[Rasch] FW: Item order
janescott at iprimus.com.au
janescott at iprimus.com.au
Fri Mar 9 11:09:52 EST 2007
Thank you everyone for your advice.
Just a bit more information that might help you understand what I am doing.
All items are on a 5 point Likert scale. Items to do with feelings and arousal
have a scale which measures intensity. So instead of SD-SA, responses are
recorded as being "never", "a couple of times", "a few times", "often" or
"throughout the movie". For other items to do with cognitive effort and pleasure/liking,
the scale is a more traditional 5 point likert "strongly disagree to strongly
I am not sure if this will affect feedback that has been given to me, but
many people have assumed that my items were dichotomous "yes/no" or that
people picked just one of those feelings - eg. good, happy, elated - to agree
with. Rather, I am interested in how people agree with, or how intensely
they feel, each item.
Finally, I am quite open to splitting the sample and having different versions
of the questionnaire. Especially since my next stage is a pilot, so I do
have some flexibility. What sample sizes do you think I would need for each
version to make testing and analysis viable?
Many thanks in advance,
>-- Original Message --
>Subject: RE: [Rasch] FW: Item order
>Date: Thu, 8 Mar 2007 11:14:11 -0500
>From: "Snider-Lotz, Tom" <tsnider-lotz at previsor.com>
>To: "Jane Scott" <janescott at iprimus.com.au>,
> <rasch at acer.edu.au>
>Actually Jane, I think it would be interesting to try two or more of
>those methods and see how they compare -- but I expect you don't have
>the time or subject pool to allow that.
>Will the subjects be giving a yes/no response or will they be applying a
>If you arrange the items as triplets, you're essentially turning the
>triplet into a single rating-scale item:
>While watching this film, I felt
>1 -- good
>2 -- happy
>3 -- elated
>Because of that, I'd go with the complete randomization, though not
>without a concern. My concern, which applies to any method you use, is
>that your results may be influenced by individual differences in
>interpreting the words you use. For example, to some people, "good" and
>"happy" may be pretty close together in meaning, while for other people
>they may be farther apart. There may even be people who think "good" is
>better than "happy."
>Ultimately, I think I'd prefer to do this study with more traditional
>rating scales, e.g.:
>How happy did this film make you feel?
>1 -- Not at all happy
>2 -- A little happy
>3 -- Happy
>4 -- Very happy
>(Not the best options, but you get the idea.)
> -- Tom
>From: rasch-bounces at acer.edu.au [mailto:rasch-bounces at acer.edu.au] On
>Behalf Of Jane Scott
>Sent: Thursday, March 08, 2007 4:21 AM
>To: rasch at acer.edu.au
>Subject: [Rasch] FW: Item order
>I am developing a multi-dimensional scale measuring people's level of
>engrossment with a film. I am hoping people can advise me with regards
>to the ordering of items, as I can't seem to find a relevant article.
>Within each dimension (let's consider feelings for example), I am
>exploring several different feelings someone might have to a film,
>building difficulty/intensity by using 3 items to explore each theme.
>So for example, I have items such as....
>I felt good
>I felt happy
>I felt elated
>which measure happy feelings
>and items such as......
>I felt apprehensive
>I felt scared
>I felt terrified
>to measure fear
>My question relates to how these should be ordered. Do I keep them in
>their triplets with ascending levels of endorsabilty / difficulty? Which
>is good because it makes the person think, "yes, I felt happy when I
>watched the film", but realising they are being pushed to the next
>level, they might more thoughfully consider, "but did this film make me
>Or do I mix the items up so "I felt good" and "I felt apprehensive" (ie.
>2 easily endorsed items) are answered first, and then the harder items
>(eg. elated, terrified) are answered last
>Or do I mix the items up completely randomly so I might have scared
>followed by good followed by something else - hence mixing up both the
>feelings and the level of difficulty
>Any advice you could give me would be most appreciated as I need this
>resolved really quite soon!
>School of Marketing
>University of New South Wales
More information about the Rasch