By Leon J. GleserUniversity of Pittsburgh, Executive Editor of Statistical Science, 1997-2000

The interviews in Statistical Science are one of its most popular features. The personal opinions and experiences given in these interviews add a human touch, while also giving insights into the intellectual history of our field.

It is always tempting to conduct an interview as a kind of oral autobiography, leading the interviewee through the stages of his or her career. Although autobiographical details that cannot be found in a person’s vita are often interesting, the real contribution that an interview can make is to give insights into the development of concepts and into the interplay of personalities and circumstances that led to such concepts. Also of great interest is how, and why, notable departments and research groups were formed. What were the needs that such departments and groups were intended to meet? What goals did the founders of such groups have? What were the obstacles that had to be overcome? How completely did these groups meet their goals? Past interviews have provided us with the memberships of many of the important departments and research groups in the United States during World War II and the 1950’s and 1960’s, but less insight into how these individuals interacted.

Finally, individuals who have been present during the development of the field of statistics are perhaps best able to provide an overview of how far we have come, and a perspective of where our field should be heading.

To summarize, we feel that the type of interview most appropriate for Statistical Science is one that focuses on ideas and their development, rather than merely upon the details and dates of an individual’s career.

The Interview

To build an interview around ideas, while still keeping some autobiographical and historical interest, requires the interviewer to obtain a good understanding of the accomplishments and interests of his or her subject. For this purpose, it might be helpful to secure in advance a copy of the interviewee’s vita. It is not unethical to explore in advance with the individual the topics that will be covered. The interviewer then should prepare a plan of how he or she would like the interview to proceed, and construct a few possible leading questions. The following are some examples used by Morris De Groot in past interviews for Statistical Science:

Q How did you initially get interested in the subject of statistics?

Q Who have been the major professional influences on your career?

Q You have published more than x papers. Which ones do you regard as your particular favorites?

Q Tell me about your relationship with x (another famous statistician or scientist).

Q What does the future hold for you? Do you have any major books or projects under way?

Q What do you like to do when you are not doing statistics?

Q What is your assessment of the current state of health of the field of statistics?

Q Where do you see the field heading?

Q Are there other topics that we should be talking about, or that you would like to talk about?

Don’t expect an immediate response to all questions. Sometimes a long pause will elicit an answer – people feel compelled to contribute when there is silence in a conversation, and an individual often needs time to muster a response. Keep in mind that you will be able to edit the interview. Allow digressions if they seem to have potential interest, and don’t be concerned if the conversation temporarily seem to flounder. Introduce questions as needed to stop digressions or move the discussion away from areas where the interviewee seems unable or unwilling to respond. Try to find topics for which the individual has strongly felt and interesting opinions. (But it would be wise to discourage personal vendettas or partisan expressions concerning relatively ephemeral political matters.)

For some individuals, their chief contributions may not be directly connected with research, but rather with the securing of resources and the establishment of groups that produced important research. For example, the individual may have established an important department of statistics, industrial research team, or government bureau. In this case, the individual might be asked what prompted them to form this group, what goals they originally had for the group, what problems and opposition they had to overcome, and so on.

For the interview, you will want to have a good tape recorder with a microphone of sufficient quality to pick up voices clearly at a distance of several feet. If the recorder is battery operated, use a fresh battery and keep at least one other battery in reserve. Also, have sufficient tape available for at least two hours. If a trained legal stenographer is available, it might be worth having the interview transcribed as well as taped, particularly to make note of names and locations that might not be clear on the recording, and to identify who is speaking at any particular time. The stenographer has a special machine that permits them to take down what people say with rapidity. Consequently, a transcript is available almost immediately, which you can then edit with the help of the tape recording. Ingram Olkin has conducted interviews with and without legal stenographers, and prefers to use stenographers when possible. However, this is not always easy to arrange.

Try to hold the interview in a place where there is a minimum of background noise, and where both you and the person interviewed will feel relaxed and at ease. Check the tape recorder and microphone for proper performance before beginning. (Nothing is more frustrating than to have completed a successful interview and find nothing on the tape.) When you know when the interview will be held, or soon after the interview was held, please notify the Executive Editor of Statistical Science. At that time, it would be helpful to both you and to the journal to suggest an approximate time at which a first draft will be ready for submission.

Photographs and Other Materials

If you can arrange this, try to secure from the person interviewed some photographs taken at various times during their career. (The person’s spouse or close colleagues are other possible sources for such photographs.) A good clear recent photograph of the individual is particularly important. It should be relatively easy to obtain photographs of the person and their family at various times in the person’s career. Of greater interest, however, would be photographs showing the individual lecturing or meeting with notable colleagues. Try to identify individuals appearing in such photographs, as well as the time and place where the photo was taken. Photographs add greatly to the interest and appeal of an interview.

You should also obtain a bibliography of at least the articles and books mentioned by the interviewee in the interview. A list of their main Ph.D. students and collaborators may also be helpful to you later.

It is a good idea to secure as many of the promised photographs and the bibliography either before the interview, or soon afterwards. It is best to get as much of this accomplished as you can around the time of the interview, because afterwards the subject may forget or mislay the needed photographs or information.

Editing

As soon as possible after an interview, sit down with the tape of the interview and make some notes for editing. Spotting unclear parts in the tapes at this time may permit you to clarify these with the interviewee while he or she is still available. Have the tape transcribed into text by a secretary.

At this point, you are now ready to edit the interview. Most of your editing will involve getting the points made in the interview into as clear a prose as possible, while trying as much as possible to convey the characteristic speaking habits and personality of the person interviewed. Do not hesitate to delete portions of the conversation that are uninformative or distracting. In general, remember that a verbatim copy of the interview is not expected; thus, you have the freedom to delete uninteresting sections or rearrange topics, as long as this does not distort your subject’s statements into something that he or she did not intend to say. To the extent that this is possible, try to make the edited interview flow logically from one topic to the other, while keeping the liveliness and spontaneity of the original conversation. A useful tool in this process is to arrange the topics covered in the interview into coherent blocks; you can then edit your own questions and comments so as to connect these blocks into a logical and well-motivated whole.

Note: It is not a good idea to show the subject of your interview the raw transcription of the interview or your early attempts at editing. They are likely to be disturbed by the grammatical errors and incoherency typically present at this stage, and may want to take the whole interview and rewrite it. The interview is then out of your control, and sometimes interviewers who make this mistake don’t ever get the interview back. Once you have a reasonably coherent draft, you can show a copy to the interviewee in order to identify words or phrases that were unclear, or get clarification of what the person intended to say. Although the interviewee is the central focus of the interview (“Conversation”), you are the one writing the interview.

Once the interview has been completed, transcribed and edited to your satisfaction, send the first draft of the edited interview to the Executive Editor (Sending the first draft by e-mail is acceptable, and encouraged, with pdf files being preferred.) Your draft will be reviewed by at least one editor of Statistical Science, and suggestions made for revision (if needed). At this time, you might want to also send the photographs that will accompany the interview. (You can assure the donors that their photos will be returned safely.)

Don’t forget that you need to write a brief biographical introduction (called the “abstract”) to begin the interview. Try not to make this overly redundant with the interview. You can get ideas for the form of this introduction from other Statistical Science interviews. Looking at these interviews will also make you aware of the style conventions (e.g. the form of references and headings) required by the journal. One convention is that we try to avoid mentioning the word “interview” either in the title (it is always “A Conversation With…”) or in the text. The spirit of this convention is that we are being allowed to overhear a conversation between the interviewer(s) and the interviewee.

Conclusion

This guide is intended to give you an idea of what is needed, and some advice. It is not intended to intimidate you or make you anxious about the interview. There is really no reason to worry. As long as you do a little planning in advance, everything will flow smoothly. Further, you will find it enjoyable to have a more human change of pace from the academic type of writing that you customarily do. Besides the gratitude of the editors of Statistical Science, you will also have the satisfaction of contributing to intellectual history and to the enjoyment of your colleagues.