How to analyze Likert survey data?

[SlideRule]

I've been asked to analyze some subjective survey data. This is a new challenge for me. My previous training and experience has been with measured engineering data.

The person who wrote the survey form made extensive use of five-value Likert scales because he, ". . . remembered from college that they are the best way to ask survey questions." Likert scales are where each participant is asked to choose a number indicating their relative agreement with a given statement. There are more than 100 such questions in the given survey.

Likert surveys are not covered in my available textbooks. I've found a few Internet sources describing ways of analyzing Likert surveys, but those sources give sparse explanation, and conflicting recommendations.

Unfortunately, I'm told that the analysis needs to be done quickly, and so I must quickly become an "expert" on Likert surveys.

Can anyone in this forum please give a good online source about analyzing Likert surveys? I need something that, with an hour or two of reading, will give me a firm start with this task. Thank you.

[SadieKhan]

http://www.ehow.com/how_4855078_use-...-analysis.html

I faced the same problem once and I couldn't get much help either, in the end I just used a reference and added in the research and made comments on the basis of cross tabulation, I will try to dig the reference out for you, in the meantime you can check the link above.
Sadie

[SadieKhan]

Ok here you go..
Cross tabulation:
Zikmund (1997) stated,“"the mere tabulation of data may answer many research questons. In fact, many studies do not go beyond examning the simple tabulation of the question by question responses to a survey. On the other hand, although frequency counts and percentage distributions, or averages, summarize a considrable amount of information, stopping at simple tabulation may not yield the full value of the research. Most data can be further organized in a variety of ways. For example, in a survey , that samples both men an women, it is common to analyze the data by separating the respondents into groups or categories based on gender. Analyzing results by groups, categories, or classes is the technique of cross tabulation" (p. 534).

ref:
Zikmund, W.G. (1997). Business Research Methods (5th ed.). Fort Worth, TX: The Dryden Press.


I am not sure of its application in statistical analysis, but it worked out for me
HTH

[SlideRule]

Here are some links that I found on the general topic of analyzing survey data. They don't address Likert scales solely, but they do look helpful for a math person like myself who is tackling survey data for the first time.

http://www.audiencedialogue.net/surveyreport.html

http://plus50.aacc.nche.edu/docs/pub...eyAnalysis.pdf

[CowboyBear]

Resolving the 50-year debate around using and misusing Likert scales

TLDR; it's reasonably justifiable to treat Likert scale data at the summated scale level as interval in nature and use conventional parametric tests (although there's a fair bit of debate here). Presumably your 100 items reflect a handful of individual scales.