Nowadays, it has become popular to base our arguments on polls, statistics or public interviews. Many news articles as well as academic publications use these statistics to either support their point or refute others. There is a saying that "numbers don't lie." But is that really true? Let's take a look.You might wonder why I chose this topic and how it relates to the office of the ombudsman, but it would not be farfetched to say that polls were born out of newspapers and journalism. We can identify the first case back in 1824. An opinion poll published in a local newspaper called The Harrisburg Pennsylvanian showed Andrew Jackson leading John Quincy Adams by 335 votes to 169 in the presidential race in the U.S. Therefore, it can be claimed that polls started as a sort of educated guess about the results of upcoming elections – whether local or national – in democratic societies. Its method evolved from sending postcards, to manning telephone lines and now we see online polls almost daily. Data mining has become easier and it is more straightforward to ask more specific questions and discover public opinion on a variety of issues. There are also many polling methods such as benchmark polls, brushfire polls and tracking polls.I would like to focus on polls other than those that predict election results, but I do have to mention their effects on the voters first. There are two primary effects of such polls – the bandwagon and the underdog effect. The former suggests that people vote for the winning candidate, while the other prefers the underdog out of feelings of sympathy. But while it appears that the bandwagon effect theory has gained support since it was first suggested in 1884 by William Safire, the underdog effect has lost its credibility over time.But while election polls are pretty straightforward, the real problem stems from other forms of public opinion. The effect of questions and how they are worded is currently a commonly voiced concern accepted by most pollsters. Some polling companies manipulate results simply by wording questions differently and thereby acquiring the answer they are looking for. If, without any prior investigation, journalists report on such polls, we become a tool for others' propaganda.In order to avoid this situation, the National Council on Public Polls (NCPP) has issued a questionnaire containing 20 questions every journalist should ask before publishing a poll in their newspapers. Here are a few of them:"Who conducted the poll?Who paid for the poll and why?How many people were interviewed for the survey?When was the poll carried out?How were those people chosen?What questions were asked?In what order were the questions asked?"By asking such questions, we can learn about the pollsters, their aims, the timing of the poll, the demographics of the people who participated and the questions themselves. This also coincides with basic journalistic questions such as when, where, why and who. If we are satisfied with the answers, we can publish our piece, but we must make sure that we inform our readers of the questions and the number of people who were interviewed in order to achieve full disclosure. This will provide credibility both for us and the firm that conducted the poll.In addition, there is the case of Internet polls. The dangers of conducting a poll online is that it can be very easy to manipulate, even by people other than the pollsters. Multiple voting and altering results through technological means are some of the issues involved. You may remember a scandal that erupted on this subject last year as part of the Edward Snowden leaks. In the leaked documents, it became apparent that a U.K. intelligence agency, the Government Communications Headquarters (GCHQ), manipulated online polls via programs called Gateway and Underpass. So it would be wise to approach them with a modicum of skepticism, whether publishing an online poll or conducting one via the website of the newspaper. Nevertheless, the science of polling remains an efficient method with which to discover public opinion, but journalists need to be able to weed out the "bad" ones since they are of little use.