High-quality data is a fundamental tool for crash analyses and research necessary for effective road safety management system. It is useful for assessing performance of road safety initiatives, informed decision making and designing effective countermeasures to road crashes.
On the other hand, suboptimal crash data can derail the realisation of safer road objectives. Suboptimal data manifests in the form of inaccurate, incomplete, duplicated, or missing information. More often than not, these defects are a product of poor-quality data analytics and want of standardised guidelines and procedures for collecting, recording, reporting and/or managing information.
The Federal Road Safety Commission (FRSC), the lead organisation in traffic management on Nigerian public roads, conducts research into the state of safety on the nation’s roads based on data aggregated from multiple sources and locations. It publishes its activities in a quarter-yearly publication, The FRSC Statistical Digest, with the maiden edition published in Quarter 1, 2015.
The FRSC positioned the Digest as an informed perspective for research and planning, with robust data management as one of its cardinal plans. However, the quality of the output analysis in the Digest and determination of appropriateness for intended use is only as good as the quality of the input data. Judging by its own metrics, analysis of the reports shows the unreliableness of the FRSC’s statistics, which is epitomised by poor-quality datasets, spelling errors and missing or mismatched information. A cursory review of the publication reveals that it is nothing but a glorified catalogue of fragmented statistical posturing, lacking standardised terms or discernible analytical tools or basis for monitoring performance. In other words, it falls short of its stated objectives.
What ails the Digest can be found in the segment, Road Traffic Crashes Involving Fleet Operators (RTC), reported by the Transport and Standardisation Office. In the inaugural edition, the Office captured 62 RTC, followed by 53 in Q2. The numbers dropped to a significantly low 36 in Q3. This is a 26-point differential from the Q1 report, and a 14.25-point variation from the year-average of 50.25. Conversely, the 62 cases recorded in Q1 are 23.8 points higher than the 38.2 average recorded in the ten-year period for which statistics are available and accessible.
The FRSC did not adduce reasons for the improved road crash performance of the magnitude recorded in Q3 2015, or the intermittent RTC fluctuations or anomalies recorded in other periods. The improvement could be attributable to a variety of reasons: stricter enforcement or road-user compliance with road traffic laws; improved road facilities; better quality vehicles on the roads, etc. Of course, none of these reasons was in play as the number of crashes spiked to 50 in Q4 2015, which is close to the year-average.
The fluctuations in the number of recorded crashes continued in the reports for 2020. From a crash rate of 36 in Q1, the number dipped to 15 in Q2 – a 58.33% drop from the figure for the previous quarter. While the statistics for Q3 are not available, the figure rose to 20 in Q4. Again, the FRSC failed to adduce any reasons for such significant dissonance in crash occurrences for the period.
Apart from dissonance in RTC entries, another problem that occurred in the Digest was data misinterpretation where statistics arising from same source were interpreted differently, and produced conflicting results or misaligned Table/Chart. In context, RTC are reported in Tables and presented in corresponding Charts. Since the statistics are based on common underlying facts without add-ons, the results are expected to align. But the Digest seems incapable of normalcy.
The misalignment arose from the mode of entering crashes in the Table and Chart. Each crash occurrence is denoted independently in a column of the Table and recorded in a separate Chart. This is so, whether the crash is a single or multi-vehicle incident. In the latter case, even if the crash-involved vehicles are more than one, they are entered in a single column of the Table and Chart, but separated by slash (/) or comma (,). A problem would arise if a multi-vehicle incident that is correctly depicted in a single column of the Table is wrongly inputted into multiple columns in the corresponding Chart. The Chart subdividing would create the illusion of unconnected crashes, and double-counting.
An example of Table/Chart misalignment occurred in Q4 2018. There, a multi-vehicle crash involving fleets operated by Mass Transit and Police Van, was entered in Table 24, Column 27 as a single incident. However, in the corresponding Chart 15, it was represented or stacked into separate layers or sub-columns as Mass Transit and Police Van. Without addressing the mediocrity in identifying the fleet operators as Mass Transit and Police Van, the subdividing created an illusion of unrelated crashes, thereby resulting in 43 counts in the Table, and 44 in the Chart.
Another faux entry of a multi-vehicle, column-separated crash, involving Health Life and Yobe Line, was recorded in Q2, 2016. Entered as a single incident in Column 28 of Table 31, it was separated into Yobe Line and Health Life in Chart 22, leading to a count of 46 and 47 in the Table and Chart, respectively. A disparity also occurred between Table 24 and Chart 15 of Q1, 2019 report. There, a crash involving Nganzai Line/NPF Mopol, in Column 26 of the Table was separately entered in the Chart as NPF Mopol and Nganzai Line for 36 and 37 counts, respectively.
The most egregious cases of Table/Chart misalignment occurred where single incidents involving the same fleet operator were carved up into multiple crashes and logged as separate occurrences. For example, Table 19 of the report for Q1, 2021, reflected 52 crashes as against 54 in the corresponding Chart 10. The reason for this error is not only obvious, but lazily so. Firstly, a single vehicle crash logged in Column 27 of the Table to EEC Construction Company was logged as Company, and EEC Construction in the Chart. This created the impression of two fleet entities – Company, and EEC Construction. Similarly, in Column 38, the involved vehicle belonged to Kaduna Transport Authority, but was split into two, and captured as Authority, and Kaduna Transport in the Chart.
There is a commonality to split or faux entries, whether involving single or multi-vehicle crashes – the Table order of vehicles is inverted in the Chart. A case in point is the earlier referenced Nganzai Line and NPF Mopol in Column 26 of Table 24, Q1, 2019. This is entered as NPF Mopol, and Nganzai Line in Chart 15. For the single-variant, the name of the fleet operator is inverted and juxtaposed in reverse order, so EEC Construction Company becomes Company, and EEC Construction, and Kaduna Transport Authority becomes Authority, and Kaduna Transport (Q1, 2021). The order inversion ignores the fact that the Chart is supposed to be illustrative, or a mirror the entry of the Table.
There are other forms of inexplicable statistical variance in the reports that can only be attributed to sloppiness. For instance, Table 27 of Q1 2021 reported 52 RTC as against 29 in Chart 15; while 50 cases shown in Table 27 of Q3 2023 were entered as 25 in Chart 15.
One of the most recurrent data weaknesses of the Digest is the lack of standardised identity for fleet vehicle operators. One of the most obvious examples is Dangote Plc., the lead private fleet operator and errant road crash mogul in the nation. For every quarter of the Digest in publication, Dangote Plc., recorded fatal crashes. Given its prominence as the lead crash generator, a credible road safety management authority would have had the operator on default entry or auto dial, and assigned an identifying name to it. Instead, Dangote PLC., is referred to variously by other names as Dangote Group, Dangote Cement, and Dangote.
Another example is Nigeria Police Force. This is also identified by other different names, from its street names like Nigeria Police, Police, NPF; and pedestrian identity like MOPOL, Police Van, SWAT Police. A similar situation applies to Nigeria Armed Forces, which is also captured as Nigeria Army, Nigerian Army, Nig. Army, N/A, NA, Military.
The most-maligned fleet operator in the misidentified category is Akwa Ibom State Transport Company. In addition to its official name Akwa Ibom State Transport Company, it is identified by nomenclatures like Akwa Ibom Transport Corporation, Akwa Ibom Transport Coy, Akwa Ibom State Transport, Ibom Transport Company, AKTC Transport Company, AKTC, Akwa Ibom Trans, Akwaibom Transpot Comp, Akwa Ibom, and Akwa Ibom Mass Transit.
There is also the case of the carbonated drink company, Seven-Up Bottling Company Plc., bottlers of 7Up and Pepsi. The beverages are produced and marketed worldwide by PepsiCo, but for reasons beyond the scope of this work, are produced and marketed in Nigeria by Seven-Up Bottling Company Plc. But in a testament to the clueless of FRSC for the job assignment, the fleet operator is reported in several incidents involving its vehicles as 7UP, Pepsi, 7UP Bottling Company, Pepsi Bottling Com, 7UP Bottling Bus, Pepsi Company.
Following the trend of Seven-Up Bottling Company Plc.; Guinness Nigeria Plc., is reported as Guiness PLC, Guinness, Guiness, Guiness Nigeria and Guinness Truck; and Dufil Prima Foods, makers of Indomie Noodles, becomes Idomomie Food Comp.
This list of misspelt or misidentified names is endless. Akwa Ibom Transport Company becomes, Akwaibom Transpot Comp; Taraba State Transport Corporation is Taraba State Transport Cooperation, and Nigeria Railway Corporation, Nig. Railway Cooperation. International Breweries Plc., acquires a new moniker, Intrenational Breweries; Eastern Mass Transit is Estern Mass Transit; Nasarawa Transport is rebranded Nasarawa Transort; and Borno Express, Borno Expres.
There are also unrelatable names like, Furtunate; African Natural Resourse & Mines Ltd; Mass Transmit; Honourable Inter Acadamy; Anamcra Comfort Line; etc. Then, there are the lazy generic names like Flight; Mass Transit; CBN Convoy, Anglican Community, Country, Police Van, Plant Hire, Sawmill, Interstate 547447, Cooperative Society Limited, Correctional, NNPC Tanker, etc.
What might seem like benign mistakes actually have great repercussions. As a result of misspellings, crashes involving same fleet operators in a reporting period are sometimes treated as separate occurrences. For example, in Q4, 2015, Nasarawa Transport was wrongly spelt as Nasarawa Transort and double-captured under the different names in the same reports. A similar occurrence occurred in Q3, 2023, where Nigeria Police Force is variably reported as NPF and Police. In Q3, 2017, Borno Express was reported alongside Borno Expres. These misspellings create overcounting of fleet operators for the respective periods.
There are quarters where the Statistical Analysis are either not available or inaccessible. Cases in point include Q2, 2019 and Q3 2020. Email requests to FRSC for access to the reports were unsuccessful as the Corps failed to respond to the messages.
Defective or suboptimal crash data is susceptible to compromised statistical inferences. It limits the ability to share data among users, or to select effective countermeasures, or support the delivery of road safety initiatives. Further, such data strains the efficacy of road safety advocacy.
Reliance on statistics derived from suspect or suboptimal crash data is susceptible to compromised statistical inferences. Questionable underlying facts will almost always negatively skew conclusions, dampen responsiveness to desired initiatives, and lead to research-consumer lethargy if users have to manually identify and correct the data.
Invariably, inaccurate crash data could limit the ability to share data among users and result in lost collaboration and funding opportunities. It could also constrain the selection of effective countermeasures necessary to support the delivery of adequate road safety initiatives, and strain the efficacy of road safety advocacy.
Perhaps realising the limit of its competence or lack thereof, it would appear that the FRSC has paused or entirely ceased publication of the Statistical Digests, with the last edition published in Q1, 2024. This may well be good for the stakeholders because, junk or doctored statistical research masquerading under the authority of the state does no one any good.