Guttman's smallest space analysis I am trying to replicate an analysis as conducted by Raven, et al. (1998). A scale is analyzed with Guttman's smallest space analysis. First of all I wasn't able to find a suitable procedure in R and it appeared to me as if SPSS might be the solution. However even though I use the same scale it is impossible to reproduce the graph as printed in the original publication:

The graph I manage to get out of SPSS looks like the one below:

I suppose there must be some settings to adjust the dimension or scales.
Any recommendation to an R package or some guidance to the SPSS settings would be appreciated.

References
Raven, B. H.; Schwarzwald, J. & Koslowsky, M., Conceptualizing and Measuring a Power/Interaction Model of Interpersonal Influence, Journal of Applied Social Psychology, Blackwell Publishing Ltd, 1998, 28, 307-332 [Link]
 A: Multidimensional scaling (in particular smallest space analysis) is a very sensitive measure, and it is very unlikely two separate data sets (regardless of mirrored variables) will produce an identical geometric space. This is because the common spaces displayed are an overall relationship between every single variable in the data set, and thus easily manipulated when analysis parameters are changed (variables added, removed etc). Thus, unless you had obtained exactly the same results as the original paper, it is highly unlikely you will get the same plot. Although holistically the same, your participants will have had individual variability different to that of the original paper. Thus the common space (e.g. correlation in simple terms) will be slightly different from the original. 
What you are looking for is similar patterns in the geometric space. Just because they are not in exactly the same place on your geometric plot, when compared to your geometric plot, doesn't mean the same grouped clusters aren't there. For what it's worth, I'm not 100% sure the groupings from the original paper are replicated in yours. Are you using the same number of variables, exactly? 
Would I be correct in saying you are comparing the Euclidean distances between a series of variables? If so I would look to do multidimensional scaling within the PROXSCAL procedure of SPSS and use the common space function. I think this is likely to offer you a much better outcome. Just because the original paper used SSA doesn't mean a different form of multidimensional scaling (PROXSCAL, cluster analysis – gosh, even factor analysis with a specified number of groups, which in this case would be 11) would be inappropriate. You are looking to confirm factors. If you are unable to get the same matched factors, then perhaps there is a methodology or theoretical explanation to it.
It's difficult without theoretical understanding of the topic to fully appreciate whether the common spaces in your output are theoretically interesting (thus challenging Raven) or whether they are inconsistently confusing and disinteresting.
