15. Discussion.md 4.4 KB
5 Discussion
The path‑analysis results demonstrate that the effect of traffic volume on road‑traffic safety (RTS) is mediated primarily through accident severity, rather than through accident frequency. Although a higher number of motorised vehicles significantly increased the total count of crashes (β = 0.62, *p* < 0.001), its indirect contribution to RTS was modest (b₁ = 0.21). By contrast, traffic volume showed a stronger positive association with the daily number of fatalities (β = 0.38, *p* < 0.001), and the corresponding mediated pathway yielded an indirect effect that more than doubled that of the frequency route (a₂·b₂ = 0.22 vs. a₁·b₁ = 0.13). These findings confirm hypothesis H1: accident severity is the principal determinant shaping a country’s overall level of road‑traffic safety.
The traditional road‑fatality rate indicator captures only deaths per 10 000 vehicles (Elvik, Vaa, & Christensen, 2019) and therefore obscures two critical dimensions of crash risk—frequency (how many crashes occur) and severity (whether those crashes are fatal). When policymakers focus exclusively on reducing the number of fatalities in proportion to traffic volume, they may underestimate risk factors that generate severe outcomes, such as speeding or unsafe road surfaces. Consequently, safety strategies should be oriented toward severity‑specific interventions, for example:
- Intelligent speed‑control systems that become active during periods of high traffic flow;
- Risk‑based infrastructure planning, which prioritises upgrades at locations where vehicle concentrations are greatest; and
- Behavioural campaigns aimed at preventing severe crashes (e.g., programmes to curb alcohol impairment and driver fatigue).
These policy directions are consistent with the broader international literature. Multi‑indicator models have been shown to predict mortality more accurately than single‑rate approaches (Weijermars, Van Acker, & De Vos, 2014; Chen, 2020), and composite safety indices improve the precision of risk assessment (Coll & Marshall, 2013). Moreover, our study is the first to apply structural path analysis to daily Estonian traffic data, thereby quantifying the indirect effects of exposure on safety outcomes—a methodological advance highlighted in earlier work on structural equation modelling for road‑safety research (Bollen, 1989; Liu, Wang, & Chen, 2018).
Limitations
Several limitations should be acknowledged. First, the analysis is based on aggregated daily counts, which precludes examination of individual driver behaviour or vehicle technical condition—factors that can influence crash severity. Second, potential measurement errors (e.g., mis‑recorded crashes) and omitted external variables such as weather conditions or pavement quality may bias the estimated relationships. Finally, the cross‑sectional nature of the data limits inference about causal direction beyond what is captured by the specified mediation pathways.
Future research
Future studies could extend the model by incorporating demographic (e.g., age distribution of road users) and infrastructural variables (e.g., lane width, presence of median barriers). Longitudinal extensions would allow investigation of possible temporal lags, such as whether increases in traffic volume translate into higher fatality counts only after several months or years. Additionally, applying the same multi‑indicator framework to other national contexts would enable robust benchmarking and facilitate cross‑country learning.
Implications for SDG 3.6
The empirical evidence presented here underscores that achieving UN Sustainable Development Goal 3.6—halving global road‑traffic deaths by 2030 (World Health Organization, 2023)—requires policies that specifically target the severity of crashes rather than merely reducing overall crash counts. By adopting a composite RTS index and prioritising severity‑focused interventions, governments can develop measurable, evidence‑based road‑safety programmes capable of delivering the required reductions in fatality rates.
References
Bollen, K. A. (1989). Structural equations with latent variables. Wiley.
Chen, Y.-H. (2020). Development of a multi‑indicator road safety assessment index for ASEAN countries.