Boards Golf Society are looking for new members for 2022...read about the society and their planned outings here!
How to add spoiler tags, edit posts, add images etc. How to - a user's guide to the new version of Boards

# Analysing Likert Scale in Excel (SurveyMonkey)

• Registered Users Posts: 41 ✭✭✭

Hi all,

I am doing quantitative research for my thesis, I ran an online survey through SurveyMonkey.

Having downloaded the result into an Excel file, I need help on correlation or analysing the "relationships" between the different pieces of data, and the conclusions that can be drawn from these relationships. How can I do this through excel and with results from a likert scale?

Some of the likert scale questions have up to 5 statements in them. Survey monkey has given me the average rating per statement.

I have been advised to Cross correlate all of my data and to get a much more indepth picture of my results. Is this right?

• A correlation matrix is usually a good first pass assessment of possible relationships. The problem with likert questions is that, depending on the number of items on your questionnaire, many are likely to be correlated with each other, and some reduction may be needed. This is usually achieved through factor analysis, but maybe park that for now until you have a good look at the raw data. Excel can do quite a bit in the way of analysis, and you should be able to google a decent guide easily. The website for Alan Bryman's social research methods textbook comes with a great free guide to analysis in excel. If you want advice on anything beyond correlation ask away - although the options you have will depend both on the nature of your data, and the research questions of your thesis.

• I ran an online survey through SurveyMonkey...

I need help on correlation or analysing the "relationships" between the different pieces of data, and the conclusions that can be drawn from these relationships.

I have been advised to Cross correlate all of my data and to get a much more indepth picture of my results. Is this right?
What type of sampling method did you use to select your online subjects? This sampling method will be important if you intend to estimate population parameters from sample statistics, or if your method allowed for such estimates when making "conclusions that can be drawn from these relationships."

For example, if you used convenience or accidental sampling methods as many do for online surveys (i.e., nonprobability sampling), attempting to estimate confidence levels or intervals, and if your sample was representative, may be problematic. Depending upon your methodology, it may be that your conclusions will be limited to those that replied, and cannot be generalised to a larger population. We need more details about your research before we can be of help.

If you used a random method to ensure that each and every member of the population had an equal chance of being selected, there will be other issues that may affect what "conclusions" you can draw from the online survey data; e.g., sample size, response rate, systematic error, etc. We can discuss these in greater depth here if you wish.

Before we can proceed analyzing the online survey data with statistical methods (e.g., correlations, etc.), and be able to draw "conclusions," understanding the sampling method you used is essential.

• You could do a Principle Component Analysis to break down the data into factors and then apply a Cluster Analysis to determine groupings (based on factor scores) from your sample population.

• You're better off using Matlab, or R instead of Excel. They're much better suited to this kind of work. As above PCA is a good choice to reduce the dimensionality of your data.

• Black Swan wrote: »
What type of sampling method did you use to select your online subjects? This sampling method will be important if you intend to estimate population parameters from sample statistics, or if your method allowed for such estimates when making "conclusions that can be drawn from these relationships."

For example, if you used convenience or accidental sampling methods as many do for online surveys (i.e., nonprobability sampling), attempting to estimate confidence levels or intervals, and if your sample was representative, may be problematic. Depending upon your methodology, it may be that your conclusions will be limited to those that replied, and cannot be generalised to a larger population. We need more details about your research before we can be of help.

If you used a random method to ensure that each and every member of the population had an equal chance of being selected, there will be other issues that may affect what "conclusions" you can draw from the online survey data; e.g., sample size, response rate, systematic error, etc. We can discuss these in greater depth here if you wish.

Before we can proceed analyzing the online survey data with statistical methods (e.g., correlations, etc.), and be able to draw "conclusions," understanding the sampling method you used is essential.

I used convenience sampling for the online survey. The likert scale was used as part of the Technology Acceptance Model using the constructs Perceived Usefulfulness, Perceived Ease of Use, Subjective Norms, Self-Efficacy, Perceived Cost, Security and Trust. Using this to measure consumer adoption and use of Mobile Banking.