# Likert scale analysis: Help!

#### JustJoe

##### New Member
Hi, I'm new here.

I'm at my final bachelor year and in order to graduate, I need to write a thesis.
The subject of my thesis regards the increase of narcissism due to time spent on an MMORPG and the amount of praise received in said MMORPG.
In other words, my independent variables are praise and time, whilst my dependent variable is level of narcissism.

The tests I used are as follows:
- NPI-40 for narcissism
- Amount of hours played per week => time

Now here comes the tricky part. I couldn't get my hands on a questionnaire about praise that fit my subject, so I made a 17 item Likert scale.
I already tested the reliability and the cronbach's alpha is 0.79.

I shuffled around the internet to look for a way to analyze the Likert scale in my context.....and suffice to say the results are very confusing.
I only understood that the way a likert scale is analyzed depends on your hypothesis.
So here's mine:
Hypothesis 1: narcissism increases as more time is spent playing an MMORPG
Hypothesis 2: narcissism increases as more praise is experienced during gameplay.

Also, I contacted my supervisor regarding this and according to him a regression analysis is appropriate. Although, he's not very sure either.

#### Injektilo

##### New Member
How are you measuring praise?

Anyway, I would do as your supervisor suggests. A regression analysis will indicate the impact of praise and the amount of time have on narcissism. You may also want to look if there's any interaction between praise and time.

#### JustJoe

##### New Member
How are you measuring praise?

Anyway, I would do as your supervisor suggests. A regression analysis will indicate the impact of praise and the amount of time have on narcissism. You may also want to look if there's any interaction between praise and time.

It's good to hear that a regression analysis would be appropriate to use and thank you for the tip about the interaction between praise and time.
Also, I'm measuring praise through the likert scale I made- it basically asks the respondents how often certain experiences (related to praise) have happened to them.
It's a 17 item 5 point likert scale varying from never, rarely, every once in a while, sometimes, to almost always.

Last edited:

#### dhvalden

##### New Member
Multiple Regression analysis should do the trick, but make sure you have some other variables to control your results. it seems to me that the questions you have needs at least a semi experimental design or a panel design and some demographic control variables, in order to know the direction of the relations of the variables and to control for effects of other important variables. E.G. maybe narcissistic people tend to report higher levels of praise, instead of being the praise that make people more narcissistic, or maybe the age is important and your hypothesis are only true in teenagers and not in adult players.