Dear Pahlaj Moolio

The 5-point Likert scale has 5 response possibilities, two extreme sides and a neutral option related to the intermediate answer alternatives. Very Satisfied, Satisfied, Neutral, Dissatisfied, and Very Dissatisfied are examples of 5-point rating scales for gauging satisfaction.

Best Regards

Dr. Fatemeh Khozaei

Dear colleague.

Maybe it will serve as an inspiration :

Conference Paper The Cognitive, Affective and Conative Components of Consumer...

A Likert scale is an evenly scale from which respondents choose the level of agreement or disagreement. It can be used to measure the product or service analysis whether the consumer is pleased or having issues with a particular product or service. 5 Point scale analysis is frequently used because respondents can easily answer questions. They don’t have to think and write a lot of lines in order to answer. They just simply can choose from already given answers, Likert scale questions help to sample opinions and get relevant information. It gives an insight into what audience is thinking.

You might find the following papers useful for your Q:

Article A pedagogical approach to integrate parametric thinking in e...

Article Cognitive Styles, Gender, and Student Academic Performance i...

Multinomial regression would be best suited for this analysis if the Likert scaled variable is the dependent variable and if used as it is. On the other hand, recoding the dependent variable is viable, especially when it is necessary (negligible numbers for one or two levels of the scale). You will thus have to use binomial regression.

I hope that is not the order that the scale points were presented in; one should always begin with the negative. And the neutral midpoint is generally a mistake. But what's done is done, so I would quantify the use of the midpoint vs. the other 4 points combined, for starters, so you have a ratio of "non-neutral" [sic] values divided by the total responses, and then do your typical frequencies, grouping the "agrees" and likewise the "disagrees."

For the record, unless I am using a categorical scale, frequency scale, or 0-10 semantic differential, for agreement questions, I am generally using a naturalistic 3-point scale of disagree, somewhat agree, and strongly agree. Anything more abstract is...abstract-- and an artifact of mood, salience, propensity to grade hard vs. easy, etc.

before you do anything you should factor analysis your survey to see if your apriori categories are correct - e.g. questions statistically cluster into (1) Behavioral component; (2) Cognitive component; and (3) Affective component. Then you can meaningfully create scales and then undertake analyses of variance (to see any differences by demographic or differentiating variables) and/or regression analysis to determine any trends and/or key drivers of your target variable (which I assume you have also measured - i.e. Attitudes towards Entrepreneurship and Entrepreneurial Education)

You may conduct descriptive analysis such as mean, median, mode, standard deviation etc. Also hypothesis testing may be conducted according to the study objectives.

Likert scales provide ordinal data - *not* interval nor ratio data. As such, statistics and analyses that require *distances* in the data (e.g., means, standard deviations, analysis of variance) are *not* appropriate.

Each respondents' replies is dependent on *their* differing values, perceptions, opinions. And so, for example, one respondent's "Agree" may be another respondent's "Disagree."

Use histograms (bar charts) to compare the distribution of percentages of each Likert rating for each item.

Multinomial logistic regression. If they are scores multiple regression provided predictors are quantitative.

median instead of mean with interquartile, Ordinal regression.

Likert scale can be use as a summated rating scale. Apart from that you can apply EFA and CFA. Because summated rating scale be treated as a continuous data (but the individual statements are ordinal one).