My sample size is 280. Should I first determine specific cases/respondents that are Extroverted, Neurotic, Agreeable, and so on? Or can I use the entire data set, and take corresponding personality trait items as IV to look for their predictive power on DV?
Your research question should guide you. Are you interested in individual items of your latent variables or are you interested in relating your latent variables to your DV?
Dear Sumiran,
You should use all the data as scale variables. I assume you are referring to allocating them to nominal groups for each of the five personality traits. You should never prematurely reduce the quality of your data like this as it will reduce the quality of your analysis.
The first step in a regression model is descriptive statistics. I would recommend creating a scatter plot of your DV against each of your IVs in turn. This will enable you to assess the assumptions of linear regression.
There are lots of other things to consider in multiple linear regression but that's how I'd advise starting.
For most purposes, you want to use the scale scores on the five traits. Do not reduce to "types" and do not bother with individual items.
First, report descriptive statistics: mean, standard deviation, and (if possible) internal consistency for each scale.
Next, report zero-order correlations among the five traits and between the traits and your DVs (something to do with Snapchat use?).
Finally, run your multiple regression equations.
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Sampling Design
- Sample Size