What test(s) should I use?

#1
Hi,

Apologies if this is a very basic statistical question, but my knowledge is limited, at the moment.
I have gathered data for via survey at work, female participants only.
I have amassed 100+ answers regarding female use of the Internet in a tertiary learning environment.
One question is, do you use your mobile device in the classroom for educational and/or private purposes?
Result was: Education only 13%; Private 8%; both 78%; never 1%.
Another set of 7 questions was asked relating to the benefits of Internet use, these answers were transformed into one variable/subscale (a known subscale in the reserach field), I did test it for reliabilty (0.709). Answers ranged from Strongly disagree (1) to Strongly agree (5). I've used SPSS to calculate the validity of the subscale.
My question is, what tests do I need to run to answer, statistically, the following question: "Women who believe the use of the Internet is beneficial are also likely to use their mobile device for educational use?.
I'm not sure what tests to run i.e. t-tests, anovas etc. Any info would be greatly appreciated :)
 

Karabiner

TS Contributor
#2
Do you have a measure for “...believe that it is beneficial“?
What you describe is a continuous score, but not a categorization
such as belief/disbelief. Would you consider to rephrase your
question, e.g. “those women who use the internet for educational
purpose have a higher mean score on the belief scale than the
remaining women“?

With kind regards

Karabiner
 
#3
Hi Karabiner,

Thanks for your suggestion. I think I should change it to that, it does read better. What tests would I need to carry out to show that the that statement is true or false. I have ran a cross-tab using the 2 variables and got a chi-square of 0.015. Do I need to run any other type of tests, in your opinion? I'm struggling to understand what variables I have. One is a likert scale (1 to 5) and the other is asking a question on usage i.e. for educational use only, private use only, both and never. I'm not sure what tests to run tbh.

Again, thanks for the reply. :)
 

Karabiner

TS Contributor
#4
Seemingly you have 1 grouping variable (educational use yes/no)
and one interval scaled variable (a composite of 7 Likert-type items)
measuring degree of belief.

You could compare the means of the scale between groups using
the independent samples t-test (preferably the independent
samples t-test with "Welch correction").

Maybe some of these sources can be helpful in the future
https://duckduckgo.com/?q=Decision+Tree+statistical+tests&t=h_&ia=web

With kind regards

Karabiner