Engineering as Social Experimentation

1. Engineering as Social Experimentation “To undertake a great work, and especially a work of a novel type, means carrying out an experiment. It means taking up a struggle with the forces of nature without the assurance of emerging as the victor after the first attack. - Louis Marie Henri Navier (1785 – 1836)

2. Role of Experimentation in the Design Process • Preliminary tests or simulations of concepts • Components and modules tested prior to detailed design • Cycle of test and modification through production • Beyond specific elements of design, each project taken in a totality can be viewed as an experiment

3. Contributors to Experimental Nature of Projects • Project carried out in partial ignorance • Parts functionality & availability • Luxury of waiting until all relevant facts are in not available (ability to work with partial knowledge is one talent crucial to an engineer’s success) • Outcomes of projects are generally uncertain • Unknown risk may attend even a seemingly benign project • Effective engineering depends on knowledge gained both before and after products are released • Monitoring cannot be limited to in-house development

4. Informed Consent • Keystone of properly conducted experiments involving human subjects • Main elements: • Volunteerism: absence of force, fraud, or deception • Knowledge: all the information needed to make a reasonable decision (not just what they request) • Competence: consenter is competent to process the information and make rational decisions

5. Engineering EthicsThe Challenger Case • Dual boosters rockets 150 Feet long and 12 feet in diameter used in first 2 minutes of launch • Dual rubber O-rings sealed four joint segments to form tank • Engineers intimately involved regarded it as an experimental undertaking at its January 1986 launch

6. Responsible Parties • Rockwell International (orbiter and main rocket) • Morton-Thiokol (booster rockets) • NASA • Marshal (propulsion) • Kennedy (launch) • Johnson (flight control) • Washington (Chief Engineer with overall responsibility for safety)

7. Pressures • Cost overruns - $14 Billion • Behind schedule • Perception by non-technical leadership that technology was not experimental • Morton-Thiokol negotiating new contract • Spec called for 40 to 90 degree temperature launce range, but freezing temperatures forecast • Decision made in matter of hours

8. Engineers Involved • Allan McDonald (M-T at Cape) concerned about temperatures and arranged teleconference • Arnold Thompson and Roger Boisjoly (M-T Seal experts) warned of temperature problems • Bob Lund (VP Engr.) and Joe Kilminster (VP boosters) agreed that there was a launch safety problem below 53 degrees • Marshall Space Flight Center was incredulous, clearly annoyed

9. Information Involved What was used What was available

10. Engineers Involved • Jerry Mason (M-T Senior VP) ask Bob Lund to take of your engineering hat and put on your management hat • Subsequent vote of management only produced finding that seals were not unsafe • Allan McDonald refused to sign recommendation to launch which occur at 11:38 at 36 degrees and took the lives of Dick Scobee, Michael Smith, Gregory Jarvis, Ronald McNair, Ellison Onizuka, Judith Resnick and Christa MacAuliffe

11. Safety Issues • No escape mechanism – too expensive • Crew was not informed of particular problems such as the field joints • Recovered booster casing had indicated that field-joints seals had been damaged, but waivers necessary to proceed with launches had become mere gestures • NASA’s unwillingness to wait out risky weather • More “Criticality 1” failure mechanisms such as field joints (no backup)

12. Lesson • Watch out for what one engineer described as the arrogance that prompts higher-level decision makers to pretend that factors other than engineering judgment should influence flight safety decisions and, more important, the arrogance that rationalizes overruling the engineering judgment of engineers close to the problem by those whose expertise is naive and superficial by comparison • Remember:Conscientiousness, Relevant Information, Moral Autonomy, Accountability

13. Morally Responsible Engineers as Social Experimenters • A primary obligation to protect the safety of human subjects and respect their right of consent • A constant awareness of experimental nature of any project, forecasting and monitoring side effects • Autonomous, personal involvement in all steps of a project • Accepting accountability for the results of a project

14. Conscientiousness • People act responsibly to the extent that they conscientiously commit themselves to live according to moral values. • Moral values transcend a consuming preoccupation with narrowly conceived self-interest • A sense of awareness is implied • A role as a social guardian but not to suggest that engineers force, paternalistically, their own views of the social good upon society

15. Relevant Information • Conscientiousness blind without factual information • Moral concern involves a commitment to obtain and properly assess all available information • Obligation to grasp the context (uses) of one’s work • Since our vision is limited and projects are experimental, ongoing monitoring is crucial

16. Moral Autonomy • Authenticity in moral conduct and principles • Kant: Moral beliefs and attitudes held on the basis of critical reflection rather than passive adoption • Commitment to action (not abstract or merely verbal) • Professional Societies such as IEEE can be a source of employee moral support

17. Accountability • Acceptance of moral responsibility for their actions • Willing to submit one’s actions to moral scrutiny • Open and responsive to assessment of others • Willing to present morally cogent reason for one’s conduct • Resistant to a narrowed sense of accountability when working under external authority that may promote fragmentation, diffusion, meeting schedules, and limited roles

18. Commitment to Safety A thing is safe if , were its risks fully known, those risks would be judged acceptable by a reasonable person in light of settled value principles.

19. Effect of Information on Risk Assessment • Imagine unusual disease expected to kill 600 • Two alternative programs to combat disease proposed • Program A: 200 people will be saved • Program B: 1/3 probability that 600 will be saved, and 2/3 probability that no people will be saved • Which program do you favor?

20. Effect of Information on Risk Assessment • Imagine unusual disease expected to kill 600 • Two alternative programs to combat disease proposed • Program C: 400 people will die • Program D: 1/3 probability that no body will die, and 2/3 probability that 600 will die. • Which program do you favor?

21. Risk Benefit Analysis • Both risks and benefits lie in the future, so we need statistical models • Expected risk might be modeled as magnitude of potential loss times the probability of its occurrence. • Near-term and future risks/gains must be valued differently • When risks and benefits have different values they cannot be compared directly, but can use ratios to compare different designs

22. Risk Related Cost in Design T = Total Cost P = Primary Cost S = Secondary Cost

23. Preemptive Design for SafetyWhich Design is Safer?