we  try  to  measure  the  most  fluid  paradigm  in  the world–human  thought. ANALYSIS OF SURVEYS Learning  Objective: Analyze  and  use  the  data generated by audience surveys. The simplest and easiest way to communicate a complicated list of data to an audience is with a picture. As journalists and as a result of your studying Chapter 6 of this training manual, you already know about the various  types  of  visual  communication  possible  for displaying data. The final section in this chapter covers understanding  the  survey  data  you  have  collected  and how  to  correctly  plug  that  data  into  visual  repre- sentations, such as graphs and charts. CORRELATIONS As  mentioned  earlier  in  this  chapter,  making decisions, based on samples of under 400 from general populations of 4,000 or less, is possible and reliable if you   are   making   simple   comparisons.   A   simple comparison is what percentage of your sample watches television. Your results might look like the following: 34  percent–0600-0800,  10  percent–0801-1000,  and  so on. But if you wanted to know what percentage of those who watch from 0600-0800 are women or children, then remember, you must have a large enough sample to keep each category you want your survey broken into over 100. This will keep the relationship between viewers watching from 0600-0800 and the sex of those viewers reliable. Correlating  viewership  with  demographics  is important,  for  example,  when  you  plan  to  run information  spots  targeting  women.  Questions,  such  as, when do most spouses watch your television station, listen to your radio station or what editions of your paper are  they  more  likely  to  read,  are  important  to  answer when you plan an information campaign.  Know Your Audience. As Rear Adm. Baker, the former CHINFO once  said:  “Work  smarter,  not  harder.” GRAPHIC DISPLAY The  last  section  of  this  chapter  ties  back  with  the scenario used in the introduction. Putting your data into a picture to give it impact is absolutely necessary and also dangerous. It is necessary because few people will take the time to study raw print-out sheets of data, no matter what the claims of the surveyor are. It is also dangerous because a mistake in how you present your survey  findings  can  ruin  the  credibility  of  all  your efforts. How  many  times  have  you  heard  the  saying, “Statistics lie”? You must display your survey findings in the most understandable way possible and still stay within the boundaries of fair representation. Perhaps the best way to understand how to present your data fairly is to see some examples of how to be unfair. These examples will also increase your ability to spot abuse of statistical  data. Let us say that as the command’s expert on survey techniques,  you  volunteered  to  supervise  Morale, Welfare, and Recreation’s facility usage survey. You have  followed  the  guidelines  in  this  chapter  and  have collected reliable and valid data and are ready to present the results to the CO. The CO’s pet project is the base racquetball courts, and he is interested to see how many sailors are using it since it was built last year. You are aware of his interest and  decide  to  graph  the  informal  aggregate  data collected  by  reviewing  the  sign-up  sheets  in  the racquetball court front office. You hope to show a steady increase in usage since your broadcasting outlet has been regularly producing information and selling spots on the new courts. Starting with January, you list each month on the bottom of your line graph. On the left, you place numbers of users in blocks of 20 and begin to place your data on the graph. An example of this is shown in figure 9-5. The   line   chart   does   indeed   show   a   steady progression of increased usage up the graph. However, because of the scale you used, the data does not really jump  out  at  you.  So  you  decide  to  display  your information in a more effective manner. To be more pleasing to the eye, you cut off the upper and lower sections of the graph to end up with the display shown in figure 9-6. The graph looks better to you now and it even shows the slight dip in the summer months when you were forced to go TAD causing, in your estimation, radio spot production to fall off. If only the data could illuminate that  fact  a  bit  more  you  might  even  get  a  letter  of appreciation  from  the  CO  for  doing  such  a  good  job  of promoting  his  pet  project.  To  highlight  the  summer  dip, you decide to go with the graph in figure 9-7. That  does  it.  Who  could  deny  that  the  base racquetball court, the CO’s pet enterprise and your personal advertising project, is not a big success? Was the data changed in the various graphs? Was there a lie in any of the charts? The answer is yes and no. There is 9-10