Irregularity of Flight and Slow-Flight Practice Evident for a Subset of Private Pilots—Potential Adverse Impact on Safe Operations

Abstract

Background: General aviation pilots are, anecdotally, referred to as “weekend warriors” due to their flying infrequency. Considering that flight skills erode with irregular practice/reinforcement, we determined whether private pilots (PPLs) fly/train sufficiently to operate safely in the context of slow flight, a skill critical for safe operations and which rapidly atrophies with <~51 h flight time/8 months per prior research. Method: Slow-flight-related aviation accidents (2008–2019) were per the NTSB Access^R database, and fatal mishap rates were calculated using general aviation fleet times. Eight-month flight histories of airplanes in single PPL ownership were captured retrospectively using FlightAware^R. PPL survey responses were collected between January and March 2025. Statistical tests employed proportion/Independent-Samples Median Tests and a Poisson Distribution. Results: The slow-flight-related fatal accident rate (2017–2019) trended downwards (p = 0.077). In-flight tracking of 90 airplanes revealed an 8-month median flight time of 6 h, which is well below the aforementioned 51 h requisite for safe operations. Of the aircraft flown < 51 h, only 9% engaged in slow-flight practice. In the online survey, only the upper quartile of 126 PPLs achieved the aforementioned time requisite for preserving slow-flight skills, but nevertheless, 89% of respondents attested to being flight-proficient. Conclusions: Persistence in slow-flight-related fatal accidents likely partly reflects PPLs’ deficiency in in-flight time/slow-flight practice.

1. Introduction

Civil aviation can be broadly categorized into air carrier (in common vernacular referred to as airlines), charter, and general aviation, each conducted under specific operational regulations [1,2,3]. Regarding general aviation, this sector of aviation is comprised mainly of single piston engine-powered light airplanes (<12,501 lbs.) [4] and in the USA operates under the federal regulations encapsulated in 14CFR Part 91 [1]. Unfortunately, the safety of general aviation pales in comparison with that of air carriers, as evidenced by up to a 262-fold higher fatal accident rate [5,6]. Multiple factors contribute to this disparity in flight safety. For example, air carrier pilots undergo much more extensive training requiring advanced certification (airline transport pilot vs. private pilot certificate) and experience (a minimum of 1500 vs. 40 h in the USA). Additionally, air carrier operations must be conducted with a minimum of two pilots [3], whereas no such regulation applies to general aviation. Thirdly, redundancy of aircraft systems is mandatory for transport category aircraft used in passenger transport (14CFR Part 25) [7], whereas, in contrast, minimal system backup is required for light aircraft certification (14CFR Part 23) [8].

The flight frequency/time accrued by private pilots (PPLs) over a given period represents another potential contributing factor to the inferior general aviation safety record. Put another way, do private pilots fly sufficiently to maintain proficiency (anecdotally, PPLs are sometimes referred to as “weekend warriors” due to their tendency to fly only over these two days)? This question has, to date, received scant attention and is pertinent since several prior studies have indicated that flight skills erode over time if practice/reinforcement is irregular [9,10,11,12]. For example, flight data monitoring of 123,140 air carrier flights demonstrated a clear impact of the COVID-19 pandemic-related flying hiatus on in-flight aircraft exceedances [12]. Degraded knowledge of flight data automation was also evident in that study [12]. Along similar lines, earlier research documented an erosion of flying skills for pilots returning from Vietnam who were assigned non-flying duties for a period of 8 months [13]. “Do PPLs accrue sufficient time in a given period to remain proficient?” is a question that represents the main thrust of the current study. Relevant to this question is an earlier report [9] in which PPLs (comprised of two cohorts—one accruing ~51 and the other ~5 h in an 8-month period post-certification) were re-evaluated for flight maneuvers tested as part of the initial PPL practical exam [14]. Perhaps not surprisingly, the group with the greater flight time (~51 h), accrued in the 8 months post-certification, showed a higher success rate in completing those flight maneuvers compared with PPLs who accrued only ~5 h over the same time frame. Notably, of all the flight tasks re-examined, slow flight (herein also referred to as maneuvering flight) represented the one skill most degraded by diminished flight time. This observation is important since this skill set is critical for safe operations during the traffic pattern/circuit when the airplane is being maneuvered at low speed and at low altitude prior to landing. In this phase of flight, poor slow-flight skills may result in an aerodynamic stall and a fatal mishap due to insufficient altitude to recover [15].

Accordingly, the question we sought to answer herein is whether PPLs fly/train sufficiently to operate safely in the context of minimizing the erosion of flight skills (slow-flight skills), the loss of which may result in fatal consequence [15]. Towards this end, we used a two-pronged strategy employing (i) in-flight tracking of aircraft and (ii) an online survey of PPLs.

2. Materials and Methods

2.1. Accidents

To identify mishaps in which slow flight (maneuvering flight) was causal or contributory, the NTSB aviation accident Access^R database v 2019 (August 2024 release) was downloaded [16] and queried for records (2008–2019) involving piston engine-powered airplanes operating under 14CFR Part 91 (i) that were coded for low altitude maneuvering (452 ***) and (ii) for which the narrative included the word “stall.” A parallel search was conducted as above but combing the narrative section for the simultaneous presence of the terms “stall” and “traffic pattern.” Records from both search results were merged. Injury severity was per the NTSB accident database. Accidents related to equipment/installation failure and pilot incapacitation were deleted. For accidents unrelated to slow flight, the aforementioned search was repeated but excluding the above-stated criteria. Mishaps related to slow flight, identified above, were then deleted from the returned accident list. To calculate accident rates, total hours for fixed-wing aircraft with one or more engine(s) per the general aviation survey (but restricted to 14CFR Part 91 operations) was used as the denominator [4]. Fleet time for 2011 was interpolated from the years 2010 and 2012.

2.2. Selection of Single-Engine Airplanes for Automatic Dependent Surveillance-Broadcast (ADS-B)-Out Flight Tracking

This procedure has been described by us earlier [17]. Briefly, since general aviation aircraft operating within Mode C terminal areas have to be ADS-B-Out-equipped [18], four such areas within the sunny belt of the United States (allowing for year-round operations of light aircraft, which rarely are equipped for icing conditions) were arbitrarily selected (Los Angeles, KLAX; Phoenix, KPHX; Houston, KIAH-KHOU; and Miami, KMIA). Single-piston-engine fixed-wing aircraft (>199 horsepower) were identified using the FAA airplane database [19]. The list was then filtered for those with a valid air registration certificate and also in single-individual ownership. Using the FAA aviator database [20], aircraft were then cross-referenced with non-instrument-rated private pilots for the corresponding states (California, Arizona, Texas, and Florida). The resulting list was thereafter further restricted to aircraft registered in ZIP codes underlying the Mode C veils of the aforementioned airports, culminating in a total of 90 airplanes.

2.3. In-Flight Tracking of Aircrafts

Eight-month airplane flight histories were retrospectively captured using FlightAware^R (Houston, TX, USA), a publicly available web-based tracking tool [21], with the study period (over which flights were undertaken) spanning 24 February 2024–8 March 2025. FlightAware^R provides ADS-B-Out-derived position, times, ground speed, and altitude data in both numerical and keyhole markup language (kml) image formats, both of which were downloaded for each flight. The resolution of ADS-B-Out altitude data is to the nearest hundred feet and is reported once or multiple times for each minute. To preserve anonymity, all aircraft were assigned unique identifiers. Aircraft involved in instrument flight training, as evidenced by either conducting hold procedures (apparent as a “racetrack” in the kml file graphically overlayed on Google Earth), instrument approaches, as described by us previously [22], or a filed IFR flight plan (per FlightAware^R), were removed from the analyses. A practice slow flight was defined [14] as any within 10 kts or less above the stall (landing configuration) speed (V_SO) of the airplane in which the maneuver was completed no less than 1500′ above the ground and which was preceded by at least one 90° clearing turn. True airspeed was determined by adjusting ground speeds (per FlightAware^R) for winds aloft using archived METAR data from the nearest weather station [23] with the aid of an E6B electronic flight computer. V_SO of the aircraft tracked were determined from various online sources, including pilot operating handbooks/flight manuals.

2.4. Online Survey of PPLs and Human Subjects Approval

An online survey of PPLs was constructed in Qualtrics^R and approved by the Embry-Riddle Institutional Review Board (IRB # 25-089). The survey was pretested by five FAA Safety Team representatives, all of whom were active general aviation pilots. Recruitment was carried out via internet aviation newsletters/forums (AOPA and AvWeb Flash and the Blancolirio Youtube^R channel). Responses were collected between 29 January 2025 and 5 March 2025.

2.5. Statistics

In this study, 2 × 2 contingency tables were tested for differences in proportions using Chi-Square and Fisher’s Exact Tests [24]. A non-parametric test (Independent-Samples Median Test) was used to test statistical differences in median values. Accident rate changes over time were tested for statistical differences with a Poisson Distribution [25]. All statistics were undertaken with SPSS (v27) IBM^R (Armonk, NY, USA), with the exception of power analyses for requisite population size, which was conducted using G*Power (v3.1.9.4).

3. Results

3.1. Slow-Flight (Maneuvering-Flight) Accident Rate and Injury Severity—A Longitudinal Study

Considering that degraded slow-flight (maneuvering-flight) skills can have lethal consequences [15] and that the Federal Aviation Administration (FAA) has advocated for [26] the regular practice of this task since 2006, we first determined whether the rate of general aviation accidents related to a deficiency in this skill diminished thereafter. We selected 2019 as the final year due to the impact of the COVID-19 pandemic [27].

The rate of general aviation accidents related to slow flight (also referred to herein as maneuvering flight) was unchanged (Figure 1) for the period spanning 2008–2016 but trended downward thereafter. Although the reduction for the 2017–2019 time frame did not reach statistical significance (p = 0.077) using the initial time period (2008–2010) as a referent, a power analysis indicated an insufficient population size. In contrast, the rate of general aviation accidents due to other causes showed a strong statistical decline for the periods spanning 2014–2016 (p = 0.016) and 2017–2019 (p < 0.001).

Next, we determined whether the lethality of accidents in which deficient maneuvering flight was causal or contributory has changed over the same time period. Perhaps not surprisingly, for all four time frames, the fraction of fatal accidents related to slow flight were statistically (p < 0.001) higher (3.3–4.1-fold) than those for fatal mishaps due to other causes (Figure 2). However, and somewhat encouragingly, the proportion of fatal accidents in which slow flight was implicated trended downwards from 0.78 to 0.53 over the study duration. The 28% reduction evident for the most recent period was statistically significant (p = 0.012) using the 2008–2010 time frame as a referent. Nevertheless, it is clear that the risk of a fatal outcome for mishaps related to deficient maneuvering-flight skills is still substantially higher than for accidents caused by other means.

The aforementioned high fraction of fatal slow-flight-related mishaps raised the question as to whether PPLs fly sufficiently (i.e., accrue > 51 h flight time in an 8-month period) and/or practice this maneuver to offset the decay in this skill as reported earlier [9]. Towards this end, two strategies were employed as described below. First, the 8-month flight histories of a randomly selected cohort of airplanes in single-individual PPL ownership were collected using a flight-tracking service. Second, a survey of PPLs as to their annual flight histories and attitudes to maintaining proficiency was conducted.

3.2. Flight Times and Practice of Maneuvering Flight as Determined with ADS-B-Out Data

In the first approach, 8-month flight histories of 90 randomly selected single-piston engine aircraft registered in four Mode C airspace areas (and thus mandated to be ADS-B-Out equipped) located in the sunny belt of the USA and in single-individual ownership were retrospectively collected using FlightAware^R [21]. The choice of an 8-month period to track aircraft was based on the earlier study [9] showing an atrophy of maneuvering-flight skills for PPLs flying less than 51 h over this time frame. In total 1276 flights were analyzed. Of the 90 aircraft, 24 (27%) were not flown over the eight months each was followed. Moreover, a median flight time of a mere 6 h was evident for this aircraft cohort (Figure 3), well below the 51 h previously shown [9] to minimize the loss of slow-flight skills (dashed horizontal line). In fact, only 5 of the 90 (6%) aircraft in single PPL ownership were flown 51 h or more across the eight months each was tracked (Figure 3).

We entertained the possibility that PPLs flying insufficiently to maintain slow-flight skills might, nevertheless, practice this maneuver to achieve the required proficiency. This possibility was examined by using ADS-B-Out flight tracking data (FlightAware^R) again. Herein, slow flight was defined as one in which the aircraft was operated at 10 kts or less above the stall (landing configuration) speed (V_SO) with the maneuver completed no less than 1500′ above the ground and preceded by at least one 90° clearing turn (Figure 4). Note that airplane airspeed was corrected for winds in effect at the time by using the nearest METAR data [23]. Of the seventy-four aircraft operated less than 51 h in the 8 months each was flight-tracked but only seven (9%) engaged in the practice of slow flight. Note that the lower count (74 vs. 85) of aircraft, in which slow flight was determined, reflected the presence of some non-type certificated/experimental aircraft in the study cohort; published stall speeds are not available for such airplanes. Thus, in the current study, by far the majority of aircraft registered to PPLs are neither flown sufficiently nor used to practice slow flight for the pilot–owner to maintain a skill necessary for safe operations.

3.3. Flight Times and Attitudes to Maintaining Proficiency—A Survey of PPLs

In a parallel strategy, a survey of PPLs was conducted as to their annual flight times and attitudes to maintaining proficiency. A total of 126 non-instrument-rated PPL respondents, drawn from across the USA, completed the survey (

Table 1. PPL survey.

		PPL Respondents Location
Column	1	2	3	4
		All US States	CA-AZ-TX-FL	p Value
Pilot Age (Yrs)	n	126	25	0.585
	Median	63	66
	Q1	53	60
	Q3	73	72
Total Flight Time (h)	n	126	25	0.688
	Median	813	550
	Q1	405	325
	Q3	1500	1420
Years Flying Since PPL Certification	n	126	25	0.487
	Median	20	15
	Q1	8	3
	Q3	34	32
Annual frequency recurrent flight training with CFI/CFII (excluding training for an advanced certificate)?	0× n (%)	53 (43%)	10 (40%)	0.800
	1× or More n (%)	71 (57%)	15 (60%)
Annual Flight Time (h)	n	126	25	0.536
	Median	60	65
	Q1	36	50
	Q3	84	100
How many hours (annually) do YOU consider requisite for operating safely in the flight environment?	n	122	23	0.244
	Median	36	50
	Q1	24	25
	Q3	50	50
In your opinion, do you fly sufficiently to operate safely in the flight environment?	Yes n (%)	109 (89%)	23 (100%)	0.223
	No n (%)	13 (11%)	0 (0%)

PPLs were queried on their flight history and attitudes toward maintaining proficiency. Differences in median values or proportions between the two PPL populations (US state-wide and California (CA)/Arizona (AZ)/Texas (TX)/Florida (FL)) were tested using an Independent-Samples Median Test or Chi-Square/Fisher’s Exact Tests, respectively (h, hours; n, respondent count; Q1 and Q3, first (25%) and third (75%) quartiles, respectively; CFI/CFII, certified flight instructor/instrument).

) between 29 January and 5 March of 2025. Survey respondents were experienced aviators, as evidenced by an accrued (median value) total flight time of 813 flight hours (Table 1, column 2) and having flown for 20 years (median value) since PPL certification (Table 1, column 2).

Interestingly, the median annual flight time (60 h) reported by PPL respondents across all US states (Table 1, column 2; Figure 5) far exceeded (p < 0.01) the 9 h (adjusting the 6 h/8 months to an annual value) determined using in-flight aircraft tracking. Despite the higher median reported flight time, only approximately the upper quartile (Q3) of respondents (Table 1) achieved more than the 77 h of flight time (51 h over an 8-month period adjusted annually) previously shown to forestall the rapid atrophy of slow-flight skills [9].

The divergence in data between that generated by ADS-B-Out flight tracking and that from the survey was surprising. We initially entertained the notion that this difference might reflect the capture of two geographically distinct PPL cohorts—US statewide for the survey in contrast with four states per the ADS-B-Out flight tracking. To address this possibility, we analyzed survey responses restricted to those states (California, Arizona, Texas, and Florida) in which ADS-B-Out flight tracking was used. However, none of the parameters were statistically different when comparing respondents across the USA (Table 1, column 2) with those in the four aforementioned states (Table 1, column 3). Thus, it is unlikely that the divergence in data between ADS-B-Out-tracked aircraft in individual PPLs’ ownership and survey respondents is caused by geographical differences between the two populations.

Encouragingly, survey respondents showed a positive safety attitude as evidenced by 57–60% of PPLs indicating recurrent flight training one, or multiple, time(s) per year with a certified flight instructor (Table 1, columns 2 and 3). When queried as to whether, in their opinion, they flew sufficiently to remain proficient, 89–100% of PPLs (Table 1, columns 2 and 3—US-wide and Sun Belt US state restricted, respectively) responded in the affirmative. Importantly, however, the median annual flight times of 36–50 h (US-wide and Sun Belt states, respectively) were cited as requisite for maintaining proficiency. These values were well below the 77 h (annually adjusted from the 8-month value of 51 h) required to slow the atrophy of slow-flight skills as determined in a prior study [9].

4. Discussion

Whether PPLs fly sufficiently/practice to remain safe in the flight environment has not previously been reported on. At least per our study cohort, the majority of PPLs fly irregularly and rarely partake in slow-flight (maneuvering-flight) practice. Considering previous studies documenting the necessity of frequent flight exposure for minimizing the erosion of flight skills [9,10,11,12], such infrequency may very well be a contributing factor to the much higher fatal accident rate evident for general aviation when compared with air carrier operations [5].

The divergence of the data between that garnered by in-flight tracking and the online survey merits discussion. Aviators in the latter cohort reported greater flight time and more frequent recurrent training. That said, it should still be recognized that the median annual flight time (60–65 h) reported by survey respondents was still below the 77 h (annually adjusted from 8 months) required to delay the atrophy of maneuvering-flight skills [9]. The disparity in flight times between that achieved by in-flight tracking and reported in the survey could reflect one or a combination of reason(s). First, admittedly, the two PPL cohorts were relatively small. Second, “over-claiming” [28] in surveys is well-recognized, and due to the anonymity of the survey, flight data for respondents could not be verified. Third, considering that the survey was advertised through online aviation forums, it may be that we inadvertently selected for a skewed population represented by pilots passionate about this activity and who consequently fly more often. Finally, in-flight tracking was restricted to aircraft with engines of 199 horsepower or greater, whereas no such criterion was applied for the survey.

We considered the possibility that PPLs who flew irregularly, per in-flight tracking, might, nevertheless, maintain their proficiency by frequent practice of the perishable slow-flight maneuver. Indeed, the Federal Aviation Administration recommends the practice of maneuvering flight and aerodynamic stall recovery every 4–6 weeks as part of the visual flight profile [26]. However, based on in-flight tracking, we saw little evidence that PPLs subscribe to this recommendation. Thus, slow flight was only evident in 9% of aircraft that are in single-individual PPL ownership and were flown less than that required for persistence of slow-flight skills [9].

How applicable are our in-flight tracking findings on irregularity of flight to the general aviation population? Although the cohort was limited in size by virtue of the methodologic approach utilized for identifying aircraft likely operated by a single individual, it should be emphasized that, with the exception of engine horsepower, airplane selection was random in nature. Nevertheless, our findings argue for future research with a larger PPL cohort to determine, with greater accuracy, the fraction of the general aviation population characterized by irregular flight/infrequent practice.

Our study was not without limitations. First, the in-flight tracking data were from aircraft, not PPLs. In this regard, although the aircraft, per study criteria, were all in single-individual ownership, it is possible that the airplane was also flown by an aviator other than the owner. However, if this is the case, recorded flight times, on a per-owner basis, would be even lower than the flight hours evident in ADS-B-Out flight tracking. Another possibility is that recognizing the potential impact of their flight irregularity on safe operations, PPLs elected to fly with a flight instructor. Indeed, per the online survey, 24% of PPL respondents indicated that they undertook recurrent training 2× or more annually Second, we cannot exclude the possibility that PPLs make use of advanced/aviation training devices to practice slow flight. Unfortunately, this question was not included in the survey and is worthy of future investigations. That said, such devices are typically unable to reproduce the 3D motions of flight and physiological cues, which can lead to inappropriate pilot control inputs [29]. Third, limiting the flight tracking to aircraft > 199 hp and to the Sun Belt states raises the possibility of geographic and sociodemographic biases. Regarding the former, PPLs of higher income brackets may have less time to fly and/or undertake recurrent training. Fourth, we accept that determining airspeed using data from the most proximate METAR to correct for winds aloft is inferior to using Skew-data. Unfortunately, the latter are not archived, precluding this type of correction. Finally, the gap between the data generated from in-flight tracking and that from the survey may very well reflect a more active and safety-conscious cohort of pilots who follow aviation forums.

In conclusion, a combination of in-flight and PPL survey data argues that a substantial number of PPLs fly insufficiently to keep safe in the flight environment. Future research should endeavor to create a post-validation, automated detection tool/algorithm that could be applied to a much larger population to advance the current findings. That said, we propose the following recommendations. First, such PPLs should be educated to dispel the prevalent perception (evident from the survey) that they fly sufficiently to maintain such proficiency. Second, aviators should be encouraged to engage in recurrent flight training on a 6-monthly schedule as mandatory for pilots-in-command exercising air carrier privileges (14CFR Part 121.433) [30]. Third, PPLs should utilize the defined minimum maneuvering airspeed (a speed of no less than 4× V_S) as a minimum airspeed for slow-flight operations in the traffic pattern/circuit, a common practice for air carriers during this phase of flight [31]. Fourth, towards preventing loss-of-control accidents, reflecting deficient maneuvering-flight skills, civil aviation authorities should encourage (i) the aviation industry to port the “bank angle” exceedance aural annunciation, a technology currently restricted to transport category aircraft, to general aviation aircraft and (ii) the inclusion of upset recognition and recovery training in recurrent training. Finally, the current study emphasizes the utility of ADS-B-Out-based in-flight tracking to inform safety deficiencies in general aviation [32,33] akin to Flight Operations Quality Assurance (FOQA) employed by air carriers [34].