From An Archive Approach to Oral History (1978, at. pp. 15 - 16) by David Lance, at the time Keeper of the Department of Sound Records at the Imperial War Museum, some "sins" which interviewers should always avoid:
1. Questions which are unnecessarily too long;
2. Questions which are not clear;
3. Questions, too frequently, which are answerable by "yes" or "no";
4. Combining several questions into one;
5. Interrupting a speaker with a secondary question before he has finished answering the first;
6. Failing to follow-up on a question which has not been fully answered;
7. Seeking, too often, for opinions and attitudes (particularly without establishing any factual basis for them);
8. Missing opportunities for follow-up questions which are "invited" by earlier answers;
9. Not asking for specific examples to illustrate general points which an informant has made; and
10. Jumping to and fro between one subject and another, or one time period and another."
Lance also offered these useful pieces of advice:
"Generally, the degree of useful information in a recording is in direct proportion to the amount of interview preparation that has been carried out."
"Interviews most conveniently follow a chronological pattern; start at the beginning and work systematically through the period which the particular project is concerned with."
"Do not hurry the interview process. The pace of an interview depends mainly on the informant's personal capacity; the length depends on the amount of useful information he has to give. There should be no other personal factors to consider in deciding how much time to devote to each informant."
"The purpose of oral history interviewing and recording is to collect interesting and significant information by questioning men and women about their personal experiences within prescribed subject areas. Interviews should be based mainly on activities or events in which informants were directly involved."
And finally, my favourite, which all SETI types and debunkers who think witness testimony is worthless should be forced to read; here, Lance quoted from Sir Basil Liddell-Hart, one of the great historians of the 20th century:
"History should be tested by the personal witness of those who took part in the [events]... The more that any writer of history has himself been... in contact with the makers, the more does he come to see that a history based solely on formal documents is essentially superficial."
Absolutely, one hundred per cent, true. However, in order to accomplish this goal, it is critical to remember that the researcher has to get it right in the field.