This study aims to assess the impacts of glass cockpits on pilot performance as perceived by respondents drawn from pilots currently flying these aircraft. The objective of this study include the identification of: i) pilot’s perceptions regarding the effect of glass cockpits on their performance; ii) the safety aspects derived from these perceptions; iii) the effect of pilot performance on aircraft safety; iv) the human factors affected by glass cockpits; v) the migration by airlines and manufacturers to glass cockpits; and vi) the human performance perspective of the advantages and disadvantages of the use of glass cockpits.
Chapters one and two introduced existing literature and previous research undertaken in relation to the impact of glass cockpits on pilot performance with a particular focus on the aforementioned objectives. This methodology chapter explains each stage of the research, data collection and analysis undertaken in conducting the research.
2 Qualitative Research
When undertaking research, there are two main research possibilities – quantitative and qualitative. Louis (1981) describes them as both excellent ways of tackling research effectively. He distinguishes between the two by naming quantitative as “inquiry from the outside” and qualitative as “inquiry from the inside”. By this, Louis is referring to the researcher’s involvement and immersion within the data collection – the “inside” approach involves the researcher more, often leaving a potential hazard of bias. Whereas the “outside” approach enables the researcher to isolate the phenomenon of bias because of the lack of feedback associated with quantitative research.
Qualitative research is often used as “a form of systematic empirical inquiry into meaning” (Shank, 2002, p. 5) – it is useful for those wishing to complete research in a methodical, ordered and planned way. Authors (Bryman et al 1988; Punch, 1998; Maykut and Morehouse, 1994) all describe qualitative research as a valued method of collecting a bettered, in-depth range of data. However, with the advantage of enriched data comes the disadvantage of researcher bias. As qualitative research was chosen, the researcher had to identify and accept that certain areas of the data collection would be open to bias, either value or objective (Hussey and Hussey, 1997).
Referring to Johnson (1994), “the selection of the research method is a crucial element in the planning of an investigation” (p.174). Johnson also suggests that using a variety of research methods is beneficial to gaining an adequate understanding of the objective results – as suggested in my research proposal. However, it became evident that my access to the great number of “glass-cockpit-flying pilots” needed to complete questionnaires successfully was limited. It was therefore decided that the best course of action was to select a qualitative approach, and to interview the pilots that met the selection criteria highlighted in the “Research Approach” section of this methodology.
Kumar (2011) defines an interview as “any person-to-person interaction, either face to face or otherwise, between two or more individuals with a specific purpose in mind”. Reasons for employing an interview to collect data a numerous. Depending on the topic discussed, it is possible to obtain in-depth information by probing; an interviewer can often add to data collected not only by verbal means, but also by non-verbal reactions displayed by the interviewee; and, because the interviewer can alter the order / wordings of the questions, it is less likely that the interviewee can misunderstand the question.
The author chose to utilise a semi-structured type interview as a hybrid of the two extremes – unstructured and structured – as this would give the ‘best of both worlds’. Although a structured interview would i) establish a good base for the interviewee to explain exactly what the author felt was needed, ii) provide uniform information, in the form of the comparability of the data collected and iii) enable the interviewer to collect the data without having the necessary interview skills for an unstructured interview, this method was discarded as it was nearly impossible for the interviewer to probe new ideas formulated by the answers of the interviewee. For many of these reasons, also, the unstructured interview was, too, rejected. Referring to point iii), this existed as the main reason to not chose the unstructured type interview as this was the author’s first attempt at an academic interview. Thus a semi-structured, or hybrid data collection model was developed: an element of flexibility was permitted to enable interviewees to include relevant information pertinent to the topic in discussion (Patton, 1990; Bernard, 1988).
In this respect, the researcher would include an interview guide – as found in a structured interview. However, this was, as said, a guide: the interviewer and the interviewee were able to establish and discuss any points that came up in conversation, some of which were new to the interviewer or perhaps the interviewee.
A predominately qualitative research approach was used throughout this study. In this respect, interviews were used as the data collection tool through which the current pilot perceptions of glass cockpits were identified. This method was selected because:
It allowed flexibility to follow unexpected ideas during research and explore processes effectively;
It gave sensitivity to contextual factors;
Also, the ability to study symbolic dimensions and social meaning;
o to develop empirically supported new ideas and theories;
o for in-depth and longitudinal explorations of leadership phenomena; and
o for more relevance and interest for practitioners.
-Bryman et al, 1988
Although it is arguable that a questionnaire would have suited this type of study, there were certain concerns with finding the multiple pilots needed to complete the surveys – particularly given the high security status post September 11th.
4 Research Sample
Consideration into finding the most applicable people for inclusion in this study was needed to ensure that the author chooses the candidates with the required information to achieve the objectives set for the study (Kumar, 2011). This strategy is useful to assess the pilot perceptions of glass-cockpits as, although much researcher has been conducted into glass cockpits, very little has been done into the perceptions of pilots regarding them.
The population of this study consisted of 5 interviewees. Purposive sampling was used to identify the particular pilots needed for interview, the selection criteria included:
Have flown a glass cockpit within the previous 6 months
Have flown an analogue cockpits previously
Willing to participate in the interview
4.1 Rationale for selection criteria
Firstly, by defining “previous 6 months” as recent, then we are able to understand that the interviewee must have recent experience in order to acknowledge the disadvantages and advantages of glass cockpits as experienced by them – the longer it has been since the stimuli, the less effects the subject will retain in memory. The rationale for selecting the second criterion was simple: although this study is an exploritive study into glass cockpits, the interviewees must have experience with the older fashioned analogue type cockpits in order to have a valuable opinion regarding their own perceptions. The final criteria suggests that every pilot’s participation is critical to this study – since the research involved potentially lengthy interviews, every interviewee would need to be willing to take the time to discuss the topic at hand (Chaffee, 1995).
Advantages of performing interviews with an acquaintance are well documented by authors. Roger (1988) implies that ‘social norms’ constrain strangers from interacting in a natural way and Coates (1996) adds that the presence of recording equipment, and people’s perceptions to it, is less obvious when dealing with a friend or group of friends – ultimately, by interviewing pilots whom were personally known the the researcher, it is possible to allow the interviewee to become more relaxed and less inclined to shy away from sensitive questions.
After designing an interview, but before putting it into practice, there was a chance to try out my interview on a potential subject. This pilot study was a opportunity to develop the interview so that, in future interviews, any problems would be ironed out and questions could be arranged in an order that would suit a naturally flowing conversation. Hundley (2001) suggests that performing multiple pilot studies is advantageous to the researcher, however, due to time constraints only one pilot interview could be undertaken before it was necessary to move onto the actual interviews. That pilot study changed the order of questions within the interview, as well as a few question wordings to make it more logical and easier for the interviewees to understand.
4.2 Administration of the Research
When beginning the interview stage of the project, a way of recording the conversations was necessary – this came in the form of a Dictaphone. This made the process of analysing results much simpler than having to take notes and understand them later. Of course, the participants were given full authority on whether or not they wanted the interview recorded – obviously it was competently explained to them that the interview was completely confidential via an interview confidentiality letter (appendix B). Where possible, the interviewees were given the questions via email (appendix C). The author decided to do this in the hopes that the interviewees would provide more concise and clearer answers.
Arranging meetings with pilots was, at first thought, something of a given. However, the pilots whom the author thought were once suitable for interview declared that they had no prior, or very limited, experience with glass cockpits. It was up to me to find suitable alternatives.
Luckily, the pilots who had no experience were able to refer me to their friends and colleagues who did have experience with glass cockpits – the author was able to arrange contact with them via email and telephone. Bennett (1994) discusses the options to improve the chances of the potential interviewees responding to my emails and calls. He states that, by explaining the following points, the interviewee will be more inclined to respond accordingly:
Imparting the conviction that the investigation is a worthwhile piece of work and the investigator a competent person to carry it out;
Explaining why the investigation seeks the co-operation of the persons or institutions being approached;
Indicating the use to be made of the eventual research material.
– Bennett (1994, p.174)
Interviews were planned to take place face to face, at a place convenient to both the interviewer and the interviewee. A suitable quiet environment was to be used, in most cases this was arranged in advance – but some were very short notice so the choice of location was somewhat improvised. This is for the interviews that had taken place in person, but as discussed earlier email interviews had been arranged. Due to the nature of email, there was no way of enforcing the setting – so it was left up to the interviewee to chose a suitable place to respond.
I was the interviewer in all bar the two email interviews, it was therefore my duty to maintain the correct track throughout as well as to ensure the interviewee’s were giving suitable responses to questions. Two of my potential interviewees had requested, due to time-zone differences, that they complete the interview via email – I was more than obliged to do so, so the interviews were sent to them along with appendix B (interview confidentiality letter). Unfortunately, there was no possible way of administrating the email respondents – other than reading through their answers and responding to them with any queries.
Each face to face interview lasted between 10 and 20 minutes and were conducted between February and March 2011. The email interviews were between 1000 and 1500 word long and were sent out during the beginning days of March for a expectant reply by the end of March. Other than the emails, no names were used during the interviews – this left the researcher an option to code the names for the data analysis: Email interview 1, Interview 1 etc…
It was essential for the researcher to transcribe the interviews as soon as practicably possible as too not forget any non-verbal communications made, as discussed earlier. It is also for this reason that the researcher did not complete any more than one interview per day, or per transcribe. i.e. the researcher would only move onto another interview when the previous one was transcribed completely.To avoid recall error as defined by Kumar (2011): “error that can be introduced in a response because of a respondent’s inability to recall correctly its various aspects when replying” when possible, and if time permitted, interviewees were given the opportunity to examine the interview transcripts for conformation and approval.
5. Research Analysis
Kumar’s(2011) four step research analysis was adopted to efficiently summarise and analyse the data collected. The steps were identification of the main themes; assignment of codes for the main themes; responses to the main themes and; integration of themes and responses into the text.
The first step involved carefully reading through the interviews collected by each interviewee, understanding the meaning that they communicate (the use of language to express themselves) and categorising the discovered ‘broad’ themes. The second step uses coding to identify the amount of times, or how frequently, a certain phrase or wording has occurred and pursuing the more specific themes from these. Step 3 analyses the interview scripts collected and assigns certain responses and phrases under the themes chosen in step 2. Finally, the fourth step uses the responses that fall within different themes collected in step 3 and integrates them into the text of the dissertation.
Blaxter et al (2006) suggested that most research could be affected by opinions, beliefs and motivations of those involved – it was essential for the undergoing researcher to remember this throughout this study because it had the possibility to bias results by means of altering the questions asked to accommodate preconceptions known by the researcher.
6. Concluding Points
To summarise, this chapter has covered the beneficial aspects and disadvantages of qualitative research, semi-structured interviews and data analysis; how purposive sampling was engaged as a way of implementing assurance that interviewee’s could offer their perceptions of glass cockpits; and the various limitations imposed onto the researcher, interviewees and ultimately, the study itself.