Dichotomising a 5-item Likert scale?

[Jean Providence NZABONIMPA]

Dear Listers,
Happy 2014 !
I am doing a comparison on various variables from two different
dataset and some levels of measurement are the same, thus making the
comparison possible. What is the best way of collapsing a 5-item
Likert scale into a dichotomous variable? Any guide or articles to
read would be so much appreciated. The Likert is structured as
follows:

Dataset 1:
                Life goals                      Very much       Much    Unsure  Very
little  Not at all
Have an interesting job               5            4                 3
           2            1
Have a career in sciencey             5            4                 3
     2          1

Dataset 2:
                         Life goals         Much mentioned      Unsure/not mentioned
Have an interesting job                        2                                1
Have a career in sciencey                      2                               1

I am inclined to collapse "very much" and "much" of Dataset 1 into the
same category while the rest would also fall under another category,
assuming that respondents who were "unsure" (the middle-point) about
their plan or goal would more likely be on the side of those with
little or no interest in the item measured. I also thought of dropping
the middle-point and unfortunately realised that the sample size would
be reduced to a great extent, thus affecting the analyses to be
performed on subgroups.

Is that right, sound statistically or methodologically?

Many thanks for your support.
Kind regards,
Jean
--
Jean Providence NZABONIMPA, PhD Student
Stellenbosch University
Cellphone: [phone]
                  [phone]
South Africa

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD