Research Proposal.

Hats_Off

Hey, I'm trying to write a research proposal for my second year research lecturer and I'm having maaajor difficulties with my Research question.
So here's what I should know but I don't.
I was contemplating going with,
"To investigate the correlation between up-tempo instrumental music on mood in adolescence"

So to measure a correlation I obviously need more than one variable. I'm confused though, are music and mood both variables?
Would the study be better if I were to compare up-tempo instrumental music with up-tempo Lyricised [is that even a word] music on mood?


Help would be greatly appreciated.

p.pete

Music is your independant variable (what you're manipulating) and Mood is your dependant variable (what you're measuring). Presumably you'd compare different styles of music as different conditions and then compare conditions. Also I presume you're actually allowed to run the project?

Hats_Off

Thank you Pete, that's really helpful.
I'm actually going to do the research on a sample of the college-going population rather than adolescence.

Would I then use an ANOVA or a t-test?
Could I introduce a physical test to test heart rate, muscle tention etc?

Cinful

Hats_Off wrote: »

Research question.

"To investigate the correlation between up-tempo instrumental music on mood in adolescence"

Research Hypothesis (positive-directional relationship): As up-tempo instrumental music increases, adolescence mood improves.

• Null Hypothesis (the one you test): There is no relationship between up-tempo instrumental music and mood in adolescence.
  
• Independent variable: music (you need to define this term, using citations)
• Dependent variable: mood (you need to define this term, using citations)
  
• Design: Solomon 4-group (2 experiment groups, 2 control groups)
  
• Sampling: Random assignment of subjects to experimental and control groups
  
• Experimental Group 1: subjects only receive up-tempo music.
• Experimental Group 2: subjects only receive down-tempo music.
• Control Group 1: subjects pick their own music without influence from you
• Control Group 2: no music during experiment
  
• Sample size: Minimum number of subjects (n=30) for parametric statistical analysis; less than 30 for nonparametric; e.g., you could have each group, experimental and control, randomly assigned with 10 subjects each, for a total of 40 (Below 10 per group not advisable for statistical analysis purposes; i.e., you could commit a Type II error as a result of insufficient sample size, by accepting the null hypothesis, when in fact it is false).
  
• Data collection: Cross-sectional (collect two times: pre-experiment and post-experiment during one session)
  
• Data collection methods: (pre-and post-survey instrument of mood)
  
• Statistical tests:
• ANOVA at p<.05 to assess if there is a statistical difference between experimental and control groups (if n=30 or more subjects; recommend Wilcox U if less than 30 subjects in experiment).
• For correlation, you could assess bivariate relationships using Pearson r, with a significance of p<.05 (if n=30 or more subjects; Spearman rho if less than 30 subjects).

Limitations:

• Experimental design has controls for comparison purposes, but is not natural.
• Significant difference and correlation are necessary, but insufficient conditions for causation.
• Small experimental and control groups for this study increases the likelihood of committing a type II error, or the likelihood of finding no significant differences or correlations (accepting the null hypothesis), when in fact there may be if the number of subjects were increased.
• May also be limitations associated with the population you randomly draw your sample from?

The above is an experimental design, or what they call a laboratory experiment; whereas, you may wish to measure your variables in a more natural setting? If so, all the above would have to change, and perhaps a quasi-experimental design would work, or something more qualitative-ethnographic in measurement? I can design these too. It's my bag.:D

Hats_Off

Wow, thank you so much.

Cinful

Hats_Off wrote: »

Wow, thank you so much.

Any time;)

Hats_Off

Cinful, we have been previously warned against creating our own survey or questionnaire. Is there one on mood that you would suggest?

p.pete

Have you done any research looking at studies that take a measure of mood? If you scan the method section it should be near the start what material they use.
Re taking heart/muscle measurements, does this study need to go past an ethics board before being run? have you a project supervisor to discuss what's feasable?

Cinful

Hats_Off wrote: »

Cinful, we have been previously warned against creating our own survey or questionnaire. Is there one on mood that you would suggest?

See attached journal article. It pertains to your research topic. Note the semantic differential at the end. Layout is poor. Can be reformatted. Useful?

Article Link: http://clearinghouse.missouriwestern.edu/manuscripts/259.asp