Likert (or similar other scales, like 'semantic differential 7-point scale) scale is usually for measuring qualitative variable. We try to quantify qualitative constructs, for further quantitative analysis. "perception', Serv-Qual, attitude measurements are common examples, where these scales are of common use.

An individual item on a five-point Likert scale would be ordinal or ranked data for which Spearman rho and the Mann-Whitney U test are suitable statistical analyses to apply. If you have multiple items with a Likert scale that comprise a summative value, the summative values are typically treated as interval data for more robust statistical tests. Either way, the data can be analyzed as quantitative rather than qualitative data.

Thanks for your reply Dr. Peter. In both cases you explained we can say that we are using a quantitative approach or method in the study. 

My question, however, is what if you use likert scale to report frequencies only, would I still call that a quantitative method.

Yes, I would, but the frequencies alone would be largely without meaning. Frequencies associated with a sample, however large, are not generalizable to the population of interest with the demonstration of statistical significance. That requires an inferential, rather than merely descriptive, statistical test.

Yes.It is used for quantitative variables.  A Likert item is simply a statement that the respondent is asked to evaluate by giving it a quantitative value on any kind of subjective or objective dimension. 

Yes, the total score for a Likert scale is considered at a high-level of measurement. Assuming the scale is valid and reliable. Thus, it can be used in inferential statistics as a DV.

Likert scales are majorly used to measure perceptions of respondents with refererence to the construct under measurement. We operationalize the construct using secondary data followed by interviews / focus groups etc which are qualitative in nature. To measure how respondents perceive these constructs a likert scale is used. A simple read https://www.simplypsychology.org/likert-scale.html

Just to get your basics in place

Good luck for your research

Likert scale is basically a quantitative approach. Whether you are using it to report only frequencies or you are doing more rigorous statistical analysis with it is still quantitative. 

Good luck my Firend

Yea its a quantitative approach but quantitative analytics can again be perceived as contextual qualitative. 

This is a good question, a perennial one, and one that always provokes misunderstanding and disagreement. Your data are qualitative, and nothing you do with the data changes them from being qualitative. While a "5" is a number, it may signify "highly satisfied" which is a qualitative measure. Thus while it's not only possible, but fun, to analyze 5s as though they are the data, they're not; they merely represent the data. The trick here, as always is to be very clear in reporting your findings to explain exactly what you did to analyze qualitative data using quantitative methods. That is, be clear and put some burden on readers, users, or clients to read what you've written. 

Likert scale items measure opinions and perceptions of participants on latent variables under investigation. Whether such a scale is in 5 or 7 point, it is a quantitative measurement. In spite of this nature of Likert scale instruments, they can be analyzed both qualitatively and quantitatively. It is qualitative when frequencies and percentage analyses are done. The analysis becomes quantitative when inferential and robust statistical tools such as ANOVA, MANOVA, and COVAS, Regressions, Path Analysis, SEM and likes are applied.      

One can certainly distinguish between data that are in numeric form (quantitative) and those that provide descriptions without the use of numbers (qualitative).  The problem is that this simple distinction ignores all the variations that exist both with different levels of measurement and the type of analysis that is conducted.  Thus quantitative data can be subjected to a range of quantitative analysis ranging from the most elementary to the highly complex.  It could also be described in qualitative formats, although it rarely is.  By the same token, qualitative data can be summarized with some quantitative techniques although this largely ignores the strengths of qualitative research.

Also Likert scales can be used to predict a dependent variable (risk, sensitivity, specificity, AUC values, etc.)

Qualitative data obtained from subjective opinions has to come from a normal sample, then, the accuracy of the conclusions will depend on the quality of the questions and the meaning of the questions. If you want to know about the bias, do an exploratory factorial experiment (better if you have an independent variable), and the software will report the variance percent that you have covered of the dependent variable. Once your bias is controlled (through factorial analysis exclude supernumerary variables), transform data into categorica/ordinal using percentil scale. I hope it helps you.

It is quantitative but requires non-parametric statistics as opposed to parametric because the data are most likely not normally distributed. Using parametric statistics in this case will be misleading.

I agree with James Riddle's description and I find it misleading when students describe their 'quantitative' research on peoples opinions - which is simply wrong in my view. My answer to students is that quantitative data are gathered on an independent scale. When opinions are gathered on a Likert Scale we can apply some quantitative methods in our analysis, but the study is essentially a qualitative one because peoples' opinions are subjective.

Likert scales are attitudinal and because of that they are qualitative in nature and as such they are non-metric so we have to use the appropriate statistical tools using nonparametrics. median for central tendency and interquartile range for variation.

It remains qualitative in nature due to its nature of gauging opinions and perceptions on questions asked. However, you can apply non parametric methods of analysis to arrive at conclusions

I must review this again

I cannot agree more with most of the views by many people above, particularly Edward Wilson Ansel & Bridget Kane's, as numbers (0, 1, 2, 3, etc.) used in Likert Scales are only representing some qualitative aspect , about a given topic e.g. knowledge about learning method.

This paper should provide some clarity:

Joshi, A., Kale, S., Chandel, S., & Pal, D. K. (2015). Likert scale: Explored and explained. Current Journal of Applied Science and Technology, 396-403.

I agree. Likert scale derived data is merely qualitative. It is just a measure/scale and mentioning 1, 2 or 3 does not imply quantitative analysis

the scale or measure you use does not change the nature of your data.