A non-chance statistical result is said to be significant.

Another way of looking at this is comparing statistical significance to the idea of the ?simple majority?. This statistical significance does not mean ?significance? in the usual sense of the word. It only refers to a result not readily attributed to chance. Technically, this working definition of statistical significance is really too simplistic. But it will work for this article.

Generally, chance for two options equals fifty percent. That is, if you flip a non-trick coin several times, you should expect 50% heads and 50% tails. If the options are more than two, you?ll usually have to adjust your percentage down. For example, three options equally weighted would give a chance result of 33 1/3 % one result, 33 1/3 % the second result, and 33 1/3 % the last result.

In surveys with true/false or yes/no options, you clearly have only two options. So statistical significance for one of the items should be 51% or more. In our example of three, we should have one result at 34% or more.

An internet survey on alien abductions was offered to interested respondents, for eight years, between 1997 and 2005 by MAAR, the Malevolent Alien Abduction Research website. It can be found at http://www.maar.us/abduction_data.html.

It was a laudable attempt. I believe the authors were sincerely looking for a legitimate consensus on this topic. It also had the appearance of a scientific research survey. But it had limits.

The survey tabulators eliminated approximately 6370 or 48% of the 13270 respondents, due to several indicators including:

- Inappropriate answers
- Jokes
- Spam
- Being under 13
- Non-legitimate return email addresses
- Faked/dishonest answers, and
- Responding more than once from the same IP address

But there were other flaws.

First, the majority of questions were of the ?yes/no/not sure? type. Although this appears to indicate a three-way split for equally weighted items, that?s not the case. There?s a psychology to surveys. Most respondents want to answer yes, if that?s an option.

Next, the survey was offered at a site with a title indicating ?malevolence? on the part of the aliens. This theory of alien intent may, in the end, be revealed as irrefutable. But titled as malevolent, the survey could be expected to indicate negative experiences throughout, which it did, to a high level of statistical significance.

Then, the respondents came to the site from web searches that were based on specific keywords consistent with the topic. I presume that this internet-based survey was chosen due to a lack of financial resources to take a survey among a larger and more general population. I can really sympathize with that! But this greatly reduced the number of respondents from the general public. It is true that no survey is unbiased at its core. But more generally administered polls typically have more objective results.

So, with all that in mind, we can conclude the following from the MAAR survey:

- Respondents were those who thought they might have been abducted by malevolent aliens.
- Respondents sought out a place to voice their negative experiences.
- Negative experiences predominated the results.

But these conclusions were revealed as a sort of self-fulfilling prophecy. The results were predicted before the data was obtained, due to questions chosen for specific responses.

However, I must again aver that similar results may be revealed and become more credible with time and more data objectively obtained.

On the other hand, the Roper survey was better administered and yielded more credible results. A summary can be found at http://www.nidsci.org/news/roper_surveys.html.

(The Roper organization is found at http://www.ropercenter.uconn.edu/)

There are several reasons for this:

- The survey sample was from a more general population. But without compulsory means, this was still limited to those willing to complete a survey.
- Only eleven questions were used. Five of the questions were considered essential.
- The survey was conducted three times each. This yielded an error rate of only +1.4%.

But the Roper polls had flaws too.

A 1991 item, ?Having vivid dreams about UFO?s?, was changed in the 1998 survey to, ?Having been abducted by the occupants of a UFO.?

The control word, TRONDANT, originally thought to be an arbitrary, made-up word, was in fact a real word. The word ?Trondant? is a title for the angel Metatron, or Sefer Ha-Heshek. (By the way, the Metatron stories and their connection to possible ancient, alien abductions and contacts, is an interesting topic in itself. But that?s another story.) So however unlikely, persons indicating familiarity with the word TRONDANT, may have been telling the truth.

Even so, the Roper poll did give more reliable results.

In the 1991 poll, only two percent of the 5947 respondents, or 119, had ?experiences known to be associated with UFO abductions?.

In 1998, this figure was cut in half. One percent of the 5955, or 60, had those kinds of experiences. Although this was less than the poll seven years earlier, it contradicted the concerns that exposure to the UFO media would spoil the survey sample. This did not happen. No reason for the downward trend was revealed by the results.

This brings me to my last point: How can we all know what we?re dealing with, until we can get a more generic response from a more general population sample? Or does this already exist? And if it already exists, where is it? Why isn?t it available to all of us?

Rick Nielsen, on the web at: http://alienabductionaccounts.blogspot.com/ and http://nilthchi.blogspot.com/.

	Search our hypnosis downloads below or browse over 300 hypnosis downloads
	Search for: