Ankur:  A Likert scale technically is restricted to agree-disagree judgments, like Strongly Agree, Agree Somewhat...) -- although the term is very frequently used incorrectly.  It sounds like you are referring to what is more generally referred to simply as a rating scale.

Ankur - it depends what 0, 1, 2, 3 are intended to mean. If they are intended to imply something that increases as one moves from 0 to 3, then one could call this an ordinal scale - which is, as I understand, what a Lickert scale in principle also attempts to achieve. I wonder what led to to use a 0-3 scale, and why not a 1-5 or 1-7 scale. 

It is better to follow a 1-5 or 1-7 scale than 0.1.2.3.

Each choice determines a consecutively higher progression in a four phased manner, therefore 0-3. The questions are based on constructs. The increased scores tend to capture an increase in stage  based on determinants that form the four choices. I am confused about the way I can treat the questionnaire.

Ankur:  A Likert scale technically is restricted to agree-disagree judgments, like Strongly Agree, Agree Somewhat...) -- although the term is very frequently used incorrectly.  It sounds like you are referring to what is more generally referred to simply as a rating scale.

Ankur: advantage of a lickert scale is you can then code your results (strongly disagree to strongly agree as mentionned by Gordon) -2 -1 +1 +2 Moreover important to notice if you use a scale with 5 or 7 choices, you have a central axis that you can code zero to let a possibility of neutrality for your respondants

i agree with Smitha because this will help u further in statistical analysis of your data too..

For opinion scales one can use 1, 2, 3 and so forth to show the differences in magnitude. This can be likened to an interval scale. But an ordinal scale just expresses relative positions and not magnitude. To better understand your problem, perhaps one would need to know what constructs you are looking at. 

0 to 3 (ore equivalently 1-4)  is just below the minimum required (five steps) to treat a rating scale (ordinal scale) as an interval scale . With an interval scale you can apply factor analysis and other multivariate analysis legitimately. Otherwise you may consider to treat your measures with categorical factor analyis (Categories from SPSS)  or use ordinal factor analysis (MPLUS does it, among others)

This is a 4-point Likert scale  [my guess ranging from 1 (strongly disagree ) to 4 (strongly agree) ?]

It has to be treated as an ordinal level scale.  From 7-point upward the scale can be treated as an interval scale

It seems like your (0 to 3) scale is a simple ordinal scale. Also, it seems like you wish to measure the construct with a unidirectional scale.

Likert scale is a particular ordinal scale that has specific response categories (strongly disagree, disagree, neither disagree nor agree, agree and strongly agree). If you are using the exact same scale then call it 'Likert scale' otherwise call 'Likert type scale' if you five ordinal response categories.

Ankur, what you describe is not a Likert scale, as others have said, Likert is strictly a set of statements to which someone gives their level of agreement. What you do have is interesting, as presumably 0 means something specific? If someone responds 0 to all q's, then this is distinguished from someone who answers 0 to some and 3 to others. However, what if someone answers 1 to some and 2 to others, and ends up with the same score as the 0 and 3? You may need some way of ensuring they are treated as different response sets. 

Are we stuck with definition problems? Wikipedia's definition:

It's important to distinguish between a Likert scale and a Likert item. The Likert scale is the sum of responses on several Likert items. Because Likert items are often accompanied by a visual analog scale (e.g., a horizontal line, on which a subject indicates his or her response by circling or checking tick-marks), the items are sometimes called scales themselves. This is the source of much confusion.

A Likert item is simply a statement which the respondent is asked to evaluate according to any kind of subjective or objective criteria; generally the level of agreement or disagreement is measured. It is considered symmetric or "balanced" because there are equal numbers of positive and negative positions.[6] Often five ordered response levels are used, although many psychometricians advocate using seven or nine levels

What hinders you from changing 0,1,2,3 to 1,2,3 and 4?

Yes, I might change to 1,2,3,4...it seems more relevant, as it will make the scale more sensitive. My problem is that the choices are not agree, somewhat agree, somewhat disagree and disagree..or on that basis...nor am I asking the respondents to mark on a scale of 1-4 or something based on their agreement. The choices are something like I haven't tried, I will try, I am trying, I have been successful. AND No, Maybe, Yes, to some extent; A lot... I am scoring these as 1,2,3,4..

I think you could consider it a Likert scale. Maybe isn't it strictly, but, on my opinion, is really close to a Likert scale

Ankur,

Your four categories "I haven't tried, I will try, I am trying, I have been successful" that you, it appears, have coded 0-3 (and now considering to recode them 1-4), are, strictly speaking, hard to justify as Lickert scale which is essentially ordinal. What you have are categorical, NOT ordinal (ie Licker scale) data. 

Thanks, Mr. Subhash Chandra. I am having a hard time trying to find the right way to treat my questionnaire for validation. I find it hard to digest that such a scale may be a likert scale, though it closely resembles one. I just don't want to carry out a research based on my own presumptions and prejudices or my lack of knowledge about a particular thing that I am confused about. I want to make sure how exactly to treat this kind of questionnaire for its validation.

hi ankur  - pls check my article on  attitude of mobile users on advisory services  (likert scale )

Likert scale is the agreement or disagreement of particular statements thereby the attitude of the person could be detected. The scoring for four choices '' I will try/ i tried/ i succeed / i lost will not be attitude statements- it will be measure of participation in the activity --more exactly you are rating something. 

Hi Ankur,

If you are having a hard time, construct a table and look at it. Perhaps you can make a dicotomy (I haven't tried, I will try = 1) and ( I am trying, I have been successful) = 2

 use it as an independet variable in a linear regression analysis or use Fisher's test or

compute a logistic regression analysis with  0 =Not trying versus  1=trying.

I would recommend the literature on the partial credit model (a generalization of the Rasch model for polytomous response sets).There is a lot of stuff written by Jürgen Rost. Especially the category characteristic functions of Andrich's rating scale model seem adequate to me while studying your question. 

Kind regards

Ulrich Frick

hi, i think, first of all you have to specify your objective for doing survey, if your objective is just to determine the perception of people about something then it does not matter which scale are you using, you can find out this just by calculating simple frequency of each response.

Waqas, from the viewpoint of a sustantive researcher your statement might earn some support. As far as I understood Ankur's question, the focus is not on a substantive interpretation of some real item set, but on a more general methodological problem. Only if this assumption is correct, the hints on the respective methodological literature may be of interest.

Otherwise, there might also exist situations,  where your plea for "quick and dirty"-inspection of item characteristics could be misleading for substantive interpretation. A famous example is given here:

Schwarz, N., & Hippler, H. J. (1987). What response scales may tell your respondents: Informative functions of response alternatives. In Social information processing and survey methodology (pp. 163-178). Springer New York.

Kind regards

UF

Given that you have already collected the data, I think the most important question at this point is where you can create a scale from the variables that will fact like an interval variable. The easies tool for assessing that assumption is Cronbach's alpha.

There have been so many questions on this topic that I have collected a set of resources to help people learn about it:

https://www.researchgate.net/post/How_can_one_help_people_create_simple_scales_from_Likert-scored_items

I would say, still yes