I'm fairly new here in this site.
I'm currently taking Bach. Psychology, and Research Methods is one of my core subjects. I have conducted an experiment titled "Gender Differences and Effect of Types of Music on Emotion".
So before presenting the problem I'm having, let me briefly describe my experimental design. So basically, it's a 2 (Male/Female) x 2 (Sad/Happy Music) Between Subjects design with 24 participants. For the first session, 12 participants (6 males and 6 females) are exposed to Sad song. For the second session, another 12 participants (6 males and 6 females) are exposed to Happy Song.
I used the Genova Emotional Music Scale (GEMS-45) for both pre-test and post-test to measure if there are any changes in Emotion after the participants listen to a music (either sad or happy song). The pre-test serves as the baseline of their current emotion.
Some confusions I'm having are:
From the first session (with sad song), I added the scores from 3 items (sad, tearful, and sorrowful). While for the second session (with happy song), I added the scores from 3 items as well (joyful, happy, and feel like dancing). Or should I take into account all 6 items (sad, tearful, sorrowful,joyful,happy, feel like dancing) for both sessions?
In the questionnaire, the Likert Scale ranges from 1-5. Now, for the sad session, I added negative values on the scores? This is to indicate that negative scores = negative emotion which is sad. So for example, a total score of a participant (in the sad session) for the pretest is 3, and 7 for the postest. The difference between the scores (postest - pretest) would be 4. It means that the participant actually felt sadder after listening to the song. So if the participant scored -3 on the pretest, and 7 on the post-test, the difference would be -4 (SAD)
Now, same goes to the happy session. If the pretest is 4, and the post-test score is 9, the difference would be 5 (or +5, which indicates positive emotion = happy) correct? The reason why i came up with the negative value for sad is to avoid SPSS being confused. But I'm not too sure about this.
Or, would I have to create a new scale, which consists of very sad to very happy on a 7-Point likert scale (1 being very sad, 7 being very happy)?
I hope to get some feedbacks about this as I'm having trouble with it. I know I can run ANCOVA (pretest as covariate, and post-test as DV) or Independence Factorial (Emotion, which is the differences between post-test and pre-test as the DV, and the gender and music as IVs), but I'd hope to enter the data correctly before running any stat-test.