Name: Jesus Salcedo
About Jesus: Jesus joined SPSS recently as an education consultant for the SPSS and Clementine® product lines. Previously, he worked as a statistical consultant and college professor. He received a PhD in Psychometrics from Fordham University. In his free time, Jesus enjoys playing baseball, hiking, and traveling.
Recoding reverse-coded items in SPSS
As a researcher, you often write several questions to assess a construct. For example, rather than asking employees to rate their overall job satisfaction (the construct), you would instead them ask about different aspects of their job satisfaction, such as their relationships with their boss and co-workers, the amount of vacation time they receive, the amount of any bonuses they receive, how well they like their health plans, etc. Asking several questions allows you to better understand a construct.
Sometimes, however, you would like an overall score for the construct. You could add the respondent’s scores across the series of questions in the survey to assess the construct and use the total as an overall score. Although this process works well, you need to be careful of any reverse-coded questions that might appear in your survey. Reverse-coding occurs when a question is phrased in a way that is opposite to most of the other questions in a survey. For example, imagine assessing job satisfaction on a 7-point rating scale with the endpoints, 1 = Strongly Disagree and 7 = Strongly Agree. You might have items like these:
|
The bonuses I receive are fair. We have good company health plan. I have a good relationship with my co-workers. I do not have a good relationship with my boss. |
This last item is reverse-coded because a response of 7 = Strongly Agree indicates low satisfaction, while with the other items, a response of 7 = Strongly Agree indicates high satisfaction.
Reverse-coded items are typically used to help ensure that respondents are reading survey questions and not just providing the same response to each question. In the example above, however, you can’t simply tally the scores on all the items to compute an overall job satisfaction score—first, you need to recode the reverse-coded item.
To recode a reverse-coded item in SPSS, click on Transform->Recode into Different Variable. Name the new variable you will create. Then, click on “Old Value” and specify all of the original values for the variable, and go to “New Value” and specify all of the new values for the variable you are creating, as shown below:
After you’ve created the new variable, you can add up the scores on all of the items to produce your overall score. If you would also like to assess the reliability of a construct, you will need to follow the same procedure above to recode reverse-coded items.
We offer SPSS courses at locations around the world.
Find a course in the location nearest to you.
Predictive Analytics
can make your organization
more
successful
Resources