Research and Statistical Support

MODULE 4

Recode a reverse coded 5-point Likert scale item

Start off by importing the ex3reverse.sav into the Data Editor window of SPSS. This data set is fictional and contains 100 cases and 7 variables (6 items & a case identifier).

For this example, a 6 item paper survey was used to assess people's agreement with a proposed national law making April 15th a holiday (e.g. day off). The example items each have the same 5-point Likert response scale. On the paper survey given to participants, numerically higher responses correspond to greater agreement. However, question number 4 (q4 in the data set) was worded in such a way as to reverse code the response set. This is to say, persons responding with a 1 (typically Disagree Strongly) are really expressing that they Agree Strongly with the idea of giving citizens the day off on April 15th. The items are listed below:

1 = Disagree Strongly          2 = Disagree          3 = Not Sure          4 = Agree          5 = Agree Strongly

q1.) I think I should have April 15th off from work.

q2.) I am bitter when I have to work on April 15th.

q3.) I do not like leaving my home on April 15th.

q4.) I enjoy working on April 15th.

q5.) I'm quick to leave work on April 15th.

q6.) I think everyone should have April 15th off from work.

Our goal here is to recode q4 so that we can sum the scores of each item to create a total score in which numerically higher numbers represent greater agreement with having April 15th off from work.

Start by going to Transform, Recode into Different Variables... (using a different variable preserves the integrity of the original data).

You should now see the recode dialog box. First, highlight and move q4 into the "Numeric Variable --> Output Variable:" box using the arrow. Next, type q4_RC into the "Output Variable" box and type q4 Recoded into the "Label" box (both on the right). Then click the "Change" button. Next, click on the "Old and New Values..." button.

You should now see the following dialog box.

(1) Start by typing a 1 in the "Value" field under "Old Value" and type a 5 in the field under "New Value" and then click on the "Add" button.

(2) Next, type a 2 in the "Value" field under "Old Value" and type a 4 in the field under "New Value" and then click on the "Add" button.

(3) Next, type a 4 in the "Value" field under "Old Value" and type a 2 in the field under "New Value" and then click on the "Add" button.

(4) Next, type a 5 in the "Value" field under "Old Value" and type a 1 in the field under "New Value" and then click on the "Add" button.

(5) Next, click on the "All other values" circle under "Old Value" and click on "Copy old value(s)" under "New Value"; and then click on the "Add" button. We do not need to change values of 3 because they substantively mean what we want them to mean.

(6) Next, click on the "Continue" button and then click on the "OK" button to complete the recode.

Now we can use a compute function to sum the items to produce a total score in which higher numbers reflect greater agreement and lower numbers reflect less agreement. As we shall see; the compute function can also be used to recode a variable or item.