One of the most useful forms of question is a rating scale, where you ask people to mark how satisfied they were with an item or a service. You can then analyse the answers to these questions to see if people are generally satisfied or dissatisfied, so you are more able to judge where to put the effort in to improve what you offer.

For example, here are five rating questions with a five point scale.

How do you rate the following?

You could use the responses to these questions to help you judge whether you should invest in a larger parking lot or better food.

This worksheet explains some of the issues about using the data as it stands, and describes how to set up a three column table and a simple bar chart to display whether people are generally satisfied or not.


When you are looking at the responses, you need to be careful that you know what is being counted, and that you compare like with like.

For example, how do you compare people’s satisfaction with parking and cleanliness if forty people answer your questionnaire:

  • ten people answer the parking question, of whom eight select Good
  • forty people answer the cleanliness question, of whom twenty select Good.

To make the comparison useful, you could:

  • ignore the people who haven’t responded, and compare the percentage of people who responded who think it is good (in this case 80% think parking is good while 50% think cleanliness is good).
  • count the people who hadn’t responded as neutral, and then compare the percentages (20% of all respondents think parking is good while 50% think cleanliness is good)

You need to then display your conclusions clearly.

Here are some of the ways you could display the rating data in tables.

Look at the responses in detail.

This tells you how many people and what percentage of people, selected which answer to which question.
AN3: q4a~q4e

Compare the questions as a whole

  • You could use confidence boxes to group the Poor and Very poor responses together as Negative, and the Good and Very good responses together as Positive and see which question scored highest on Negative or Positive responses.

    Counts Respondents

  • You could work out a value for all the responses for a question and see which question did best on average. One of the clearest ways of showing this is to give the responses a score ranging from negative to positive.

    AN8: Q4a... Scoring system from -2 to +2

  • You can reduce the table to three columns using derived variables, to make it easy to see the general trends. This worksheet describes how to do this.

    Table: Percentage ratings scale from derived variables

Summary of steps

This worksheet uses the satisfaction survey data supplied with Snap, and shows you how to create derived variables from the rating questions, so you can display the ratings in a three-column table, as Positive, Ok, or Negative.

The rating scale questions in the snSatisfaction survey supplied with Snap 10 are questions Q4.a to Q4.e. These appear as image maps in the survey.

The worksheet explains how to:

Step 1: Create derived variables for the three columns

You need to create a new three code variable to represent each five-level rating scale.

    1. Click Variables window button on the Snap toolbar to open the Variables window.
    2. Click New button on the Variables window toolbar to create a new variable.
    3. Set the Name to V4a to remind you that it is derived from Q4.a.
    4. Set the Type to Derived. (The values in this variable are derived from the answers to the rating scales questions.)
    5. Set the Label to Service (to remind you that it is the Speed of service question).
    6. Set the Response to Single.
    7. Click in the Label area for code 1

      VD: variable with code label 1 circled in red

    8. Type Negative. This is going to be the heading of the column in your analysis.
    9. Click [Tab] to move to the Values column.
    10. Enter an expression to select the Poor or Very poor responses to the rating question. In this example, these are the responses 1 (Very poor) and 2 (Poor). The expression that selects these responses to the appropriate question (Q4a in this example) is Q4a=(1,2). I.e., if the answer to Q4a is either 1 or 2, this variable will have an answer of Negative.

      Code list Q4...

    11. Click [Tab] to move to the Label column for code 2 and type OK.
    12. Click [Tab] to move to the Values column and type Q4a=3 (i.e, the answer to Q4a is code 3, OK).
    13. Click [Tab] to move to the Label column for code 3 and type Positive.
    14. Click [Tab] to move to the Values column and type Q4a=(4,5)(i.e, the answer to Q4a is code 4 or 5, good or very good).

      VW: Dervied variable window dialog showing ratings

    15. Click 1 2 3 button on the Variables window toolbar to check the number of responses for the parts of your new variable. This displays the correct counts.

      VD: counts shown for new derived variable

    16. Click Save button to save your variable.
    17. Highlight the variable that you have just created in the Variables window, and click Clone button to clone it.
    18. Edit the new cloned variable. Change the variable Name to V4b, the Label to Cleanliness and change the question number used in the code values from Q4a to Q4b.
    19. Click Save button to save your variable.
    20. Repeat cloning and editing the variable for Q4c (Parking), Q4d (Quality of food) and Q4e (Choice of food).

Step 2: Create an analysis table from the derived variables

You can now create analyses based on the new derived variables.

    1. To create a simple table click Table button on the Snap toolbar to open the Results Definition window for a table.
    2. Enter V4a ~ V4e as the analysis expression. (If the variable names are not in sequence, you will have to enter the variables individually, in the format V4a, V4b, V4c, V4d, V4e.)
    3. Check the Analysis Percents box, so you can see what percentage of people chose each answer.

      Definition tab

    4. Click [OK].

      Table: Percentage ratings scale from derived variables

Step 3: Create a stacked bar chart from the derived variables

Although the table displays all the information you need, it might be easier to see it in a chart. This step shows creating a stacked bar chart to display the same results.

    1. Click Chart button on the Snap toolbar to define a chart.
    2. Set the Style to Horizontal Stacked Bar 2D Labelled.
    3. Select the analysis term you used for the table from the drop down list by Analysis.
    4. Check Counts if it is not checked.
    5. Check the Transpose box.

      RD: section of def for stacked bar chart for satisfaction

    6. Select the Notes/Title tab and type Satisfaction ratings as the title of your chart. Clear the values for the Analysis and Break chart axis titles.

      RD: Notes title tab showing satifaction ratings title

    7. Click [Apply] to display the chart. (You may need to drag the definition window away from the chart window.)

      Chart: stacked bar chart showing all figures

      This shows the satisfaction ratings as a single bar for each question. The section of the bar representing negative ratings is red. The positive ratings section is green. It is easy to see that people are most dissatisfied with Service and most satisfied with Parking.The number of responses for each rating is given on that section of the bar.


This worksheet has provided a brief introduction to working with derived variables to display rating information in different ways. You can use derived variables in many other ways. For further information, see the section on Derived variables in the manual and online help.

You can find more information about creating the different tables shown at the beginning of the worksheet in:

If there is a topic you would like a worksheet on, email to