without the visual details with which we human beings ori-
ent ourselves?
E-mail can work well for follow-up, or if all you need is a
brief expert reaction to someone else’s research, but the
e-mail “interview” tends to devolve into mere information:
dry little packets of fact. I find that a piece written that way
acquires the hollow, depersonalized sound of an encyclope-
dia entry, and I would abandon a story idea rather than de-
pend on e-mail interviews.
Telephone interviews may be a feasible alternative. Person-
ality transmits on the phone, and you and the scientist can
develop significant rapport and a coherent line of thought.
Interview in the morning, on the other person’s turf, and
never in a restaurant. Timing is no deal-breaker, but if you
can, avoid the midafternoon droop: make your appointment
for the morning.You’ll get a better interview when both of
you are bright-eyed and bushy-tailed, as my mother used to
say, and when the day’s distracting little crises have not yet
arisen.
Never interview in a restaurant. Silverware clicking on
plates and the conversation at the other tables will obliterate
every sound on the tape, nor are you well situated to take
notes. Distractions abound.
At the person’s lab or office, by contrast, things are quiet,
the scientist feels comfortable, and serendipity can happen. If
the subject widens, you and the scientist can scamper down
the hall to look in the lab or talk to someone else. If there’s a
reprint or photograph the person wants to give you, you’re
in the right place. (What you go home with that day, you
can be sure you have.) Because you two can see each other’s
faces, you will avoid many small misunderstandings, and

You want to leave room for the new and exciting. “I’m sure
you know anything I’ll ask right off the top of your head”
can be a good way to say it, because it is so patently true.
A brief description of your purpose should be enough.
For example:
“To talk about your work on protein folding for a news item
in the Weekly Blather.”
“To talk about the implications of your work on protein fold-
ing, possibly for a feature article in the New York Times.”
“I understand you are a gatekeeper, one of those people who
always knows what’s going on. I’d like an hour at your con-
venience to hear about whatever is exciting people in the
field right now.”
Even though you did not send questions, you will some-
times find, when you get to the appointment, that the per-
son has prepared a speech anyway. In that case you should
listen. Sometimes people must fulfill their own agenda be-
fore they can pay attention to yours.
If their agenda is incompatible with yours, you will need
to be gentle but forthright: “Not of interest to our readers”
is the time-honored way to phrase it. Or you can blame the
editor: “I will discuss your idea with my editor [my teacher],
but I know that what intrigued her was the protein folding.
Perhaps we could go on to that now?”
Research
and the
Interview
51
And—sometimes there’s no extraneous agenda. Sometimes
the scientist knows better than you what you should be ask-

comes out of your mouth will be a direct response to some-
thing the scientist said.You’ll be tackling the subject in an
order governed by his train of thought and in language that
reflects his—in short, your questions and comments will be
better than what you wrote down.
Nothing good will happen, however, if you do not prepare
coherent questions and write them down in some sensible
order. Once again, the wisdom of Louis Pasteur applies:
“Luck favors the prepared mind only.”
Ideas
into
Words
52
In preparing your questions, stay simple and straightfor-
ward, like Bill Moyers.Your purpose is to elicit the lively
explanations you need for the piece, not to impress the sci-
entist or to fill in gaps in your own education. Doing your
work well is the best way to be impressive.
Many inexperienced writers are afraid the scientist will
think they are dolts, so they work up long, elaborate ques-
tions, the sort of scenario-setting stuff that looks well in-
formed on paper. Don’t do it. The questions you see in
printed interviews were usually written for structural rea-
sons, to make the interview come out sensibly after the tran-
script was cut and rearranged. No one actually said anything
so long-winded, as you’ll know if you try to speak one of
those three-thought mindbenders.
If you are afraid the scientist will think you are a dolt, you
can always say, “Our readers will want to know etc.”
For starters, make sure your questions cover the news-

are also your future sources, come the day when both you
and they are distinguished in your own right, and they have
much to say already. Why not meet them now?
With regard to What, make sure you know where the back-
ground explanation stops and this particular research begins.
You’d be surprised how easily that line can blur in your mind.
Why and When? Why this line of research and not some
other? Why now and not before? Guaranteed, there must have
been other ways to approach the issue, so what was the advan-
tage of this one? The answers are always part of a larger pic-
ture, about either the science or the research strategy or both.
How, precisely, was the research performed? You want to
know at about this level: “We do this because A, then that to
create B. Then we put the B in the glomerator and wait.
What we hope will have happened after 24 hours is . ” Or,
“We assembled a control group of 230 people. The group
had to be that big because X. It was important that they all Y
because Z.” The root of the matter is implicit in these mun-
dane details, so knowing them will clarify your thoughts.
Occasionally, you will even want to hang the entire piece on
a narrative structure: “The team wondered X, so they did Y.
Unexpectedly, results were M. So they started again, doing
H.” And so on.
“Why should the reader care?” can sound hostile and you won’t
want to phrase it that way. I usually say, “What are the long-
term implications of this work?” A friend of mine likes to
ask, “If you had to state the significance of this work in
twenty-five to thirty words, what would you say?” He and the
scientist sit there and work on the statement till they get it just
right, a process they both seem to find fun and illuminating.

Give particular thought to your first question, which has
several jobs: it should start the conversation off in the right
direction at the right level, and it should be a big fat juicy
one, right over the middle of the plate—something the sci-
entist can hit out of the ballpark.You want her to feel satis-
fied with both herself and you. (“Oh yes, she’s okay and I can
handle this. I’m hot today!”)
The opening question should not be personal. While Bar-
bara Walters asks people things like “What kind of tree
would you be if you were a tree?” the rest of us find it works
better to go straight to the science.
Simple can be best, especially if time is short. For example:
“I have studied the material you sent me, and it seems to me
the heart of the matter is [
FILL IN THE BLANK]. Is that correct?
How would you phrase it? . . . How did you get the idea?”
And you’re off. Don’t try that one unless you are dead sure
you understood the material, however.Your purpose is to si-
multaneously reassure, indicate the level of discourse, and
cut to the chase, not to expose yourself as unprepared.
Curiously, it’s okay to be semi-informed so long as you
show you’re aware of it. “I have studied the material you
sent me and I got stuck on X, as I think the readers might. I
don’t understand [
WHATEVER YOU DON’T UNDERSTAND, IN A CRISP
TWO SENTENCES
].” This formulation also cuts to the chase and
indicates the level of discourse. It’s efficient. It is even some-
what reassuring, in that you show yourself to be a person who
thinks with enough precision to know what you do not know.

be able to adapt.
Did you have a Eureka! moment in doing this research?
What new possibilities do you now see?
Do others in the field generally accept your data? Do they
generally agree on your interpretation? What is the cen-
tral issue?
What was the biggest surprise you got in this work?
How did you first get the idea?
What comes next? Is the next experiment already clear
to you?
What is the next important question?
Do you have any intuition about what the answer might be?
Ultimately, where do you expect this line of research to go?
“Do you have any preliminary results?” Once you’ve
launched into implications and future work, look for some-
thing concrete. Scientists almost always know more than
Ideas
into
Words
56
they can prove, because to get the next grant, they must offer
evidence that what they propose is worth doing. In effect,
then, much research funded by grants actually nails down
results from preliminary work that the team had already
done, late at night or sketchily, on bootlegged funds.
A full experiment will not necessarily confirm preliminary
results. Even if it does, the most prestigious academic jour-
nals often refuse to print what has appeared elsewhere—yes,
even in a local newspaper. From the scientist’s point of view,
then, prepublication blabbing carries a heavy price.

We do not learn by abstract sets, however.We learn by start-
Research
and the
Interview
57
ing from something concrete that we know or can imagine, to
which we attach the new idea.Your reader is learning from
your article, so you must start with the familiar.You might
start explaining surface tension, for example, in terms of the
familiar water spider, which appears to skate upon the sur-
face. How can it do that? Well thereon hangs the story.
The scientist you are interviewing is long past needing any
such crutch, of course. He can visualize the abstractions with
no problem, so may not think to give an example or analogy.
In that case, you must ask for them as needed.
If an analogy occurs to you as you listen, you might offer
it and ask if it works. In general, however, the best analogies
will come from the scientist because they derive from a
grandly simple view of the field. The scientist may have such
a view, but you almost certainly do not.
“Compared to what?” Hold this question at the back of
your mind, so it will automatically pop out as required. Let’s
say such-and-so operation produces a 70 percent survival
rate after five years. Great! How does that compare with
other treatment strategies? With similar operations for other
cancers? Is the difference statistically significant?
“Tell me more” is a magical, all-purpose question because
it is open-ended, giving the speaker permission to take the
idea wherever he or she wants. It’s a good way to go deeper
when you can tell something is important but don’t yet

basic they’re easy to forget—make sure you have eaten and
that your bladder is empty. Few things are so distracting as
hunger pains, which is why you might keep a nutrition bar
handy. It can save your bacon.
Dress appropriately, meaning professionally but not what
used to be called “power dressing.”Your goal is to make the
other person feel relaxed and comfortable, not intimidated
or even much impressed. Dazzle ’em with your mind and
your writing, not your fashion sense, especially if you are a
woman. Keep the transaction professional.
In an academic or computer-lab setting, professional attire
may be very informal indeed—for the natives. As a visitor,
your style of informality needs to stay respectful. For men, a
tie and sports jacket (even if both are later taken off) will not
hurt. For women, the specifics of “appropriate” vary wildly
over time: do the right thing.
Having taken care of all important trivia, you can now
begin the interview with no need to think about yourself.
Rather, your undivided attention will be available for the
other person and the fascinating things you are about to
hear. As you walk through the door, mentally give your
troubles a kiss and skootch them over. They will wait.
Once in the room, there’s no rush to start. Allow a few
moments for the two of you to get used to being in the
same room, the process that I call “dog-sniffing.” Dogs need
to sniff each other, and people need to ask whether you had
trouble finding the building, is it still snowing, where
should we put the tape recorder, do you want any coffee?
It’s all dog-sniffing, two to three minutes of adjustment time.
(I have sometimes wondered whether in literal fact the need