compartment of a ship or in a barracks where one line is  often  the  only  telephone  connection  to  scores  of people. Another  important  factor  is  that  many  people  will not  feel  free  to  make  candid  comments  within  earshot of  their  shipmates. Personal Interview Surveys There   are   many   advantages   of   face-to-face interviewing, and it is still considered the most accurate form  of  gathering  data  for  a  survey.  However, interviewers must be properly trained so as not to let their  personal  habits  or  motives  influence  the  sample respondent.  The  interviewer  is  supposed  to  be  like  the ultimate statesman or the ideal father or mother figure. Interviewers  must  be  understanding,  patient,  modest, well dressed (but not too well dressed), soft spoken and above ail, good listeners. Do you know anybody like that? The point here is that whether doing a face-to-face interview or talking on the telephone, keeping the survey data free of any interviewer effect is important and difficult. For instance, if you are a chief petty officer doing an interview with a seaman and the question deals with the command’s response to recent incidents on base vandalism,  the  outcome  might  be  affected  by  the disparity  of  rank.  For  the  most  part,  face-to-face interviewing will seldom be used because of the time needed to do it, the inherent difficulties of the technique and  the  labor  cost  of  moving  an  interviewer  from respondent  to  respondent. Mail  Surveys The  most  widely  used  and  traditional  DoD  survey is the mail survey. It has two very serious disadvantages though. The first and foremost disadvantage is that mail surveys most often only have a 10-to-15 percent rate of success in terms of surveys being returned. In most academic  and  professional  polling  circles,  this  is  a useless effort. To base a decision on a mail survey with a response rate of only 15 percent is asking for a decision based on nothing better than hearsay. It is not “better than nothing” as is often heard in defense of such efforts. It is worse than nothing! The reason it is worse than nothing is because the hearsay data now has an air of fact about it that it does not deserve. It would be better not  to  conduct  a  survey  at  all,  and  base  your programming   or   public   affairs   efforts   on   your observations  of  the  informal  survey  methods  mentioned above. The second big disadvantage is that a mail survey can be filled out by anyone at the address to which it is mailed. This means if you are trying to find out what the working  spouses  in  your  audience  would  like  the  base to do about day care, you might get a large percentage of married military members simply answering for their spouses in the way they think the spouses would have, or  should  have  answered.  This  will  produce  biased results,  and  again,  decisions  may  be  made  that  might exacerbate the problem. Now that we have put mail surveys in their place, let us look at why we use them. One reason is because they  are  cheaper  than  personal  interviews  of  each sample respondent. Also, the low response rate can successfully be attacked. If you can get the response rate into  the  80-plus  percentile  range,  you  have  a  good audience  representative  survey  on  your  hands. The following is a list of proven ways to get into that response  range: 1.  Keep  the  questionnaires  short.  Ideally,  this should mean one page with 20 or so questions. This is not a lot of questions, but remember, you are trying to get  accurate  information,  not  a  bunch  of  responses  from only those people who have the time to fill out a lengthy questionnaire. Also, keep in mind that much of your demographic   information   on   military   and   DoD personnel can be gathered from existing personnel tiles. Gathering  data  from  these  files  also  eliminates  a respondent’s  fear  of  being  identified.  With  shorter questionnaires, the dreaded survey project will be easier to complete. This means you will probably do more of them as you become familiar with the methods. Shorter, more  frequent  surveying  of  your  audience  gives  you  a better picture of  what is happening and that is what it is all about. 2. Put a stamped, self-addressed envelope inside the survey mailer. (One creative journalist let the survey questionnaire itself serve as the return envelope. All the respondent  had  to  do  was  fold  the  questionnaire  over, staple it shut and put it in the base guard mail system.) 3. After you make your random selections (this will be explained later), stick to them and get their response.   Three   or   four   follow-up   mailings   to nonrespondents are often needed to get an 80- or even a 90-percent return ratio. (See fig. 9-1.) 4. Do not continue to mail surveys to the general population until you get a magical number of responses. This  will  produce  a  survey  that  might  have  a disproportionate  segment  of  your  audience  represented, and is considered to be a biased survey. An example of 9-4