Kelly,
You might see a reason to split up the questionnaire and do the analysis twice. This would increase your participants to items ratio. (only if you have a good reason though - for instance that you asked it on two separate time slots and one of the questions was overly repetitive?) See Barki, H., & Hartwick, J. (1994). Measuring User Participation, User Involvement and User Attitude. MIS Quarterly, 18 (1), 24.
Best wishes
Caroline Wilson
Institute of Energy and Sustainable Development
De Montfort University
------------------------------
Date: Tue, 24 May 2011 13:18:17 +0100
From: Richard Morrisroe <[log in to unmask]>
Subject: Re: Factor analysis question
Kelly,
Some useful rules of thumb are that you need 300 participants, or that
you need 20 people for each item.
That being said, it really depends on the communalities (squared
correlations) between the variables, as when these increase, then you
need less people.
Personally, I would say that you have too few, but it might be worth
trying. If you intended to use the questionnaire in a regression context
(to predict some other variable) then using principal components with
orthogonal rotation would give you uncorrelated predictors, which will
simplify the interpretation of the outcomes and prevent
multicollinearity issues et al.
A very good paper on Exploratory Factor Analysis is Fabrigar et al 1999,
which discusses many of these issues in detail.
http://www.statpower.net/Content/312/Handout/Fabrigar1999.pdf
I hope this helps.
Best Wishes,
Richard.
On Tue, 2011-05-24 at 12:05 +0000, Kelly Tate wrote:
> Hi,
>
>
>
> I have produced a correlation matrix of some questionnaire data and
> notice that there exist several significant correlations amongst the
> questionnaire items. I am considering running an exploratory factor
> analysis on this data set, however I am concerned that my sample size
> is not adequate. There seem to be numerous ideas out there about what
> constitutes a large enough sample size for factor analysis, with some
> researchers suggesting that is it more about the ratio of sample size/
> variables measured. My data consists of 65 participants who each
> completed an 11 item questionnaire, does anyone have an opinion on
> whether this number is sufficient to produce a meaningful factor
> analysis?
>
>
>
> Thanks in advance
>
>
>
> Kelly Tate
>
> PhD researcher
>
> Univeristy of Manchester
>
> School of Psychological Sciences
>
>
>
> Sustainable Consumption Institute
>
> http://www.sci.manchester.ac.uk/
>
>
>
>
--
Richard Morrisroe
Irish Research Council for Humanities and Social Sciences PhD Scholar
2009-11
School of Applied Psychology,
Enterprise Centre,
North Mall,
University College Cork,
Cork,
Ireland
086 349 1678
[log in to unmask]
[log in to unmask]
------------------------------
Date: Tue, 24 May 2011 15:43:56 +0100
From: Ben Haysom-Newport <[log in to unmask]>
Subject: Re: PsyPAG 2011 - Conference Bursaries
Hi Folks
Is anyone from the Portsmouth area going to this, and would somebody act as
my carer (I am physically disabled in a wheelchair).
Thanks
Ben
On 23 May 2011 13:23, Chair PsyPAG <[log in to unmask]> wrote:
> *Apologises for any cross posting*
> **
> Dear all,
> For those of you who are planning on attending our conference in July
> (6th-8th, Bangor University) and looking for some funding, we have
> conference bursaries (up to £100) available. The deadline for applications
> for the bursaries is this Friday, 27th May. More details on the bursaries
> and a link to the application form can be found on the conference website (
> www.psypag2011.co.uk) - under the Registration tab.
> Best wishes,
> Sarah
>
> Sarah Wood
> [log in to unmask]
>
--
Many thanks
Ben Haysom-Newport BSc (Hons), MBPsS, MSc
------------------------------
Date: Tue, 24 May 2011 19:47:48 +0100
From: "Brian K. Saxby" <[log in to unmask]>
Subject: Re: Factor analysis question
Hi Kelly,
The rule of thumb I quoted in my thesis to support my PCA was:
'The recommended minimum in the literature is a 5:1 ratio and a sample
size of at least 100, although it is generally accepted that a ratio of
10:1 and a sample
of 200 is more satisfactory' (Child D. The essentials of factor
analysis. 3rd ed. London: Continuum
International Publishing Group; 2006.)
Also very helpful for PCA with Varimax rotation is: Kline P. An easy
guide to factor analysis. Cornwall: TJ Press (Padstow) Ltd;1994.
Hope that helps,
Brian
On 24/05/2011 13:05, Kelly Tate wrote:
>
> Hi,
>
> I have produced a correlation matrix of some questionnaire data and
> notice that there exist several significant correlations amongst the
> questionnaire items. I am considering running an exploratory factor
> analysis on this data set, however I am concerned that my sample size
> is not adequate. There seem to be numerous ideas out there about what
> constitutes a large enough sample size for factor analysis, with some
> researchers suggesting that is it more about the ratio of sample size/
> variables measured. My data consists of 65 participants who each
> completed an 11 item questionnaire, does anyone have an opinion on
> whether this number is sufficient to produce a meaningful factor analysis?
>
> Thanks in advance
>
> Kelly Tate
>
> PhD researcher
>
> Univeristy of Manchester
>
> School of Psychological Sciences
>
> Sustainable Consumption Institute
>
> http://www.sci.manchester.ac.uk/
>
------------------------------
Date: Tue, 24 May 2011 18:41:05 +0000
From: MICHAEL LIAM PETER RICHARDS <[log in to unmask]>
Subject: Community Psychology Survey
Hello
I’m Michael Richards, the ‘representative elect’ for the community psychology section at PsyPag.
In September, the European Congress of Community Psychology will be taking place at York St John University. We are considering whether we should have a pre-conference workshop about community psychology for postgraduates; however, we want to know what you think first. Below is a link you can access to give us your opinion on what you might like to get out of a workshop on community psychology i.e. is there a particular speaker you would like to listen too? There are only 3 questions, so it shouldn’t take you too long.
The deadline for responses is Tuesday 7th June 2011
http://www.surveymonkey.com/s/JYPVBY8
In addition, below is a link to a website that will give you a good idea about what community psychology sets out to do if you feel unsure or have never heard of it.
http://www.compsy.org.uk/
Many Thanks
Michael Richards and Liz Freeman
------------------------------
Date: Tue, 24 May 2011 11:55:47 -0700
From: Jeremy Miles <[log in to unmask]>
Subject: Re: Factor analysis question
Hi Kelly,
It's a confusing area, but I agree with Richard and Brian: you almost
certainly don't have enough cases. However, check your KMO, and
Bartlett's test. If your KMO is good, you could argue that you do.
Jeremy
On 24 May 2011 05:05, Kelly Tate <[log in to unmask]> wrote:
> Hi,
>
>
>
> I have produced a correlation matrix of some questionnaire data and notice
> that there exist several significant correlations amongst the questionnaire
> items. I am considering running an exploratory factor analysis on this data
> set, however I am concerned that my sample size is not adequate. There seem
> to be numerous ideas out there about what constitutes a large enough sample
> size for factor analysis, with some researchers suggesting that is it more
> about the ratio of sample size/ variables measured. My data consists of 65
> participants who each completed an 11 item questionnaire, does anyone have
> an opinion on whether this number is sufficient to produce a meaningful
> factor analysis?
>
>
>
> Thanks in advance
>
>
>
> Kelly Tate
>
> PhD researcher
>
> Univeristy of Manchester
>
> School of Psychological Sciences
>
>
>
> Sustainable Consumption Institute
>
> http://www.sci.manchester.ac.uk/
>
>
--
Jeremy Miles
Support Dan and Alex's school: Vote for Goethe Charter School to
receive a grant from Pepsi to help build a library:
http://www.refresheverything.com/gicslibrary
------------------------------
Date: Tue, 24 May 2011 21:33:06 +0100
From: Lelanie Smook <[log in to unmask]>
Subject: 2 day course in Trauma focused CBT in Buckinghamshire
Hi everyone
We have some places available on a top notch 2 day PTSD/Trauma CBT course presented by Dr Heather Sequiera, taking place in Buckinghamshire on 7th and 8th of June (two full days). Cost £150.00 per person. This discounted price is available because we are all in training. (Normal cost £300.00). Please email [log in to unmask] for further details. This course is offered independently by Dr Sequiera , not through any specific university arrangement.
Dr Sequiera wrote:
These are the details again for the course I ran recently at the BPS Head Office. This is a 2 day course that gives you the skills to apply Trauma focused CBT to PTSD. It is based on the Ehlers & Clarke model - considered by UK CBT therapists to be the 'Gold standard' in PTSD treatment'.
http://www.bps.org.uk/conferences-and-events/event-listing/events$/2011/march/using-cbt-to-work-with-ptsd/using-cbt-to-work-with-ptsd_home.cfm
Lelanie
------------------------------
End of PSYCH-POSTGRADS Digest - 23 May 2011 to 24 May 2011 (#2011-98)
*********************************************************************
|