Desktop version

Home arrow Management

  • Increase font
  • Decrease font


<<   CONTENTS   >>

Creating Future Plans: Shared Understanding

Mental models and the manner in which they represent a shared understanding of the world were discussed in a previous chapter. In Excerpt 8.5, San Juan АТС told the crew to expect the ILS approach to Runway 10. This was in accordance with the flight plan filed by the crew before departure. The statement made by the controller was confirming that the expected plan of action still held (i.e. the runway in use had not changed since the crew departed Mayagiiez). Excerpt 8.8

continues the interaction we have already seen in Excerpt 8.3, where the FO gave his approach briefing. The captain interrupts with a comment about the more likely runway.

Excerpt 8.8

C: Approach briefing [Control]

FO: Uhhh, we’re doin’ the ILS ten [Explicit future goal]

C: Yup. You’ll probably circle to eight [Implicit goal]

FO: Ok. [Verification]

C: Probably be Lagoon [Restated implicit goal]

The Lagoon Visual approach is a published arrival procedure that, initially, brings the aircraft inbound on the ILS Runway 10 heading and then, passing over the western edge of San Jose Lagoon, the aircraft is turned left onto the approach to Runway 08. Later, in Excerpt 8.9, as they begin their approach, the captain effectively negotiates a change of runways.

Excerpt 8.9

C: San Juan Tower Eagle 401 ILS runway one zero [Explicit goal]

with eight in sight. [Implicit goal]

АТС: Eagle flight 401 San Juan Tower cleared to land

runway one zero {...} [Explicit goal]

C: ‘kay. One zero and eight is not available? [Implicit goal]

АТС: Eagle flight 401, uh, confirm you have [Verification of

eight in sight? implicit goal]

C: Affirm [Verification]

АТС: Eagle flight 401 change runways, cleared [Goal state change]

lagoon visual runway eight approach, runway eight cleared to land.

The company gates at this airport are adjacent to Runway 08 and, so, a landing on this runway, as opposed to Runway 10 saves a long taxi. It is highly probable that the captain always intended for the FO to land on Runway 08 as he had even mentioned that this was common practice before they had even taken off from Mayagiiez. He failed to make his intentions clear to the FO.

Social Chat

The study by Gougen and Linde commented on the role of non-operational utterances in sustaining alertness during low-workload phases of flight. Given that communication seems to have evolved to support group functioning, it is hardly surprising that social chat is a normal part of human interaction. Communication can reduce tension and build a positive group atmosphere, but unfocussed chat at the wrong time can be a distraction. The ‘sterile cockpit’ concept, which tries to limit non-essential chat at critical stages of flight, is an example of a bureaucratic response to what is considered to be inappropriate behaviour.

The crew of the ATR 72 engaged in extensive social chat, most of it from the captain and directed at the FO. We see the crew discussing points of technique (the FO had trained on the ATR 42 and only had 20hours in total on the ATR 72), differences between training and line practice (work as imagined and work as enacted?), and conveniences such as the use of flight deck power outlets to recharge mobile phones and to power DVD players to watch movies in flight.

What ‘Normal’ Communication Looks Like?

I mentioned that the San Juan accident was one of those rare occasions where we have an almost complete transcript of the crew’s interaction and, in Table 8.3,1 present an analysis of the communication acts using my four-factor taxonomy. Of course, just as there is no such thing as a ‘correct’ error rate, there is, equally, no preferred distribution of speech acts. Although context-dependent, it is rather assumed that an efficient crew will use communication as a tool in an appropriate manner, as dictated by procedures and circumstances. However, I think we can draw two significant observations about the San Juan data. First, given that we saw earlier that the captain’s implicit assumption about the probable runway at San Juan was a plausible contributory factor in the accident, the frequency of directed speech acts intended to clarify goals and intentions was the lowest of all. Sexton and Helmreich (2000), analysing transcripts from a simulator study, suggest that increased communication between crew will reduce error because of the increased predictability of behaviour induced. Predictability can only flow from a shared understanding of the goal. The second observation concerns the high frequency of social speech acts. In particular, 29% of speech acts during the descent and final approach phases fall into this undirected category. While social interaction serves a purpose, the operational focus at the appropriate time is essential.

What can line operations safety audit (LOSA) tell us about ‘normal’ communication? In the sample of 100 sectors discussed in the previous chapter, 42.57% of

TABLE 8.3

Distribution of Speech Acts

Ground

T/O & Climb

Cruise

Descent

App/land

Total

Time (%)

14.9

16.8

12.4

43.7

12.04

External (АТС/ cabin/Ops)

14

17

7

46

9

93 (20.08%)

Control and verification

53

14

7

35

33

142 (30.66%)

Option

selection

4

8

4

8

40

64(13.82%)

Clarify goals and intentions

0

2

9

6

8

25 (5.3%)

Social

22

20

21

51

25

139(30.02%)

Total

93 (20.08%)

61 (13.17%)

48(10.36%)

146 (31.5%)

115 (24.8%)

463

‘errors’ related to communication. This high rate simply reflects the fact that communication forms a significant part of the work of the crew and, importantly, it is also accessible to observation and, thus, audit. Unfortunately, LOSA gives us an incomplete understanding because it only captures departures from proceduralised communication and does not capture the quality of communication in the sense that it has been discussed so far in this chapter.

Given these caveats, communication was either within the crew (endogenous), such as procedural call-outs, or between the crew and external agencies (exogenous), such as АТС calls and acknowledgements. The ratio of endogenous to exogenous communication was 2:1. This simply reflects the density of the АТС environment the flights were operating in. Shorter sectors increase the time spent dealing with АТС. Looking at endogenous behaviour first, of the four factors in my model, control and verification were easily seen, as were formal structures associated with clarifying future goals. Controlling action (calling for a checklist) and verifying progress (calls associated with configurations and status) make up the largest cluster of interactions. Control can be immediate (a call) or delayed (actions arising from a prior briefing). Briefings, of course, represent opportunities to clarify future goals. In the sample, there were two points in each flight where a briefing was mandated: before departure and at the top of decent, prior to the arrival. In the sample, we observed 200 briefings. In 7% of briefings, either the structure of the briefing was incomplete or it was delivered at the wrong time (typically in the descent because of an intervention by АТС). Errors do reveal something of the fragility of performance that flows from individual variability (incorrect phraseology, incomplete or incorrect messages) and from the challenge of balancing competing demands. Dealing with АТС while completing other tasks being a classic example. Missed procedural call-outs were the largest category of endogenous errors. Interestingly, observations of ‘social’ communication acts comprised just 1.7% of all episodes, but I am inclined to think that this value is depressed simply because the crew knew they were being audited.

When dealing with external agencies (exogenous), typically АТС, we see the impact of contextual facts such as local accents, available technology, traffic densities and, hence, controller workload. In this instance, the notional ‘error rate’ is as much a measure of regional infrastructure as it is of crew performance. Incorrect read-back was the largest cluster of exogenous errors.

We saw in Chapter 5 that LOSA allows you to tag errors in terms of crew responses. Did they actively respond or did they passively ignore the error, for example. The discussion so far has only looked at errors where the crew responded in an active sense, which comprised 62% of all communications errors. If I turn now to those episodes where the crew response was passive, we see that the endogenous to exogenous ratio changes to 7:1. In both cases, missed procedural call-outs were the largest cluster (44.5% of active and 59.8% of passive errors). Interestingly, missed calls have been cited as evidence of poor ‘monitoring’ in some studies, but it seems to me that crew elect to expend effort based on a number of possible factors: alignment with the intent behind the requirement (why call if the automation will do the job?); workload (trading off a verbalisation against a more important use of attention); crew synergy (we both know what’s going on so why chatter needlessly?).

Normal work, then, affords space for the crew to modify responses - the extent of compliance w'ith prescribed communication protocols - in accordance w'ith local circumstances.

 
<<   CONTENTS   >>

Related topics