After quite an elaborate search on the web I had to admit not to be able to find a suitable way to represent questions using the Semantic differential (a.k.a. the Osgood scale as well).

Having both SPSS and MS Excel to my disposal I cannot find an appropriate way to show the output of this kind of questions (like http://allpsych.com/researchmethods/ima ... ential.gif on the webpage of http://allpsych.com/researchmethods/res ... tware.html ). In short questions aks the people's opinion of certain (experiential) elements of which every aspect is put on (usually a 7-point) scale ranging from this aspect to its opposite, e.g. questions about a store where the personnel is 1) friendly - unfriendly 2) professional - amateuristic, etc., all ranging from 1 - 7.

My students pose me questions how to use this kind of questions and although it's not difficult to ask people this kind of questions in a survey (paper), the workflow of processing these data in SPSS is still a riddle to me. And after browsing several hours on the web I get the impression it is not a well-known recipe in SPSS or Excel, so I hope someone here knows how to manage this problem.

Sorry, not clear as to your question.

Are you asking how to enter the results into SPSS?
Are you asking how to create differential questions?
Are you asking how to analyze such questions?

Sorry, let me clarify. Although I've entered the data similar to Likert scale questions (in this particular case the Osgood scale questions range from 0 - 4) I am unable to generate a chart that shows the results of a series of Osgood scale questions in an output type that is usual for this question type. So maybe the setup of my questions is not appropriate to generate this kind of chart, but that should not be the problem since recoding could solve that.

Thus, in essence my question is about the representation, how to generate a chart where questions of this Osgood scale (aspects of what is to be researched) are shown appropriately? The simple bar chart where separate variables are shown appears a bit like this, but instead of bars I need lines and instead of the variable names I need the two opposites of each aspect, one on top, one on the bottom, and (sorry, this sounds so demanding, but this Osgood scale simply works like this) this chart should be rotated 90 degrees, so that the opposites of all apsects are on the same row. Well, the website with the graph on my original post shows it better than a bunch of words.

Data example
Case Friendly Professional Variety
1··········3············4············2
2··········2············2············3
3··········3············4············1
4··········1············3············1
5··········2············3············1

So the means will be:
-·········2.2·········3.2·········1.6

And the graph should show something like (with the X's connected by lines), the dots (·) are used for the purpose of spacing only.

····················|·0·······2········4·|
Unfriendly·······|············X·········|·Friendly
Amateuristic····|················X·····|·Professional
Simple···········|·······X··············|·Varied

Well, imagine those '|' exactly aligned as perpendicular lines but that's a bit difficult in HTML here.

I use both versions 15 and 16. Hopefully this makes sense.

Hello,

I am student and am currently doing my dissertation in which I think I am having the same problem.

I have semantic differential scales within my questionnaire and am unsure how to analyse the data in such a way that will give me a graphic output such the one attached? I have version 16.0 if this helps?

Many thanks