Re: Greetings everyone!

Devi Jankowicz (
Thu, 28 Jan 1999 23:58:12 +0000

Jonathan Lee writes:

>The main problem that I face is that when doing the triadic
>comparison (i.e., How are 2 of these alike and different from the third)
>students will have a tendency of being broad with their constructs (e.g.,
>they will state that writing an essay and studying for an exam are
>similiar because they are both academically relatedin relation to playing
>cards) instead of supplying
>constructs that will help us to understand why they procrastinate on these
>tasks (e.g., writing an essay and studying for an exam are "personally
>umeaningful"). We are hoping to try and elicit the latter types of
>constructs because they will help us get a better understanding of why
>academic procrastinators delay on their school work. ARRRGGHHH! I'm losing
>sleep thinking about this thesis!

A few suggestions:

a) that depends partly on the breadth of the realm of discourse you offer
them. If you start off with the purpose of assessing "what students do in
general" then yes, you might find that you have "problems" arising from
the fact that one of the elements you've provided was "playing cards",
when you subsequently decide that you're _really_ interested only in the
distinctions they make between their academic activities! On the other
hand, if you're clear from the start that it's only "academic tasks faced
by students" that you're interested in, which presumably doesn't include
the element "playing cards", then you'll you'll exclude that element and
seek to find out about why they procrastinate within the _academic_ realm
of discourse you've provided them.
(Is the problem that you feel I've made a big prejudgement, laying my own
construct that playing cards is not an academic activity onto the
respondents? Aha and alas! All measurement involves the researcher's
prejudgements, and these aren't eliminated in a pcp approach! The issue
is the general one, "Do the judgements you've made reflect your research
objectives?" You're expected to make design decisions in any research!

b) try remining them of the realm of discourse you're interested in, by
tagging on a "from the point of view" phrase as you offer each triad of
e.g. "Writing an essay; studying for an exam; preparing for a tutorial:
in what way are two of these alike, and one different, from the point of
view of why you might, or might not, procrastinate in doing them?"

c) work with each student individually, rather than working with a whole
class of students. What I have in mind here is that the grid is just one
form of structured interview (unique because open-ended, of course); and
all the stuff you know about rapport, probing, empathy, exploring what
the interviewee actually intends to say, assisting him/her in expressing
distinctions that are pre-verbal before they're uttered, _negotiating_
meaning with the interviewee, similarly applies.

d) in fact, more so. How about not writing anything down as soon as your
interviewee has offered a construct, but doing some laddering down first?
e.g. if the construct that's offered is "very academic" as opposed to
"more practical", you might ask "can you say a little more about that?
What is it about them that you recognise as "academic"? What is it that
makes them "academic", that the odd-one-out doesn't have?" That's a good
way of avoiding a bunch of "motherhoods": a way of obtaining more
behaviourally specific constructs, for example, if that's what you're
looking for. Difficult to do if you're doing a group grid in the sense I
described in c) above, and a very good argument for drawing up grids
individually, student by student, instead.

That's timeconsuming, I know, but, consider: if you're planning to obtain
more than just 3 or 4 grids (which can be analysed individually without
your reader getting lost among all the detail), you're possibly going to
want to do some sort of content analysis; and for that, you don't really
need more than 25 or so respondents x 10 to 12 constructs each. Since
your unit of analysis is the construct, not the respondent, 250-300
constructs is _ample_ for a content analysis, which limits your
respondent sample size to something that's handleable through individual

But, ultimately, on this issue of time, it's the same old story as for
any research: the more effort you put in to understanding what your
interviewee intends, the more useful the results will be. "Nowt for
Nowt", as we say in the north of England!

Kind regards,

Devi Jankowicz