I am doing my dissertation about brand recognition. I have 5 questions with 7-point Likert Scales that measure brand recognition. I want to combine those variables into one in order to conclude if the brand has recognition or not. How can I combine the variables? Will I go to Transform - Compute Variables in SPSS and will I combine them by the median? And after creating a new variable can I do the descriptive analysis? (like calculating the frequencies and the median)
the seven items are not variables but they are indicators about brand recognition. If your data are normally distributed, calculated the mean and the standard deviation for each of the seven items, then calculate the overall mean and sd. If the data are skewed ly distributed, find the median for each of the seven items and calculate the overall median. In calculating the overall mean in SPSS, go to transform then go to compute variable, when you are here add all the responses of your total respondents divided by the number of respondents you have in your study. The result is now your overall mean. For overall median, combine all ratings of theyour respondents then determine the median using descriptive statistics.
Five questions are not variables rather items. You are right for computing variable to one, that can be through average or total. Later on, that one variable will be ready to perform any type of descriptive or inferential analysis
Yes, you are correct except, I combine via the "mean". Refer to the following SPSS steps: Transform / Compute variable / in “Function group” select “Statistical” / in “Functions and special variables” double click on “Mean” / in the “Numerical expression” the “Mean” function appears in the box, replace question mark with variable names with commas by “blocking” the ? and double click on the variable names (in the “Type and label” section) / double click one by one and type comma in between. Type label in “Target variable”, the construct name, then OK. Then repeat for the other constructs!
the seven items are not variables but they are indicators about brand recognition. If your data are normally distributed, calculated the mean and the standard deviation for each of the seven items, then calculate the overall mean and sd. If the data are skewed ly distributed, find the median for each of the seven items and calculate the overall median. In calculating the overall mean in SPSS, go to transform then go to compute variable, when you are here add all the responses of your total respondents divided by the number of respondents you have in your study. The result is now your overall mean. For overall median, combine all ratings of theyour respondents then determine the median using descriptive statistics.
If you like you can send your data to me and I will process it based on the objective of your study.
Thank you so much for your answers. I realized that my data is not normally distributed so I found the median of each of the five items. So now that I will combine those items into one variable, in order to do then the descriptive analysis, I will go to compute variable and then I select the median when I click in Statistical? I didn't understand that part if I select the mean, total or median in order to combine the items into one variable.
Dear Joao Cruz ,
If you designed the questions yourself you should start with a reliability analysis using Principal Component Analysis. If this works you can save the regression model and double check with a Cronbach's alpha calculation. Please see my study guide for more information.
- Badges
- Science topic
- Similar topics
- Medical Science Nursing
- Dissertations