Displaying the means of your semantic scale question

This example shows how to produce a representation of the means of a grid of semantic scale questions.

It uses the example pictured below (Q2.a to Q2.c) and scores the left as 0 and the right as 100, and then displays the average response on a scale.

It uses a semantic chart style supplied with Snap.

The style assumes that:

• You are only displaying the responses to the semantic scale questions
• These are not broken down in any way
• You have a derived variable All that is true in all cases. This is used to position the mean on a continuous bar.
• You create a Weight to convert the responses into a numerical value. (How you create the weight will depend on the number of response codes you have in your semantic scale. This example works with five.)

Creating the derived variable

1. Click on the Snap toolbar to display the Variables window.
2. Click on the Variables window toolbar to add a new variable.
3. Specify the Variable details:

   Set the Name to All.

   Leave the Label blank as it is not required.

   Set the Type to Derived.

   Set the Response to Single.

   Leave the Code Label blank .

   Set the Value to true.

   

4. Click to save your variable.

Creating the score weight

1. Open the Weights window by clicking or pressing [Ctrl]+[W].
2. Click to create a new weight. A weight is simply a way of matching values to codes for analysis.
3. Name the weight SM100 so you know what it is. Give it a label explaining what it is used for.
4. Type 0 next to Code 1 (code 1 is the value on the extreme left) and click [Tab] to create a new code value. You'll notice that Snap attempts to guess what value you might want to use.
5. Enter 25 next to code 2. Click [Tab] again.
6. Repeat and enter 50 next to code 3 and 75 next to code 4.
7. Code 5 was the extreme left response .Give it a score of 100 (the maximum).

   

8. Click to save your score.

Creating the table that is used to set out the semantic data

1. Click to display the Analysis Definition dialog for a table. Type the names of your semantic variables in the Analysis field (e.g. Q2a~Q2c)
2. Type All:All in the Break field.
3. Check the Transpose box
4. Select Means and Significances in the Calculate field.
5. Enter the weight (SM100) in the Means box.

   

6. Click [OK] to view your table.

   

Setting up the bar length

The semantic scale chart style uses the first All row to show the position of the average response, marked with a circle. It uses the second All row to show the full length of the scale. You must overtype the second All row values with the full length of the bar. You could use 100, but use 102 to give space for the circles when the mean is at the extreme end of the scale.

1. Double-click the first cell value in the second All row (Strong/Weak in the example). The Overtype Analysis Value dialog appears.

   

2. Enter 102 in the space and click [OK]. Repeat for all the other cells in that row.

   

3. Click to save your modified table.

Convert the modified data table to a chart

1. Click on the table window toolbar to open the Analysis Definition window.
2. Change the Type of analysis from Table to Chart.
3. Change the Style of analysis to Semantic Scale.

   

4. Select the Notes/Titles tab and change the Title to something that represents your chart.
5. Click [Apply] to display your chart.

   

6. Click to save your semantic analysis.

See Also Creating tables of your semantic scale question Creating a simple means chart of your semantic scale question