Hi Peter,
Don't know if of much use, read you needed it by Feb 8th (sorry L). As you're aiming at surveying 1/3 of your population results might be statistically significant for such universe unless there's remarkable diversity within it; though the key issue lies with the goals of your research, why you're conducting it and what you expect to achieve. . If you want to ensure a representative outcome sampling or addressing the 30 people in your universe would be advisable, particularly for further research beyond that population. If this includes well-defined groups (e.g. according to job roles) you might define your sample according to these rather than applying full traditional techniques but this depends on your design. . Alternatively, you could use a second "simplified" survey/questionnaire among the remainder of the population to contrast your original results. In this case, you could reduce the number of "full interviewees", again linked to design. . Or you may realise statistical significance is not that prime for your study; or that a qual/mixed methods approach fulfils your objectives better. Info is scarce so can't be more specific. If you need further help, please drop me a line. Cheers, Alex P. Real De: [email protected] [mailto:[email protected]] En nombre de Wayne Mackintosh Enviado el: lunes, 08 de febrero de 2010 5:36 Para: [email protected] Asunto: Re: [WikiEducator] Small Sample Size - Non-parametric tests? Hi Peter, That's a good question. mmm ... I'm stepping out of my zone of comfort here, I'm not a statistician or applied mathematician -- so you'll need to check my response with a professional. Your population is small --- so not sure whether sampling techniques are relevant or appropriate in this instance. With a small population like this -- I'm not sure how one one would figure out how your population fits any parametrized distributions. So my uninformed opinion is to go with non-parametric statistics. I could be totally wrong here ;-) Let me ask one or two WikiEducators with better statistical knowledge and point them to your question. Cheers Wayne On 7 February 2010 21:31, Peter McCrory <[email protected]> wrote: Can anyone help? I am looking to conduct some research into the effects of 5 mental skills on self-regulation, and I will be selecting a sample size of approx 10-12 people from a total population of approximately 30 people. Am I required to use non-parametric statistics? Or is the sample large enough to use parametric tests? Many thanks Peter McCrory T. +44 1522 514422 M. +44 7812 080853 E. [email protected] Skype. peter.mccrory Twitter: @Peter_RAFUltra www.wikieducator.org/User:Wikipeedy Follow our progress in preparing for the World's Toughest Bike Race at www.raf-ultra.co.uk -- You received this message because you are subscribed to the Google Groups "WikiEducator" group. To visit wikieducator: http://www.wikieducator.org To visit the discussion forum: http://groups.google.com/group/wikieducator To post to this group, send email to [email protected] To unsubscribe from this group, send email to [email protected] -- Wayne Mackintosh, Ph.D. Director, International Centre for Open Education, Otago Polytechnic, New Zealand. Board of Directors, OER Foundation. Founder and Community Council Member, Wikieducator, www.wikieducator.org Mobile +64 21 2436 380 User Page: http://wikieducator.org/User:Mackiwg Skype: WGMNZ1 Twitter: OERFoundation, Mackiwg -- You received this message because you are subscribed to the Google Groups "WikiEducator" group. To visit wikieducator: http://www.wikieducator.org To visit the discussion forum: http://groups.google.com/group/wikieducator To post to this group, send email to [email protected] To unsubscribe from this group, send email to [email protected] -- You received this message because you are subscribed to the Google Groups "WikiEducator" group. To visit wikieducator: http://www.wikieducator.org To visit the discussion forum: http://groups.google.com/group/wikieducator To post to this group, send email to [email protected] To unsubscribe from this group, send email to [email protected]
