During the last 3 years presented FS2004 to a group of +70 students with ages between 20 and 25. They had some interested in aviation but specially in the telecommunications systems, they were in the last bachelor year in engineering.Only one asked questions after the presentation and in fact started using FS. Of course this approach was not with the objective of studying the use of flight simulators but to show examples of telecommunications systems interaction, it can indicates a very low retention/interesting in simulator use.

Edited March 4, 201214 yr by jcboliveira

Sorry, but I cannot resist.

...a web browser executing a copy of the actual avionics software.

That, ladies and gentlemen, would be "as real as it gets"!Aaawwwesome.

I find it particularly difficult to believe that such a statistically attuned culture as Microsoft would not have done their maximum amount of objective statistical due diligence on both the FLIGHT delivery and interface before release. If the sample size is adequate to establish a satisfactorily accurate probability, and if there was not some kind of self selectability at work within an otherwise random exercise, then the findings may be significant. But I can not see where such a small sample size by a circle of friends can be any form of validation of any indications, let alone such stated conclusions.Accurate statistical studies are very complex. For example, in order to establish reasonable accuracy with political polling and opinion surveys each separate question has to be individually tested to validate its' constancy when presented in the greater context of multiple queries. Even then, the error factor is significant when the random sample size is less than 2,000 + or -.I would like to see the project notes to evaluate the likelihood of there being any form of reliable probability factor here. I suspect, all good intentions aside, it remains invalidated or at least unverifiable.Kind regards,

I find it particularly difficult to believe that such a statistically attuned culture as Microsoft would not have done their maximum amount of objective statistical due diligence on both the FLIGHT delivery and interface before release. If the sample size is adequate to establish a satisfactorily accurate probability, and if there was not some kind of self selectability at work within an otherwise random exercise, then the findings may be significant. But I can not see where such a small sample size by a circle of friends can be any form of validation of any indications, let alone such stated conclusions.Accurate statistical studies are very complex. For example, in order to establish reasonable accuracy with political polling and opinion surveys each separate question has to be individually tested to validate its' constancy when presented in the greater context of multiple queries. Even then, the error factor is significant when the random sample size is less than 2,000 + or -.I would like to see the project notes to evaluate the likelihood of there being any form of reliable probability factor here. I suspect, all good intentions aside, it remains invalidated or at least unverifiable.Kind regards,

:Thinking: I would have assumed a minimum of 500 people.

Edited March 5, 201214 yr by HiFlyer'

I find it particularly difficult to believe that such a statistically attuned culture as Microsoft would not have done their maximum amount of objective statistical due diligence on both the FLIGHT delivery and interface before release. If the sample size is adequate to establish a satisfactorily accurate probability, and if there was not some kind of self selectability at work within an otherwise random exercise, then the findings may be significant. But I can not see where such a small sample size by a circle of friends can be any form of validation of any indications, let alone such stated conclusions.Accurate statistical studies are very complex. For example, in order to establish reasonable accuracy with political polling and opinion surveys each separate question has to be individually tested to validate its' constancy when presented in the greater context of multiple queries. Even then, the error factor is significant when the random sample size is less than 2,000 + or -.I would like to see the project notes to evaluate the likelihood of there being any form of reliable probability factor here. I suspect, all good intentions aside, it remains invalidated or at least unverifiable.Kind regards,

:Thinking: I would have assumed a minimum of 500 people.

If my statistics studies haven't gotten too rusty... any sample of size of greater than 30 is considered statistically significant. That is assuming the sample is a SRS from the population studied.

If my statistics studies haven't gotten too rusty... any sample of size of greater than 30 is considered statistically significant. That is assuming the sample is a SRS from the population studied.

But was it? A sample group gathered by previous participants?If I was asked as a beta tester to find some friends for further beta tests, I don't think my selection would be random. I would be selecting for people I thought would be amenable......Of course, as beta testers, we also self selected as interested parties, but now MS is backing up their results from us with usage data from actual customers.....And that's above and beyond focus groups, consultants and God knows what else they brought in before they even green-lighted this project.....

Edited March 5, 201214 yr by HiFlyer'

But was it? A sample group gathered by previous participants?If I was asked as a beta tester to find some friends for further beta tests, I don't think my selection would be random. I would be selecting for people I thought would be amenable......Of course, as beta testers, we also self selected as interested parties, but now MS is backing up their results with usage data from a much larger sample.......

A true SRS isn't exactly easy to come by so I would say no, it was not an SRS.