	2011	2012	2013	2014	2015	2016	2017	2018	Mean 2009–2018
TOTAL	4263.4	4000.1	3633.8	2948.9	3009.2	3111.3	3252.0	3364.0	3408.8	3408.7	3432.40
Andalusia	4662.1	4261.3	3923.6	3174.7	3359.3	3466.6	3654.9	3778.8	3846.6	3838.0	3922.90
Aragon	3516.4	3442.0	3172.0	2631.2	2694.7	2740.2	3001.6	3150.6	3247.1	3303.3	3021.40
Principality of Asturias	5170.6	4892.0	4402.7	3515.7	3522.2	3669.0	3729.8	3554.8	3413.7	3444.5	3988.50
Balearic Islands	5281.9	5018.8	4714.7	3885.1	4070.1	4357.4	4571.1	4884.0	5025.0	4996.3	4685.80
Canary Islands	4693.1	4385.3	4145.5	3364.1	3415.5	3587.0	3608.7	3675.7	3618.2	3530.6	3528.50
Cantabria	3829.3	3479.2	3227.5	2592.7	2658.1	2847.3	2980.6	3017.9	3019.6	3078.6	3053.00
Castile-La Mancha	5070.0	4673.1	4199.8	3457.8	3524.8	3696.3	3896.4	4149.1	4153.8	4083.9	4090.20
Castile and Leon	4291.8	3976.2	3459.1	2786.8	2783.2	2936.0	3085.6	3246.1	3306.0	3392.0	3109.40
Catalonia	4215.0	4014.0	3587.3	2909.6	2992.3	3117.3	3192.5	3339.1	3283.7	3313.4	3129.70
Valencian Community	3722.2	3499.7	3181.8	2621.7	2722.9	2728.5	2907.0	2999.4	3111.0	3171.3	3074.60
Extremadura	4445.7	3979.9	3529.3	2988.7	3084.5	3393.7	3577.5	3582.8	3733.3	3804.0	3511.50
Galicia	4720.8	4376.0	3824.9	3005.1	2956.9	3095.2	3269.4	3295.6	3627.1	3514.1	3466.80
Community of Madrid	3681.0	3513.8	3234.7	2604.9	2575.3	2602.7	2688.1	2778.2	2757.5	2784.2	2808.90
Regionof Murcia	4080.2	3756.8	3519.2	2843.2	2973.0	3111.0	3413.8	3475.3	3586.3	3514.4	3402.10
Navarre	4171.0	3878.3	3641.1	2813.4	2876.9	3072.6	3185.3	3437.1	3541.3	3746.6	3320.90
Basque Country	4448.2	4289.2	3880.0	3292.1	3243.5	3244.4	3409.1	3482.7	3596.0	3390.2	3484.90
La Rioja	4389.7	4023.2	3743.9	2944.0	2968.0	3049.7	3404.7	3552.8	3668.9	3619.9	3429.00
Ceuta and Melilla	4708.6	4362.6	3566.9	2900.7	2928.0	3223.9	3402.4	3421.0	3295.7	3248.7	3276.30

). These findings underscore the need to analyze specific factors influencing occupational accident rates in each territory, as well as the importance of adapting preventive policies to the realities of each autonomous community.

Although comparative analyses across multiple Spanish regions would be insightful, the Balearic Islands constitute a representative case. Their accident incidents regularly surpass the national average (Table 1), while their economic dependence on tourism and construction makes them particularly sensitive to policy and market fluctuations. This regional focus allows for a detailed examination of underlying mechanisms that may also operate in other tourism-driven economies.

The annual evolution of the incidence rate of construction-related accidents in the Autonomous Community of the Balearic Islands, during the 2009–2018 period, consistently exceeds the national average trend (Figure 2).

1.2. Economic, Statistical, and Legislative Context of the Case Study

The study period from 2009 to 2018 was defined based on three key factors: the trends of economic data, changes in the statistical monitoring of occupational accidents, and legislative reforms, specifically Law 8/2012 of July 19 [27] and Decree-Law 1/2016 of January 12 [28].

Regarding the economic context, it is notable that the Balearic Islands reached a historic peak in GDP growth in 2008. However, that same year marked the beginning of an economic recession that lasted until 2013. This period was shaped by two concurrent but distinct crises: on one hand, the global financial crisis triggered by the collapse of the international banking system following the bankruptcy of Lehman Brothers in September 2008; on the other, a severe real estate crisis in Spain, caused by the bursting of the construction and mortgage credit bubble. Although the crisis had global reach, its effects were particularly pronounced in the Spanish economy, especially in regions heavily dependent on tourism and construction, such as the Balearic Islands. From 2014 onward, an economic recovery phase began, driven largely by a tourism boom between 2014 and 2017 [29] (

Table 2. Year-on-year variation in Gross Domestic Product (GDP) in the Balearic Islands, (2009–2018) [29].

	2009	2010	2011	2012	2013	2014	2015	2016	2017	2018
Construction	−8.54	−18.02	−10.61	−5.88	−5.52	1.91	9.92	7.19	1.94	−0.35
Gross Domestic Product at Market Prices	−4.06	−0.6	−0.06	−1.26	−1.2	3.22	4.31	4.24	3.00	3.00

). These economic fluctuations directly affected the evolution of occupational accident rates. During the recession, reduced activity and employment may have led to a decrease in workplace accidents, whereas the subsequent recovery and employment growth, particularly in construction and tourism, increased exposure to risk and, consequently, the number of incidents. Therefore, the evolution of occupational accident rates must be interpreted within the framework of these economic cycles, avoiding the interpretation of changes solely to preventive or regulatory factors.

In 2019, the statistical definition of occupational accidents was updated to include cases involving self-employed workers, which significantly impacted accident rate statistics. This expansion introduced a break in methodological consistency that prevents direct comparison between data before and after that year and must be considered in any trend analysis. Regarding legislative changes, the following regulations are of particular relevance:

• Law 8/2012 of July 19, on tourism in the Balearic Islands [27]: introduced urban planning exceptions for tourist establishments in the region, permitting an increase of 20% to 40% in buildable area for five-star hotels or higher, provided the project had special interest and received a favorable report from an expert committee. This measure aimed to renovate the hotel infrastructure and enhance the competitiveness of the sector, which accounts for more than 40% of the Balearic GDP and employs nearly 200,000 people [27].
• Decree-Law 1/2016 of January 12, on urgent urban planning measures [28]: limited expansions in hotel establishments, reducing the maximum buildable percentages to 10%, extendable to 15% or 20% only in hotels rated three stars or higher. It also limited exemptions and strengthened controls over the quality and justification of interventions, thereby intensifying urban planning criteria applicable to the tourism sector [28].

This contextual framework frames the analysis of the period by considering economic, statistical, and regulatory factors that may be related to the findings of the study. Ultimately, to properly interpret the evolution of occupational accident rates in the Balearic Islands between 2009 and 2018, it is essential to account for the effects of economic cycles on activity and employment, as well as the legislative reforms that influenced construction activity in the tourism sector. Ignoring these elements may lead to misleading interpretations about actual safety trends and the effectiveness of preventive policies during the analyzed period.

1.3. Construction Typologies Analyzed in the Case Study

This study focuses on the analysis of occupational accident rates in the field of residential construction in the Autonomous Community of the Balearic Islands, choosing this category due to its relevance in regional productive development and its direct link to the tourism sector. Within this category, two clearly differentiated construction typologies are identified:

• Private-use residential, which includes single-family and multi-family dwellings.
• Public-use residential, which encompasses hotels, aparthotels, and tourist apartments.

Although there are no studies that directly compare accident rates between public-use and private-use residential construction, the specific characteristics of the Balearic tourism sector make it possible to identify significant organizational differences. In public-use residential construction, the imperative to complete establishments before the onset of the high tourist season results in compressed execution schedules and a heightened dependence on subcontracting. This dynamic leads to the involvement of larger, more diverse, and less cohesive work crews. Conversely, private-use residential construction projects typically unfold over extended timelines, allowing for greater continuity and stability within work teams. These organizational distinctions give rise to markedly different risk exposure profiles, even when the same companies are responsible for executing both types of projects.

The selection of these typologies is based on the hypothesis that differences in construction processes, execution timelines, and working conditions, particularly shaped by tourism seasonality, which significantly affects the economy of this autonomous community; may generate distinct patterns of occupational accident rates.

1.4. Research Objective

The objective of this study is to examine how macroeconomic fluctuations, and legislative changes have influenced occupational accident rates in the construction sector of the Balearic Islands. The analysis focuses on two construction typologies: private-use residential buildings, including single- and multi-family dwellings, and public-use residential buildings, such as hotels, aparthotels, and tourist apartments. The study also assesses whether accident rates differ by company size (micro, small, medium, and large enterprises) and how these differences relate with each construction typology. In addition, it investigates how accident types, such as overexertion, falls, and contact with materials, vary across construction typologies and over time.

Occupational accident rates are assessed according to company size, based on the classification established in Law 5/2015 [30]: microenterprises (MiE, fewer than 10 workers), small enterprises (SE, up to 49), medium enterprises (ME, up to 250), and large enterprises (LE, more than 250). This categorization allows for the analysis of whether the incidence rate varies depending on company size and construction type throughout the study period.

2. Methodology

This research aims to understand occupational accident rates in the residential construction sector in Spain through a comprehensive analysis of data recorded in the main tool used to report and register workplace accidents: the Delt@ System, managed by Social Security [24]. Its use enables the exchange of information between companies, managing and collaborating entities, and Public Administrations, serving as a primary source of data by recording accidents occurring in companies classified according to the CNAE (in Spanish: Clasificación Nacional de Actividades Económicas, and in English: National Classification of Economic Activities). This hierarchical coding system, used in Spain to classify the economic activities of companies and self-employed workers, groups these activities into section, division, group, and class levels, ensuring uniformity in the collection and statistical analysis of data related to the national economic structure [31]. Each economic activity is assigned a numerical code that allows for its identification in official records and statistical studies [32].

2.1. Study Design and Data Selection

The selection and processing of data focused on accidents recorded under CNAE code 412, corresponding to residential building construction. Within this category, additional filters were applied using the keywords “HOTEL,” “HOT,” and “TEL” to identify potential cases related to lodging establishments. Each record was then geolocated via the workplace address field and verified using Google Maps to confirm whether the site corresponded to a hotel or aparthotel. In cases where the address information indicated a lodging establishment but was incorrectly referenced, accidents were reclassified as public use to correct evident misclassifications. All remaining records were classified as private-use residential construction. This procedure ensured consistent case classification and aligned with the legal and economic distinction between private-use and public-use residential housing in the Balearic Islands, by consistently applying objective and reproducible criteria throughout (Figure 3).

For the analysis of occupational accident rates, accidents were classified according to the type of associated risk, following the coding established by the Delt@ System and the ESAW methodology. This classification allows for the identification of patterns of occurrence and the establishment of links between working conditions and recorded accidents. The selected risk types are as follows:

• Physical overexertion (71).
• Impact resulting from a fall from different levels (31).
• Impact resulting from a fall on the same level (32).
• Other contact from group 3 (impact against a stationary object, moving worker) (39).
• Collision or impact resulting from struck by fragments (41).
• Collision or impact resulting from struck by falling objects (42).
• Contact with cutting material (knife, blade, etc.) (51).
• Contact with piercing material (nail, sharp tool, etc.) (52).
• Contact with scratching or hard material (grater, sandpaper, etc.) (53).
• Other contact from group 5 (cutting, piercing, or hard material) (59).
• Other contact not coded in the current classification (99).

2.2. Variables and Segmentation

The study covers two main construction typologies: private-use residential buildings and public-use residential buildings. The data were broken down by company type (MiE, SE, ME, and LE) and by accident form, according to the official classification of the Delt@ System (physical overexertion, falls from different levels, falls on the same level, struck from objects or tools, contact with cutting materials, among other categories coded by the system). For each combination of construction typology, company type, and accident form, annual absolute values and relative incidence rates were calculated, allowing for a longitudinal and comparative analysis of accident rates.

This segmentation structure makes it possible to identify the most vulnerable company profiles and the most frequent accident forms within each construction typology, providing a detailed view of risk patterns and supporting the design of focused prevention strategies. In all tables presenting the number of accidents and their incidence rates, the distinction between both is made using parentheses: incidence rate values are shown in parentheses.

2.3. Statistical Analysis

Accident incidence rates were processed and segmented by construction type, graphically represented to identify notable trends in their temporal evolution. Descriptive statistical methods were applied, including measures of central tendency (mean, median) and dispersion (standard deviation, range). Additionally, 95% confidence intervals were calculated, and correlation and regression analyses were performed. The confidence interval was calculated using the standard Formula (1), based on Student’s t-distribution, as described in the specialized literature [33].

$${IC}_{95\%}=\overline{x}\pm t_{\left(\propto /2\right)}·\left(s/\surd n\right)$$

(1)

where $\overline{x}$ is the annual mean, $s$ the standard deviation, $n$ the sample size (years) and $t_{\left(\propto /2\right)}$ Student’s t-value corresponding to a 95% confidence level and 9 degrees of freedom (t ≈ 2.262).

Due to the variation in the number of accidents recorded across typologies, a standardized incidence index was established to enable consistent comparison. This index is calculated as the proportion of accidents occurring in a given year relative to the total number of accidents recorded during the study period, multiplied by 100, according to Equation (2).

$$Accident Incidence Index=(Accidents Number per Year/Total Accidents Number per Period)·100$$

(2)

Additionally, time-series analyses and inferential tests were applied, such as Student’s t-test for independent samples and Pearson’s correlation analysis, with a statistical significance level set at p < 0.05. Data processing and analysis were performed in Excel [34].

2.4. Modeling and Advanced Analysis

Two complementary approaches were used for the temporal trend analysis:

• Simple Exponential Smoothing and Holt Models [35]: These models enabled the estimation of a smoothed trajectory of the incidence rate, giving priority to the most recent data and reducing the impact of short-term fluctuations. A non-seasonal parametrization with additive trend was applied, enabling short-term (3 years) and medium-term (5 years) projections under the assumption of structural continuity.
• Multivariable Linear Regression by construction typology and company type [36]: Regression models were built using the following Equation (3).

$$y_t={\beta }_0+{\beta }_1·t+{\epsilon }_t$$

(3)

where $y_t$ is the incidence rate, $t$ denotes the year, and ${\beta }_1$ is the annual trend slope. Categorical variables were included using Dummy Coding, a technique commonly used in regression and other statistical models to incorporate qualitative variables as numerical predictors. Interactions and shared patterns were analyzed, and model fit was assessed using R² coefficients and residual analysis.

2.5. Presentation and Data Quality Assurance

The results are presented through tables and figures that summarize and visualize the identified patterns, supporting the interpretation of temporal evolution and the comparison between different construction typologies. Methodological quality and data anonymity were ensured in accordance with current regulations on personal data protection, specifically Regulation (EU) 2016/679 of the European Parliament and Council (General Data Protection Regulation—GDPR) [37] and Organic Law 3/2018 on Personal Data Protection and Guarantee of Digital Rights [38].

2.6. Added Value of the Methodological Approach

This methodological approach provides a sound analytical framework for the comparative study of occupational accident rates and the development of preventive strategies tailored to the specific context of residential construction. The breadth of the analysis, the use of advanced statistical techniques, and the integration of graphical visualizations enable the identification of key patterns and significant relationships between accident rates across different construction typologies (private-use and public-use residential), accident forms, and company types. Altogether, this constitutes a reliable methodological foundation for designing occupational health and safety policies aimed at reducing accident incidence in the sector.

A limitation of the present study lies in the uneven sample distribution, derived from the difference in the number of accidents between private-use and public-use residential construction. While the overall trends remain reliable, this imbalance must be considered when interpreting the statistical results. To ensure the validity of the analysis, the results were first evaluated independently for each typology (private-use and public-use) and subsequently integrated into a comparative perspective. This procedure makes it possible to account for the effect of the imbalance in sample size, ensuring that the conclusions are based on stable patterns across both contexts.

This study is based exclusively on quantitative statistical analysis of accident records, without incorporating qualitative organizational or cultural factors. This scope definition defines the scope of the research and ensures internal consistency of statistical comparisons. Nonetheless, these dimensions are recognized in the discussion as key avenues for future research.

3. Results

The analysis of results was structured into six thematic sections: each focused on a specific aspect of occupational accident rates in residential construction in the Balearic Islands during the period 2009–2018.

3.1. Analysis of the Number of Accidents and Incidence Rate

Between 2009 and 2018, a total of 16,658 occupational accidents were recorded. These data were analyzed in terms of absolute values and annual incidence rates, differentiating between construction typologies: private residential and public residential buildings (

Table 3. Annual Number of Occupational Accidents and Incidence Rate and GrowthRate by Construction Typology (Private vs. Public Residential), Balearic Island (2009–2018) [24].

RESIDENTIAL CONSTRUCTIVE TIPOLOGY	NUMBER OF ACCIDENTS (INCIDENCE RATE) GROWTHRATE %
	2009	2010	2011	2012	2013	2014	2015	2016	2017	2018
Private use	2243	1612	1296	912	916	1187	1431	1666	2123	2195
	(14)	(10)	(8)	(6)	(6)	(8)	(9)	(11)	(14)	(14)
		−28.13%	−19.60%	−29.63%	0.44%	29.59%	20.56%	16.42%	27.43%	3.39%
Public use	9	31	62	69	88	88	185	167	220	158
	(1)	(3)	(6)	(6)	(8)	(8)	(17)	(16)	(20)	(15)
		244.44%	100.00%	11.29%	27.54%	0.00%	110.23%	−9.73%	31.74%	−28.18%

).

During the analysis period, notable variations were identified in the number of accidents reported within the private residential construction typology. In 2009, a total of 2243 accidents were reported, marking the beginning of the time series. However, a downward trend was observed in the following years, reaching its lowest point in 2012, with only 912 incidents recorded. From 2013 onwards, a shift in the trend was evident, with a progressive increase in the number of accidents, culminating in a peak of 2195 incidents in 2018. This sustained growth in the final stage of the study period may reflect greater exposure to risks in residential contexts, an increase in construction activity, or improvements in accident reporting and registration systems. Regarding accidents recorded in residential facilities intended for public use, a marked increase was documented throughout the period. In 2009, only 9 accidents were reported in this category, contrasting with the peak of 220 accidents observed in 2017. However, in 2018, a slight reduction in the incidence rate was observed, with a total of 158 accidents reported. Although the overall incidence is significantly lower compared to private residential construction, the upward trend may be associated with increased activity in public-use buildings or changes in regulations and accident reporting procedures (Table 3).

Between 2009 and 2018, marked differences were recorded in the annual number of occupational accidents between the two residential construction typologies. Private residential construction presented an annual average of 1548 accidents, with a standard deviation of 501, a median of 1431, and a range of 1331 accidents. In contrast, public residential construction showed a substantially lower annual average of 108 accidents, with a standard deviation of 70, a median of 88, and a range of 211 accidents (see

Table 4. Annual Statistical Summary of Occupational Accident Numbers in Private and Public Residential Construction, Balearic Island (2009–2018) [24].

RESIDENTIAL CONSTRUCTIVE TIPOLOGY	ANNUAL AVERAGE	DEVIATION	MEDIAN	RANGE	IC 95% MINIMUM	IC 95% MAXIMUM
Private use	1548	501	1431	1331	1237	1859
Public use	108	70	88	211	64	151

). To assess the reliability of these estimates, 95% confidence intervals were calculated. For private residential construction, the interval for the annual average ranged from 1237 to 1859 accidents, while for public residential construction it ranged from 64 to 151 accidents. These intervals reflect the natural variability of the data and allow us to affirm, with 95% confidence, that the true annual average of accidents lies within these ranges. The greater spread observed in private residential construction indicates a more significant variability in the annual occurrence of accidents compared to public-use construction. However, the latter shows a relatively steeper upward trend in the final years of the period analyzed. These results highlight a higher incidence of occupational accidents in private residential construction, as it presents a substantially greater number of incidents throughout the entire 2009–2018 period.

However, the temporal evolution suggests that public-use residential construction is increasing its accident incidence rate at a faster pace. Although private-use residential construction maintains higher values in most years, the public-use typology shows a more accelerated increase starting in 2014, even exceeding the values of private-use construction between 2015 and 2017 in terms of incidence rate. This pattern indicates a relatively notable rise in accidents within the public-use construction typology (see Figure 4). Using the incidence rates in Table 3, the average incidence in private-use construction exceeded that of public-use prior to the 2012 inflection point (2009–2011 means: private 10.7 vs. public 3.3). After 2012, which coincides with a period of regulatory easing, the pattern is inverted (2012–2018 means: private 9.7 vs. public 12.9). This observed shift, also visible in Figure 4, suggests that public-use residential construction exhibited a structurally elevated risk profile during the expansion phase, while private-use did not.

3.2. Analysis of Temporal Evolution

Annual variations in accident rates were analyzed, identifying patterns of increase or decrease over time. The changes were interpreted in relation to the economic and regulatory context, and differences between the two construction typologies were assessed. Regarding temporal evolution, it is noted that the private residential construction typology showed an increase in accident rates following an initial period of decline, while the public residential construction typology displayed a steady rise throughout the decade, with fluctuations in recent years (Table 3). In this regard, it is important to consider the trend curves described by both time series (Figure 5).

In the case of private residential construction, the evolution follows a curved trajectory, with an initial decline between 2009 and 2012, followed by a sustained increase until 2018, suggesting a recovery in construction activity and a possible increase in associated occupational risk. In contrast, public residential construction shows a generally upward trend, although with occasional fluctuations in recent years, indicating a more consistent pattern of accident rate growth throughout the decade.

This interpretation allows for an understanding not only of the absolute accident figures, but also of the distinct temporal patterns between the two construction typologies. Data analysis reveals that, between 2009 and 2012, accidents in the private residential typology were significantly more frequent, presenting a higher incidence rate. However, from 2012 onwards, a trend reversal is identified, with the incidence rates of accidents in the public residential typology beginning to exceed those of the private typology. This difference continues in subsequent years, reaching a situation of near equivalence between both typologies by 2018 (Table 3, Figure 4).

The analysis of accidents in public-use residential construction shows a steady increase starting in 2012, corresponding with the relaxation of urban planning regulations, despite the ongoing economic recession. In 2014, a year of economic recovery, a more marked increase is recorded, with the number of accidents doubling and remaining stable until 2017, when the regulatory restrictions introduced in 2016 begin to take effect (Table 3, Figure 4). In contrast, trends in private-use residential construction appear to be driven primarily to economic factors, as urban planning regulations do not directly affect this construction typology.

3.3. Analysis of Accident Types

Accidents were classified according to their form (physical overexertion, falls, contact, impact, etc.), and annual averages, trends, and temporal correlations were calculated for each type. This section allowed for the identification of the most frequent accident types and the analysis of their differential evolution across construction typologies.

Regarding accident types, code 71 (physical overexertion) was the most frequent in both construction typologies. This type of accident far surpasses all others, showing a consistent trend over time. For the remaining accident types, the period between 2009 and 2012 reveals marked variability, with highly variable patterns, as some types of accidents were almost nonexistent in certain construction typologies or showed pronounced differences in relative frequency. However, from 2012 onwards, an alignment in the behavior of the different types of accidents is identified, both in terms of typology and frequency. The data presented in

Table 5. Number of Accidents and Incidence Rate and Growth Rate by Accident Type in Private and Public Residential Construction, Balearic Island (2009–2018) [24].

			CONSTRUCTION TYPOLOGY FOR RESIDENTIAL USE
YEAR	ACCIDENT FORM		PRIVATE USE			PUBLIC USE
	CODE	DEFINITION	Accidents Number	(Incidence Rate)	Growth Rate	Accidents Number	(Incidence Rate)	Growth Rate
2009	71	OVER-EXERTION	888	(40)		5	(56)
	31	FALL FROM DIFFERENT LEVELS	214	(10)		0	(0)
	32	FALL ON SAME LEVEL	255	(11)		0	(0)
	42	STRUCKBY FALLING OBJECT	116	(5)		0	(0)
	59	ANOTHER CONTACT FROM GROUP 5	112	(5)		0	(0)
	51	CONTACT WITH CUTTING MATERIAL	0	(0)		2	(22)
	53	CONTACT WITH SCRATCHING MATERIAL	0	(0)		1	(11)
	99	OTHER NOT CODED	0	(0)		1	(11)
2010	71	OVER-EXERTION	620	(38)	−30.18%	11	(35)	120.00%
	31	FALL FROM DIFFERENT LEVELS	168	(10)	−21.50%	2	(6)
	32	FALL ON SAME LEVEL	138	(9)	−45.88%	2	(6)
	42	STRUCK BY FALLING OBJECT	86	(5)	−25.86%	1	(3)
	51	CONTACT WITH CUTTING MATERIAL	78	(5)		5	(16)	150.00%
	99	OTHER UNCODED CONTACT	0	(0)		3	(10)	200.00%
2011	71	OVER-EXERTION	556	(43)	−10.32%	28	(45)	154.55%
	31	FALL FROM DIFFERENT LEVELS	120	(9)	−28.57%	9	(15)	350.00%
	32	FALL ON SAME LEVEL	123	(9)	−10.87%	5	(8)	150.00%
	42	STRUCK BY FALLING OBJECT	90	(7)	4.65%	1	(2)	0.00%
	39	ANOTHER CONTACT FROM GROUP 3	61	(5)		1	(2)
	41	STRUCK BY FRAGMENTS	0	(0)		3	(5)
2012	71	OVER-EXERTION	368	(40)	−33.81%	24	(35)	−14.29%
	32	FALL ON SAME LEVEL	94	(10)	−23.58%	8	(12)	60.00%
	31	FALL FROM DIFFERENT LEVELS	83	(9)	−30.83%	7	(10)	−22.22%
	42	STRUCK BY FALLING OBJECT	56	(6)	37.78%	2	(3)	100.00%
	52	CONTACT WITH PIERCING MATERIAL	50	(5)		1	(1)
	41	STRUCK BY FRAGMENTS	0	(0)		5	(7)	66.67%
2013	71	OVER-EXERTION	351	(38)	−4.62%	28	(32)	16.67%
	31	FALL FROM DIFFERENT LEVELS	119	(13)	43.37%	11	(13)	57.14%
	32	FALL ON SAME LEVEL	76	(8)	−19.15%	10	(11)	25.00%
	42	STRUCK BY FALLING OBJECT	68	(7)	21.43%	7	(8)	250.00%
	51	CONTACT WITH CUTTING MATERIAL	52	(6)		5	(6)
2014	71	OVER-EXERTION	439	(37)	25.07%	33	(38)	17.86%
	31	FALL FROM DIFFERENT LEVELS	136	(11)	14.29%	15	(17)	36.36%
	32	FALL ON SAME LEVEL	125	(11)	64.47%	5	(6)	−50.00%
	42	STRUCK BY FALLING OBJECT	78	(7)	14.71%	7	(8)	0.00%
	51	CONTACT WITH CUTTING MATERIAL	66	(6)	26.92%	3	(3)	−40.00%
	52	CONTACT WITH PIERCING MATERIAL	18	(2)		5	(6)
2015	71	OVER-EXERTION	559	(39)	27.33%	78	(42)	136.36%
	31	FALL FROM DIFFERENT LEVELS	175	(12)	28.68%	18	(10)	20.00%
	32	FALL ON SAME LEVEL	130	(9)	4.00%	13	(7)	160.00%
	42	STRUCK BY FALLING OBJECT	106	(7)	35.90%	24	(13)	242.86%
	51	CONTACT WITH CUTTING MATERIAL	76	(5)	15.15%	8	(4)	166.67%
2016	71	OVER-EXERTION	649	(39)	16.10%	66	(40)	−15.38%
	31	FALL FROM DIFFERENT LEVELS	218	(13)	24.57%	20	(12)	11.11%
	32	FALL ON SAME LEVEL	174	(10)	33.85%	19	(11)	46.15%
	42	STRUCK BY FALLING OBJECT	108	(6)	1.89%	11	(7)	−54.17%
	51	CONTACT WITH CUTTING MATERIAL	93	(6)	22.37%	9	(5)	12.50%
2017	71	OVER-EXERTION	825	(39)	27.12%	80	(36)	21.21%
	31	FALL FROM DIFFERENT LEVELS	305	(14)	39.91%	37	(17)	85.00%
	32	FALL ON SAME LEVEL	194	(9)	11.49%	17	(8)	−10.53%
	42	STRUCK BY FALLING OBJECT	163	(8)	50.93%	16	(7)	45.45%
	51	CONTACT WITH CUTTING MATERIAL	124	(6)	33.33%	8	(4)	−11.11%
2018	71	OVER-EXERTION	811	(37)	−1.70%	52	(33)	−35.00%
	31	FALL FROM DIFFERENT LEVELS	337	(15)	10.49%	20	(13)	−45.95%
	32	FALL ON SAME LEVEL	194	(9)	0.00%	18	(11)	5.88%
	42	STRUCK BY FALLING OBJECT	182	(8)	11.66%	18	(11)	12.50%
	51	CONTACT WITH CUTTING MATERIAL	143	(7)	15.32%	14	(9)	75.00%

show that, from that year, the proportions of accidents due to falls (codes 31 and 32), impacts (code 42), and contact with materials (codes 51 and 52) tend to level off and exhibit more similar patterns between private and public residential construction. Since 2012, the relative frequencies of accidents tend to converge across both construction typologies, indicating the influence of shared preventive measures and labor reorganization processes (Table 5).

Between 2009 and 2018, the main types of occupational accidents were analyzed in both private and public residential construction, applying measures of central tendency, dispersion, confidence intervals, and temporal trend analysis for each category. Physical overexertion (code 71) was the most frequent accident type in both typologies, with an annual average of 606.6 accidents in private residential construction (standard deviation: 180.2; 95% CI: 495.0–718.2) and 40.5 in public-use construction (standard deviation: 26.2; 95% CI: 24.3–56.7). In the private-use typology, this category showed a declining trend (r = −0.72) and a more uniform distribution, while in the public-use typology the trend was increasing (r = 0.88), with greater variability and year-to-year variability, corresponding with contextual or regulatory changes (

Table 6. Descriptive Statistics and Trends of Occupational Accidents (Absolute Values) by Accident Type in Private and Public Residential Construction, Balearic Island (2009–2018) [24].

ACCIDENT FORM	TYPE OF USE	ANNUAL AVERAGE	STANDARD DEVIATION	IC 95%	RANGE (MIN–MAX)	CORRELATION WITH YEAR (r)	TREND
Over-Exertion (71)	Private use	606.60	180.20	495.00–718.20	351.00–888.00	−0.72	↓Decreasing
	Public use	40.50	26.20	24.30–56.70	5.00–80.00	+0.88	↑Increasing
Fall From Different Levels (31)	Private use	187.50	75.50	140.60–234.40	83.00–337.00	+0.77	↑Increasing
	Public use	13.90	10.60	7.30–20.50	0.00–37.00	+0.97	↑Increasing
Fall on Same Level (32)	Private use	150.40	49.30	131.10–169.70	76.00–255.00	−0.04	→Stable
	Public use	9.70	6.30	6.20–13.20	0.00–19.00	+0.88	↑Increasing
Struck by Falling Object (42)	Private use	104.90	34.50	92.10–117.70	56.00–182.00	+0.67	↑Increasing
	Public use	8.30	7.40	4.30–12.30	0.00–24.00	+0.83	↑Increasing
Contact with Cutting Material (51)	Private use	63.20	50.00	42.10–84.30	0.00–143.00	+0.85	↑Increasing
	Public use	6.40	4.30	3.70–9.10	0.00–14.00	+0.64	↑Increasing

).

Falls from different levels (code 31) and falls on the same level (code 32) also were notable due to their frequency and variability. In private residential construction, falls from different levels had an average of 187.5 accidents per year (standard deviation: 75.5; 95% CI: 140.6–234.4) and an upward trend (r = 0.77), while falls on the same level had an average of 150.4 accidents (standard deviation: 49.3; 95% CI: 131.1–169.7), without a clear trend. In public-use construction, the annual average was 13.9 accidents for falls from different levels (standard deviation: 10.6; 95% CI: 7.3–20.5; r = 0.97) and 9.7 for falls on the same level (standard deviation: 6.3; 95% CI: 6.2–13.2; r = 0.88), both showing strong increasing trends (Table 6).

Struck by falling objects (code 42) and contact with cutting materials (code 51) also showed notable frequency and high variability, especially in the private-use typology. Struck by falling objects recorded an average of 104.9 accidents per year in private-use construction (standard deviation: 34.5; 95% CI: 92.1–117.7; r = 0.67) and 8.3 in public-use construction (standard deviation: 7.4; 95% CI: 4.3–12.3; r = 0.83). Contact with cutting materials had an average of 63.2 accidents in private-use construction (standard deviation: 50.0; 95% CI: 42.1–84.3; r = 0.85) and 6.4 in public-use construction (standard deviation: 4.3; 95% CI: 3.7–9.1; r = 0.64), both showing upward trends (Table 6).

The variability analysis shows that codes 71, 31, and 42 show higher variability across both construction typologies, indicating that they occur irregularly over time. The accident histogram confirms that codes 71 and 31 are the most frequent, representing a significant percentage of the total recorded incidents. The less frequent accident types (codes 39, 41, 52, 53, 59, and 99) showed low and sporadic values, which limits their analysis to a brief descriptive overview.

In summary, occupational accident rates are higher in private residential construction, although the rate of increase in public-use construction are more pronounced, particularly in the categories of falls and overexertion. These differences reflect differentiated trends, which may be relevant for the design of tailored preventive strategies.

The analysis of average annual incidence rates between 2009 and 2018 revealed marked differences both among accident types and between the two construction typologies (private and public residential). The most frequent accident was physical overexertion (code 71), with similar averages in both sectors: 39.0 in private use and 39.2 in public use, although the standard deviation was substantially greater in the latter (7.16 vs. 1.76). Other noteworthy types included falls from different levels (code 31) and falls on the same level (code 32), with annual averages ranging between 8.0 and 11.6, also slightly higher in the private-use typology. Accidents involving contact with cutting materials (code 51) and being struck by falling objects (code 42) showed lower values but pronounced differences in dispersion and trend between the two construction typologies (

Table 7. Descriptive Statistics and Trends of Occupational Accidents (Incidence Rate) by Accident Type in Private and Public Residential Construction, Balearic Island (2009–2018) [24].

ACCIDENT FORM	TYPE OF USE	MEDIA ANUAL	STANDARD DEVIATION	IC 95% (MINOR)	IC 95% (MAJOR)	RANGE (MIN–MAX)	CORRELATION WITH YEAR (r)	TREND
Over-Exertion (71)	Private use	39.00	1.76	37.74	40.26	37.00–43.00	−0.458	↓Decreasing
	Public use	39.20	7.16	34.08	44.32	32.00–56.00	−0.507	↓Decreasing
Fall From Different Levels (31)	Private use	11.60	2.12	10.08	13.12	9.00–15.00	0.866	↑Increasing
	Public use	11.30	5.21	7.57	15.03	0.00–17.00	0.645	↑Increasing
Fall on Same Level (32)	Private use	9.50	0.97	8.80	10.20	8.00–11.00	−0.245	→ Stable
	Public use	8.00	3.59	5.43	10.57	0.00–12.00	0.552	↑Increasing
Struck by Falling Object (42)	Private use	6.60	1.07	5.83	7.37	5.00–8.00	0.785	↑Increasing
	Public use	6.20	4.13	3.24	9.16	0.00–13.00	0.808	↑Increasing
Contact with Cutting Material (51)	Private use	4.60	2.50	2.81	6.39	0.00–7.00	0.66	↑Increasing
	Public use	6.90	7.05	1.86	11.94	0.00–22.00	−0.435	↓Decreasing

).

In terms of temporal evolution, the correlation with year shows a declining trend for overexertion (71) (r = −0.458 in private use and −0.507 in public use) and for contact with cutting materials (51) in public use (r = −0.435). In contrast, strong increasing trends were observed in falls (31) (r = 0.866 in private use) and in being struck by falling objects (42) (r = 0.785 and 0.808 for private and public use, respectively) (Table 7).

3.4. Analysis by Company Size

Between 2009 and 2018, the occupational accident distribution was analyzed according to construction typology (private and public residential buildings) and categorized by company size: microenterprise (MiE), small enterprise (SE), medium enterprise (ME), and large enterprise (LE). For each group, central tendency metrics, variability, 95% confidence intervals, time-based correlations, and linear regressions were calculated.

Regarding the accident distribution by company type and construction typology, it is observed that in private residential construction, the majority of accidents took place in MiE. This predominance persisted until 2016, when accidents in SE reached similar levels, while ME showed a marginal incidence. In contrast, in public residential construction, during 2009, the majority of accidents were recorded in MiE. However, starting in 2010, a gradual rise in accidents was noted in SE, followed by ME, and to a lesser extent, MiE. This pattern remained consistent throughout the rest of the period, extending until 2018 (Figure 6).

Some outliers were identified that showed substantial deviation from the overall trend, both in private residential construction—where LE, consistently recorded as 0.00, could be considered statistically atypical, and MiE, although showing only a minor decline, followed a stable pattern over the years—and in public residential construction, where MiE showed a marked decrease between 2009 and 2018, which may be interpreted as an anomalous pattern. On the other hand, ME no outliers were detected in 2011 and 2018. The only quantitative outlier in the series is that of 2010 (39), given its unusually high value relative to other years (Figure 6).

The analysis identified clear trends in occupational accident rates according to construction typology and company size. In private residential construction, a stable trend was noted in the MiE category in absolute terms, although incidence rates revealed a gradual decline. In contrast, SE showed an upward trend in both metrics. The ME and LE categories remained stable overall, with small variations and no major changes (

Table 8. Number of Accidents and Incidence Rate and Growth Rate by Company Type in Private and Public Residential Construction, Balearic Island (2009–2018) [24].

YEAR	COMPANY TYPE	PRIVATE USE RESIDENTIAL			PUBLIC USE RESIDENTIAL
		Number of Accidents	(Incidence Rate)	Growth Rate	Number of Accidents	(Incidence Rate)	Growth Rate
2009	MiE	1010	(45)		5	(56)
	SE	885	(39)		3	(33)
	ME	346	(15)		1	(11)
	LE	2	(0)		0	(0)
2010	MiE	807	(50)	−20.10%	8	(26)	60.00%
	SE	608	(38)	−31.30%	11	(35)	266.67%
	ME	192	(12)	−44.51%	12	(39)	1100.00%
	LE	5	(0)	150.00%	0	(0)
2011	MiE	711	(55)	−11.90%	14	(23)	75.00%
	SE	467	(36)	−23.19%	30	(48)	172.73%
	ME	116	(9)	−39.58%	17	(27)	41.67%
	LE	2	(0)	−60.00%	1	(2)
2012	MiE	510	(56)	−28.27%	21	(30)	50.00%
	SE	353	(39)	−24.41%	28	(41)	−6.67%
	ME	49	(5)	−57.76%	20	(29)	17.65%
	LE	0	(0)	−100.00%	0	(0)	−100.00%
2013	MiE	509	(56)	−0.20%	29	(33)	38.10%
	SE	329	(36)	−6.80%	44	(50)	57.14%
	ME	76	(8)	55.10%	15	(17)	−25.00%
	LE	2	(0)		0	(0)
2014	MiE	614	(52)	20.63%	26	(30)	−10.34%
	SE	472	(40)	43.47%	42	(48)	−4.55%
	ME	100	(8)	31.58%	17	(19)	13.33%
	LE	1	(0)	−50.00%	3	(3)
2015	MiE	656	(46)	6.84%	36	(19)	38.46%
	SE	587	(41)	24.36%	89	(48)	111.90%
	ME	174	(12)	74.00%	60	(32)	252.94%
	LE	14	(1)	1300.00%	0	(0)	−100.00%
2016	MiE	728	(44)	10.98%	33	(20)	−8.33%
	SE	781	(47)	33.05%	91	(54)	2.25%
	ME	152	(9)	−12.64%	43	(26)	−28.33%
	LE	5	(0)	−64.29%	0	(0)
2017	MiE	943	(44)	29.53%	54	(25)	63.64%
	SE	918	(43)	17.54%	107	(49)	17.58%
	ME	260	(12)	71.05%	59	(27)	37.21%
	LE	2	(0)	−60.00%	0	(0)
2018	MiE	905	(41)	−4.03%	34	(22)	−37.04%
	SE	1015	(46)	10.57%	95	(60)	−11.21%
	ME	275	(13)	5.77%	29	(18)	−50.58%
	LE	0	(0)	−100.00%	0	(0)

). In public residential construction, stronger trends were detected: both absolute values and incidence rates showed a continued decrease in MiE, while SE exhibited steady growth, especially in the total number of accidents. In ME and LE, although there were fluctuations, the overall trends remained stable, with low levels and no sustained increases.

Table 9. Descriptive Statistics and Temporal Trends of Absolute Occupational Accident Values by Company Type, Balearic Island (2009–2018).The parameters included in the main row are as follows: Company Type (COMPANY); Construction Typology (Private Residential and Public Residential) (RESIDENTIAL); Annual Mean (MEAN); Standard Deviation (SD); Standard Error (SE); T-Score (T-S); Minimum and Maximum Value Range (RANGE); Correlation with Year (r); and Observed Trend (TREND) [24].

COMPANY	RESIDENTIAL	MEAN	SD	SE	T-S	IC-MI	IC-MA	RANGE	(r)	TREND
Microenterprise (MiE)	Private Use	739.30	175.12	55.38	2.262	614.03	864.57	509.00–1010.00	0.067	→Stable
	Public Use	26.00	14.68	4.64	2.262	15.50	36.50	5.00–54.00	0.900	↑Increasing
Small Enterprise (SE)	Private Use	641.50	244.69	77.38	2.262	466.46	816,.4	329.00–1015.00	0.432	↑Increasing
	Public Use	54.00	38.05	12.03	2.262	26.78	81.22	3.00–107.00	0.958	↑Increasing
Medium Enterprise (ME)	Private Use	174.00	95.49	30.20	2.262	105.69	242.31	49.00–346.00	0.080	→Stable
	Public Use	27.30	20.18	6.38	2.262	12.86	41.74	1.00–60.00	0.757	↑Increasing
Large Enterprise (LE)	Private Use	3.30	4.14	1.31	2.262	0.34	6.26	0.00–14.00	0.075	→Stable
	Public Use	0.40	0.97	0.31	2.262	−0.29	1.09	0.00–3.00	−0.038	→Stable

summarizes the results based on the absolute number of accidents, allowing for an assessment of the overall magnitude of the issue in each construction typology.

In private residential construction, most accidents were concentrated in MiE, a trend that persisted until 2016, when SE reached comparable levels. ME showed a residual incidence, and LE maintained very low and stable values. The annual average number of accidents was 739.3 for MiE, 641.5 for SE, 174.0 for ME, and 3.3 for LE. Dispersion was highest in SE (standard deviation: 244.69; 95% CI: 466.46–816.54). Temporal correlations in private residential construction were low, indicating that accidents did not show a clear trend of increase or decrease over time, except for SE, which exhibited a moderate positive correlation (r = 0.432) and a growing trend. Visually, this corresponds to trend curves close to a horizontal line, with no significant slope (Table 9).

These results suggest that although accident rates are higher in private residential construction across all company types, the growth in accidents in public residential construction is more pronounced, especially among smaller companies. Each company typology faces distinct challenges in accident prevention, which may justify specific approaches in future safety policies, prioritizing oversight in small and medium-sized enterprises, where the annual increase is more marked.

In public residential construction, most accidents in 2009 were recorded in MiE, but from 2010 onwards, an increase in SE was observed, followed by ME, and to a lesser extent, MiE. This pattern was consolidated through 2018. Although absolute values were lower than in the private sector, temporal correlations were high for MiE, SE, and ME (r = 0.900; 0.958; and 0.757, respectively), indicating a growing trend in the number of accidents (Table 9). Regression slopes show that SE increased by an average of 12 accidents per year, MiE by 4.4, and ME by 5.0, while LE showed very low values with no relevant trend (Table 9).

Given the different magnitudes of accidents observed across the analyzed typologies, the incidence rate is used as the primary metric, as it allows for comparing the relative evolution of accident rates among groups of different sizes and facilitates the identification of patterns and temporal trends in both construction typologies. This approach enables a robust comparative analysis, focused on relative changes rather than solely on absolute accident figures, and is therefore essential for accurately interpreting the dynamics of accident rates in heterogeneous contexts.

Table 10. Descriptive Statistics and Temporal Trends of Occupational Accident Incidence Rates by Company Type, Balearic Island (2009–2018) [24].

COMPANY	RESIDENTIAL	MEAN	SD	SE	T-S	IC-MI	IC-MA	RANGE	(r)	TREND
Microenterprise (MiE)	Private Use	48.90	5.61	1.77	2.262	44.89	52.91	41.00–56.00	−0.547	↓Decreasing
	Public Use	28.40	10.72	3.39	2.262	20.73	36.07	19.00–56.00	−0.623	↓Decreasing
Small Enterprise (SE)	Private Use	40.50	3.81	1.20	2.262	37.78	43.22	36.00–47.00	0.785	↑Increasing
	Public Use	46.60	8.22	2.60	2.262	40.72	52.48	33.00–60.00	0.870	↑Increasing
Medium Enterprise (ME)	Private Use	24.50	8.25	2.61	2.262	18.6	30.4	11.00–39.00	−0.033	→Stable
	Public Use	10.30	2.98	0.94	2.262	8.17	12.43	5.00–15.00	0.018	→Stable
Large Enterprise (LE)	Private Use	0.10	0.32	0.10	2.262	−0.13	0.33	0.00–1.00	0.174	→Stable
	Public Use	0.50	1.08	0.34	2.262	−0.27	1.27	0.00–3.00	−0.119	→Stable

summarizes the results based on the incidence rate.

In private residential construction, the analysis of incidence rates shows that MiE initially exhibited a growing trend until 2012, followed by a slight decline in subsequent years (Figure 6). The regression slope (−1.012) indicates a gradual decrease in values for this category. SE showed a stable trend, with minimal fluctuations, but with a slight upward slope (+0.998). ME displayed a slightly increasing trend, although changes were moderate, with a slope close to stability (+0.018). LE maintained consistently low values, practically zero, with no significant changes (+0.018).

In public residential construction, MiE showed a sharp drop in values during 2010, followed by marked fluctuations in the following years (Figure 6). The regression slope (−2.206) indicates a more pronounced decline compared to the private-use category. SE exhibited a variable pattern, with a more significant upward trend (+2.364), reflecting a clear and steady increase. ME showed greater temporal variability, with notable fluctuations and a slight downward trend (−0.091), although no significant changes were observed over the entire period. LE, as in private residential construction, maintained very low and nearly constant levels throughout the period (−0.042) (Table 10).

The analysis of the collected data allowed for the identification of clear relationships between company size and the evolution of occupational accidents. Both MiE TCU-Private and MiE TCU-Public —where TCU refers to Construction Typology of Use—showed a declining trend in incidence rates, which was more pronounced in the public-use typology (r = −0.623). However, in absolute terms, MiE TCU-Public showed sustained growth (r = 0.900), suggesting a relative reduction within a growing volume.

On the other hand, SE TCU-Public showed a strongly increasing trend in both metrics (r = 0.958 in absolute values and 0.870 in incidence), while SE TCU-Private showed a more moderate growth (r = 0.432 in absolute values and 0.785 in incidence). In ME TCU-Public, incidence rates remained stable (r = 0.018), although absolute values indicate a slight increase. In contrast, ME TCU-Private showed overall stability (r, close to zero in both metrics), with annual fluctuations lacking a clear direction. Finally, LE, in both private and public contexts, showed low and constant accident levels, with no significant changes during the analyzed period (Table 9 and Table 10).

3.5. Evolution of the Incidence Rate by Company Type and Construction Typology

According to the analyses conducted, which include exponential smoothing techniques and multivariable regression models, consistent patterns have been identified in the evolution of the occupational accident incidence rate, differentiated by company type and construction typology (Figure 7).

In private residential construction, both MiE TCU-Private and MiE TCU-Public show a clear downward trend, although this decline occurs at a more moderate pace in recent years. In contrast, SE TCU-Public and SE TCU-Private maintain a steady upward trend, which is more pronounced in the public-use residential typology. The ME TCU-Public category reveals a variable evolution, with some stabilization and a slight reduction in its growth rate. Meanwhile, ME TCU-Private remains at relatively stable levels, with no significant linear trend. Finally, the LE category remains unchanged, reflecting its low activity volume and virtually constant accident rates.

3.6. Trend Projection Analysis by Company Type

Using exponential smoothing models applied to the 2009–2018 time series, short- and medium-term projections (3 and 5 years) were generated based on company size. A 10–15% reduction in accident rates is expected for MiE TCU-Public, while SE TCU-Public could exceed an incidence rate of 70 by 2025. ME TCU-Public is expected to stabilize between 20 and 22, and LE would remain at nearly null levels. These projections assume continuity in the conditions observed during the analyzed period.

A multiple linear regression model was fitted to predict SE TCU-Public based on other variables. The model showed an excellent level of accuracy (R² = 0.99), explaining 99% of the observed variability in incidence rate data. The results indicate a negative relationship between MiE TCU-Public, ME TCU-Public, and LE TCU-Public with SE TCU-Public, meaning that as the other categories decrease, SE TCU-Public increases. In private-use typologies, such as MiE TCU-Private, SE TCU-Private, and ME TCU-Private, significant negative coefficients were also observed, although with less explanatory weight.

The data show a progressive replacement of MiE by SE, especially in the public-use residential typology, where this transition is associated with a more pronounced increase in accident rates. Although the private-use residential construction typology maintains higher absolute accident levels, the relative growth in the public-use typology is more marked, with a convergence between both typologies starting in 2013, driven by the increasing participation of SEand ME in public-use construction.

4. Discussion

The analysis of occupational accident rates in residential construction in the Balearic Islands between 2009 and 2018 has enabled the identification of significant patterns related to economic evolution, regulatory changes, and shifts in the sector’s business structure. The results revealed clear differences between private-use and public-use residential construction typologies, as well as between different company types.

4.1. Critical Interpretation of the Results

The study reveals significant differences in accident rates depending on construction typology and company size. The private-use residential construction typology shows the highest accident burden, with marked year-to-year variability influenced by economic factors. In contrast, the public-use residential construction typology, despite lower absolute figures, shows a steady increase since 2012, associated with regulatory easing and increased construction activity.

Accident rates in private residential construction show a decline between 2009 and 2012, followed by a gradual rebound until 2018, mainly linked to economic recovery and improvements in reporting systems. Since this typology is not affected by urban planning relaxations, the economic context appears to be the main explanatory factor, consistent with studies that directly associate economic growth with increased occupational accident rates [39].

In public residential construction, a gradual increase in accident rates has been observed since 2012, coinciding with urban planning relaxations, even in a recessionary context. In 2014, during the economic recovery, accidents doubled and then stabilized until 2017, when regulatory restrictions introduced in 2016 may have contributed to a slight reduction. This pattern suggests a direct relationship between regulatory changes, increased construction activity, and accident trends, along with improvements in reporting systems.

This divergent trend between the two construction typologies cannot be explained solely by internal sector factors but rather appears to respond to broader structural transformations that have affected the Balearic economy. The increase in accidents in the public-use residential construction typology, for example, may be interpreted not only as a direct effect of regulatory changes, but also as a result of public policies aimed at economic stimulus, which relied on investment in infrastructure and tourism facilities as a tool for post-crisis recovery.

These differences highlight the importance of applying tailored approaches to occupational risk management, considering both organizational characteristics and external factors that affect safety, depending on the construction typology.

The partial liberalization of urban land in some municipalities of the Balearic Islands promoted construction projects with limited technical and preventive oversight, which may have contributed to the increase in accident rates, especially among SE, which often have limited management capacity. This phenomenon, along with the sustained growth in accidents among SE in both public and private residential construction typologies, suggests a transition toward a weaker and less formalized business structure in terms of occupational safety.

Since 2012, incidence rates in public-use residential construction have begun to match or exceed those of private-use construction, demonstrating the need to adapt preventive strategies to each construction typology, considering not only the magnitude but also the growth patterns. The observed variability suggests a complex environment influenced by economic, regulatory, and business factors, reinforcing the importance of future research on demographic, socioeconomic, and environmental variables to design more effective and context-sensitive strategies. Although the absolute number of accidents was very unequal between private-use and public-use construction, this aspect was methodologically addressed through a separate analysis for each typology, followed by an integrated comparison. This dual approach allowed us to identify consistent patterns and reinforced the robustness of the conclusions, ensuring that the imbalance in sample size did not distort the interpretation of the results.

It should be noted, however, that the present study did not directly account for cultural and organizational factors, such as safety training, subcontracting practices, or managerial commitment. Nevertheless, the specialized literature consistently emphasizes that their interaction with economic and regulatory conditions decisively influences how risks are managed on construction sites. The interaction between cultural and organizational factors and broader economic and regulatory influences plays a decisive role in shaping occupational accident rates. While economic cycles and legislative reforms create structural conditions that either increase or mitigate exposure to risk, their effects are filtered through the internal dynamics of companies. Managerial commitment is another critical factor: during periods of regulatory relaxation, firms with proactive leadership may voluntarily adopt higher standards of prevention, thereby attenuating the rise in accident rates; conversely, in companies where prevention is seen as a mere compliance obligation, regulatory leniency often translates into a deterioration of safety conditions. These interactions underscore that economic and regulatory variables cannot be understood in isolation, but rather in constant interplay with organizational culture and practices, which ultimately determine how risks are managed on the ground. For this reason, future research should integrate these dimensions to provide a more comprehensive understanding of the mechanisms underlying the persistence of certain accident patterns in the sector.

These typologies not only represent statistical differences but also correspond to organizational and contextual characteristics. Public-use residential projects are heavily affected by tourism seasonality, which results in compressed project timelines and leads to greater reliance on temporary or subcontracted labor. By contrast, private-use residential projects generally involve smaller crews, extended construction periods, and more stable workforce structures. These structural differences in workforce organization and equipment utilization plausibly explain the divergent accident patterns observed between the two typologies. While these organizational mechanisms were not directly captured in the current dataset, they offer a reasonable interpretive framework aligned with the observed patterns and should be explicitly examined in future research. We did not have direct data on equipment allocation or crew composition; consequently, these mechanisms are suggested as testable hypotheses for future mixed-methods research.

Between 2009 and 2018, physical overexertion (code 71) was the most frequent accident type in both typologies, with a downward trend in private-use construction and an upward trend in public-use construction. Since 2012, an alignment in accident type distribution between both typologies has been observed, suggesting the influence of common factors, such as standardized procedures, organizational improvements, and cross-cutting preventive measures.

Falls from different levels (code 31) and falls on the same level (code 32), being struck by falling objects (code 42), and contact with cutting materials (code 51) account for a considerable share of accidents, especially in private-use construction, where they show greater variability. The sustained increase in falls in both typologies suggests persistent deficiencies in collective protection and training. In public-use construction, although the figures are lower, the relative increase is more evident. Also noteworthy is the rise in accidents involving falling objects (42), indicating deficiencies in load handling at height.

The variation among the most frequent accident types, overexertion (71), falls (31), and struck incidents (42), reinforces the need to tailor preventive strategies to each typology. Despite their lower volume, accidents involving cutting materials (51) are rising in private-use construction (r = 0.660), possibly due to the improper use of tools. These patterns underscore the urgency of strengthening training, improving working conditions, and implementing targeted measures based on risk type and construction context.

To address the rising incidence of falls from height and accidents caused by falling objects, preventive strategies should prioritize reinforced collective protection systems, enhanced training for safe work at heights, and stricter protocols for securing loads. Strengthening on-site inspections and sanctioning mechanisms for non-compliance are also crucial for ensuring enforcement.

The analysis reveals significant differences in the evolution of accident rates according to company type and construction typology. In private-use construction, MiE shows a downward trend, while SE exhibits a slight increase. In the public-use sector, a marked reduction is observed in MiE, alongside a steady increase in SE. ME and LE remain stable, with minor fluctuations. These patterns, confirmed by autocorrelation analysis, reflect structural rather than cyclical changes, highlighting the growth of SE since 2013 as an indicator of a notable business shift.

The multivariable regression analysis confirms a structural shift in construction typology, with a progressive replacement of MiE by SE, especially in the public-use sector, where correlations are stronger. This business restructuring process reinforces the need to adapt preventive management to the specific characteristics of each company type and work environment.

The evolution of incidence rates in MiE TCU-Public shows a plateau since 2012, suggesting that external factors, such as regulatory changes, economic fluctuations, or demand dynamics, may affect accident rates. The sporadic fluctuations in SE and ME within public-use construction appear to reflect situational rather than structural factors. This general stability, without a significant reduction in risk, highlights the need to strengthen preventive measures and to investigate the impact of variables such as internal organization, staff training, and the implementation of safety measures.

The redistribution of accidents in the public-use residential construction typology, with a shift from MiE to SE and ME, appears to be influenced by economic factors, legislative changes, and shifts in demand, particularly following the enactment of Law 8/2012. This structural change highlights a divergent trend in accident rates: while MiE shows a consistent decline, possibly due to greater oversight or reduced activity, SE exhibits sustained growth, reflecting vulnerabilities linked to its expansion without robust safety systems. ME remains stable, and LE reports almost negligible figures, which may be attributed to high prevention standards or possible underreporting mechanisms or outsourcing of hazardous tasks.

These findings reinforce the need to design safety strategies tailored to each company type and construction typology, taking into account their organizational capacities, levels of risk exposure, and available resources for preventive management.

4.2. Comparison with Previous Research

The findings of this study align with a broad body of literature demonstrating how economic cycles directly influence occupational accident rates, particularly in labor-intensive sectors such as construction. Recovery following a recession is often accompanied by an increase in accidents due to production pressure and the weakening of preventive mechanisms, a phenomenon described as the “collateral effect of expansion” [39,40,41].

In the case of the Balearic Islands, following the 2008 crisis, an economic model was promoted based on tourism growth and public investment in construction as a driver of recovery [42]. This context accounts for the sustained increase in accident rates in public-use residential construction from 2012 onward, especially among SE with limited preventive capacity. The recovery of tourism since 2013 generated indirect demand for infrastructure, reinforced by more flexible urban planning regulations such as Law 8/2012 [27], without a proportional strengthening of occupational safety conditions.

Furthermore, a transformation in the business structure is observed: MiE has gradually given way to SE, resulting in a progressive structural shift. Although SE theoretically have more resources, they exhibit persistent deficiencies in training, preventive culture, and organizational control [43,44], which may explain their growing contribution to accident indicators, consistent with the findings of this study.

4.3. Theoretical Implications

From a theoretical standpoint, the findings reinforce approaches that conceptualize occupational accident rates as a multidimensional phenomenon, shaped by economic, regulatory, organizational, and cultural factors. This study demonstrates that it is not sufficient to analyze accidents in absolute terms or through single indicators; it is essential to contextualize them according to construction typology, company size, and temporal framework, integrating both macro- and microeconomic variables into explanatory and predictive models.

The convergence observed in accident types since 2012, particularly in codes 71, 31, 42, and 51, suggests risk homogenization processes linked to subcontracting, task standardization, and shared precariousness across typologies. This phenomenon raises important questions about the structural transfer of risk, supporting future interdisciplinary research that integrates labor economics, work sociology, and preventive management.

Additionally, the study provides evidence of divergent accident patterns based on company size: a decline in MiE, sustained growth in SE, and stability in ME and LE. This distinct distribution calls for a reconsideration of safety management models through a segmented approach, tailored to the organizational capacities and resources of each type of company.

Another relevant factor is technological innovation, such as digital monitoring tools and new safety equipment, which, although not fully developed during the study period, are increasingly shaping accident prevention strategies. This emerging dimension reinforces the need to incorporate technological variables into future theoretical models of occupational risk.

4.4. Practical and Methodological Implications

From a practical perspective, the findings highlight the need for preventive policies tailored to construction typology and company size. The sustained increase in accident rates among SE, particularly in public-use residential construction, calls for enhanced labor inspections, technical support, and specialized training. In contrast, the stabilization of indicators in MiE may reflect both a reduction in activity and a shift in risk towards more precarious forms, such as self-employment. In private-use residential construction, the homogeneity of accidents enables the implementation of standardized protocols focused on ergonomics, load handling, and fall prevention.

Policymakers can ensure that preventive strategies are effectively tailored to company size and typology by moving away from a “one-size-fits-all” approach and adopting differentiated frameworks. In this study, policy implications in this study refer specifically to regulatory aspects, broader political considerations. As shown in Figure 4, regulatory changes had a significant effect on public-use residential construction, highlighting the need to tailor preventive strategies to the legal frameworks of each sector. These regulatory insights are grounded in empirical data. In fact, the comparison between typologies reveals a reversal in average incidence rates after 2012 (private 9.7 vs. public 12.9 for 2012–2018), a change that corresponds with the timeline of urban-planning reforms outlined in the case study. For microenterprises (MiE), this may involve simplified compliance tools, technical assistance, and targeted inspections that recognize their limited managerial and financial capacity. For small enterprises (SE), policies should prioritize training programs, access to shared prevention services, and mechanisms that strengthen organizational culture, since this group increasingly concentrates accident rates. Medium-sized enterprises (ME) require strategies that promote the institutionalization of safety management systems, while large enterprises (LE), although generally reporting lower accident rates, should be monitored to prevent risk externalization through subcontracting.

In parallel, preventive regulations should incorporate flexibility to adapt to different construction typologies. Public-use construction, more exposed to regulatory shifts and seasonal dynamics, demands specific oversight mechanisms and coordinated planning with tourism policies, whereas private-use construction benefits more from protocols addressing ergonomics, falls, and individualized training. By combining regulatory flexibility, differentiated support, and oversight mechanisms adapted to company profiles, policymakers can enhance the effectiveness of prevention strategies while avoiding homogenized solutions that fail to address structural diversity.

From a methodological standpoint, the use of the incidence rate has enabled comparative analysis of accident trends across company types and construction typologies, addressing the limitations of absolute-value analysis. However, notable limitations are present in the database, such as the exclusion of self-employed workers prior to 2019 and the absence of sociodemographic data. Additionally, the text-based classification of typologies may not capture all relevant cases, and the causal link between economic factors and accident rates requires dedicated longitudinal studies.

Finally, it is recommended to expand the analysis to other key sectors in the Balearic Islands, such as hospitality, transportation, and healthcare, and to broaden the study period to assess long-term structural trends. It is also suggested to incorporate variables such as age, contract type, and nationality, and to adopt mixed-method approaches integrating quantitative and qualitative analysis. The very low accident rates in LE should be interpreted with caution, considering possible biases related to underreporting or task outsourcing. Overall, the study advocates for segmented preventive strategies and continuous monitoring systems to support sustained improvements in occupational safety.

Although the findings of this study are robust for the Balearic Islands, their generalization to other regions must be approached with caution. The Balearic economy is highly dependent on tourism, with construction activity closely linked to seasonal demand and to specific regulatory frameworks, such as urban planning policies designed to promote hotel renovation. These structural characteristics differentiate the Islands from other Spanish regions and from most European contexts. Nevertheless, the mechanisms identified (such as the substitution of microenterprises (MiE) by small enterprises (SE), the divergent trajectories of private-use and public-use construction, and the link between economic cycles, regulatory changes, and accident incidence) may be relevant in other territories with comparable dynamics. In particular, regions such as the Canary Islands, the Costa del Sol in Andalusia, the Valencian Community (especially Alicante and Castellón), Murcia (the La Manga area), and certain coastal areas of Catalonia (Costa Brava and Costa Dorada) share a strong dependence on tourism as an economic driver and a significant weight of construction activity related both to second homes and to the hotel industry. These territories also exhibit a productive fabric dominated by micro and small construction enterprises, high labor seasonality, and an urban planning framework shaped by the need to manage tourism pressure and promote hotel modernization or renovation plans. Thus, while numerical results cannot be directly extrapolated, the explanatory patterns observed here provide useful insights for comparative studies, are especially relevant for testing in these contexts, and should be validated through research in regions with different economic and regulatory conditions, thereby reinforcing the external validity of the findings and identifying common patterns or regional specificities.

This regional approach should be interpreted as both a limitation and a strength: while it restricts direct generalization, it ensures an in-depth understanding of the mechanisms of driving accident rates in a highly exposed territory. Future research should expand the analysis to other Spanish regions to test the external validity of the patterns identified here.

Preventive frameworks must also incorporate resilience to future economic crises and regulatory changes. Adaptive regulations, permanent monitoring systems, and flexible prevention mechanisms are necessary to ensure that safety standards remain robust even in times of structural change.

In particular, the role of seasonality and work organization in hotel construction, which is closely linked to the dynamics of tourism demand, represents a relevant factor that could not be fully addressed in this statistical analysis. This aspect should therefore be considered a priority for future research, ideally through mixed methods of approaches that combine quantitative indicators with qualitative evidence from case studies and interviews.

A mixed-methods approach combining statistical data with interviews and case studies is recommended for future research. Such approaches would allow the identification of organizational and cultural mechanisms that are not fully captured by statistical correlations, thereby deepening the understanding of why certain accident patterns persist.

5. Conclusions

This study concludes that economic factors and regulatory changes have had a distinct impact on occupational accident rates in the construction sector, depending on the construction typology analyzed. In private-use residential construction, accident trends were mainly influenced by economic cycles, whereas in public-use residential construction, a steadier upward trend was observed, associated with regulatory changes that heightened risk. This trend leveled off in the years following the implementation of regulatory limitations (2016), demonstrating that urban planning policies can effectively influence accident rates in this construction typology.

Physical overexertion (code 71) was the most frequent cause of accidents in both construction typologies, although with contrasting trends: a downward trend in private-use construction and an upward trend in public-use construction. This finding highlights the importance of implementing targeted preventive strategies based on organizational context.

A sustained increase was observed in falls from height (code 31) and in injuries caused by falling objects (code 42), indicating persistent deficiencies in collective protection and in the organization of work at height. The gradual convergence in accident profiles between both construction typologies since 2012 suggests a potential standardization of labor processes, potentially linked to widespread subcontracting and task homogenization.

The positive and negative correlations identified among different types of accidents support the presence of internal structural dynamics, such as task specialization, risk prioritization, and the influence of organizational factors on accident rates.

Company size proved to be a key factor in shaping occupational accident patterns. In private-use residential construction, most accidents had predominantly occurred in MiE, although a progressive decline in incidence rates was observed throughout the study period. SE showed a steady increase, which may be interpreted as a structural transformation in the sector’s sectoral structure.

In public-use residential construction, this shift was more pronounced: accident rates gradually shifted from MiE to SE, with a sustained increase in incidence rates over the analyzed period. ME, while showing occasional fluctuations, maintained a generally stable trend, with no evidence of consistent growth. LE, meanwhile, reported consistently low accident rates across both typologies, likely due to greater preventive capacity or, in some cases, outsourcing of higher-risk tasks.

These findings highlight a gradual replacement of MiE by SE as the primary agents in the sector, particularly in public-use residential construction. This phenomenon is interpreted as a structural reconfiguration of the construction sector in the Balearic Islands over the past decade.

The methodological approach adopted, combining descriptive, inferential, and predictive analysis, provided a comprehensive understanding of occupational accident rates in residential construction, based on construction typology. This perspective demonstrates that accident rates are shaped not only by external variables (economic or regulatory) but also by internal factors, such as business structure, work organization, and the implementation of preventive measures.

The patterns identified show that occupational accidents present distinct patterns depending on the combination of construction typology and company size, suggesting that preventive strategies should be adapted to the specific conditions of each group.

Occupational accident rates in residential construction in the Balearic Islands between 2009 and 2018 display complex patterns shaped by structural, regulatory, and organizational factors. While absolute accident figures remain higher in private-use residential construction, the behavior of public-use construction exhibits a more dynamic trend, particularly among SE, warranting focused preventive attention.

This research provides a solid empirical foundation for the development of tailored public policies, responsive to the actual needs of each construction typology. The identified structural changes should be accounting for monitoring and inspection programs, as well as in training strategies designed for companies based on their typology. Likewise, it emphasizes the need for evidence-based preventive planning, taking into account the particularities of the economic and regulatory context, as well as the organizational characteristics of each business type.

Improving occupational safety in the construction sector cannot be addressed solely through regulatory measures; it requires the active involvement of all stakeholders, the promotion of a preventive culture, and a long-term strategy based on robust evidence of the factors shaping accident rates.

The unequal magnitude of accident records between private-use and public-use construction conditions is the direct comparability of results. This limitation does not invalidate the observed trends but requires cautious interpretation.

Although the findings of this study are robust for the Balearic Islands, their generalization to other regions must be approached with caution. The mechanisms identified—such as the substitution of microenterprises by small enterprises, the divergent trajectories of private-use and public-use construction, and the influence of economic cycles and regulatory changes—may be relevant for other territories with comparable dynamics. However, their extrapolation to other regional or international contexts requires additional comparative research to validate these mechanisms.

Future studies should also explore the specific impact of seasonality and work organization in hotel construction, given its strong influence on labor demand and accident exposure in public-use residential projects.

Preventive frameworks must prioritize measures against falls from height and falling objects, reinforcing collective protections, training, and enforcement mechanisms.

A relevant avenue for further progress lies in adopting a mixed-methods approach. Although the present study is mainly based on quantitative statistical techniques, future research could be enriched by incorporating qualitative perspectives, such as interviews with workers, managers, and inspectors, or detailed case studies of specific accidents.

Preventive frameworks must also incorporate resilience to future economic crises and regulatory changes. Adaptive regulations, permanent monitoring systems, and flexible prevention mechanisms are necessary to ensure that safety standards remain robust even in times of structural change. For future research, it is particularly relevant to analyze the role of organizational and cultural dimensions in accident prevention. Aspects such as a strong safety training culture, effective management of subcontracting practices, and genuine managerial commitment may function as buffers against the pressures derived from economic cycles or regulatory relaxation. In this regard, promoting preventive leadership and embedding safety values within company structures should be considered a priority research line to support more effective policies and programs aimed at reducing occupational accident rates in the construction sector.

The study demonstrates that the distinction between private-use and public-use residential construction is well-founded but based on regulatory, organizational, and economic factors that define distinct occupational risk profiles. Nonetheless, we recognize that further advances will require the integration of additional data on workforce composition, subcontracting practices, and equipment utilization. Incorporating these dimensions into future mixed-methods research would enable a more comprehensive understanding of the persistence of accident patterns across construction typologies.