If you just have 3 scale points, it is trivial. If there are more then it depends why you are doing because for many purposes you probably should not. So, why do you want to?

thank you very much for the answers, what i am trying is to create is a literacy index based on the answers i collect form people on likert scale, what statistical method should i use. e.g. i have 10 questions in a questionnaire all liker scale, i have responses from 30 people. everyone has different answer. now i need to validate the questionnaire and also show the scores and decide what is low, medium and expert awareness groups. i also want you to help me how will i decide that at what level its low, at what level its medium and experts etc. what is the method please help me to identify. i am using SPSS.

Start with Cronbach’s Alpha criterion to validate the data

https://stats.idre.ucla.edu/spss/faq/what-does-cronbachs-alpha-mean/

Hi, thanks for your answer, i want to show something like this after all calculations, please have a look at the attached picture. help me to conclude the raw data into literacy index.

It depends what you mean by validity, and with n=30 that will be tough. Common method is if you developed the scale with a certain structure in mind (say 3 groups of items, which will all be correlated) then you might be confirmatory factor analysis of this and compare the fit with other plausible models. That said with just 30 this will be tricky. What is the purpose of you validation?

the purpose of validation is that my scale is able to measure one's environmental awareness overall based on his/her Attitude, Knowledge, Skills and Involvement related environment. after that i want to show groups based on their low, medium and high environmental literacy. my question is how do i reach to the decision of for instance 1-4 is low, 4.1-8 is medium and 8.1-10 high groups of environmental literacy, how do i decide these ranges, how do i reach to these numbers? n=could be anything 30 or more. i am not sure if i can clear you what i am saying :(.. also di you see the picture i posted above, have a look may be you get a bit more clear idea. again i am so thankful that you are willing to help me .

I think what you are looking for is methods for "cut-scores", so you can google that. These occur in education where it is required to have a single value over-which to say someone has passed. This makes sense for a driving test, where you either give someone a licence or not, but it has obvious psychometric problems for other purposes because everyone below the pass-rate is counted the same. For this reason threshold tests should be constructed with items providing most information near that threshold (so for the driving not really hard or really easy items). When three groups are necessary (in the US some regulations have asked for % proficient and % near proficient [they sometimes label these differently]). This tends to be just for reporting, but for any research purposes because why lose the information in the scale to categorization (see all the arguments against dichotomizing). This is often done with using subject matter experts (SMEs) or prediction (e.g., the value of X on a high school test that predicts getting a B in college a year later, here is an example: Article Regression methods for estimating cut scores

thanks for the detailed answer let me have a closer look at this.

It is difficult to do what you are trying to do. What qualitative difference is there between the high point of low literacy and the low point of medium literacy? You run the risk artificial dichotomization (trichotomization in this case). Doing so will cause you to lose statistical power.

what about i had two days on researching this. i spoke to the Dr. who created this literacy index. he told me that it is his assumption only where he broke down the scores. i watched videos on youtube it suggest that i can use frequency distribution and set a range of the given data and analyse it where the score falls high, low or medium based on mean scores. and rest is validation of the items using cronbach alpha and further regression analysis etc. do you think it is a justifiable way to calculate scores? Ajit kumar Roy

Hi Sultana Ashiq . There is a large literature that describes problems using the mean or median to categorize responses. A good source is http://www.psychology.sunysb.edu/attachment/measures/content/maccallum_on_dichotomizing.pdf . The big question is why do you want to categorize the variable. Sometimes you want to do this, if for example you have different interventions in a study for each tertile, but if this is at the final analysis stage there are fewer reasons for the categorization. If you think the variable (or variables) have an underlying construct that is for different groups (or taxa) rather than dimensions, there are several procedures (google taxometric analysis) to explore this.

Hi Daniel, thanks a lot for taking interest in my work. I am trying to show that different level of awarness has different impacts on other variables. I am using awareness as a demographic moderator in the model for instance difference in age shows different usage results, similarly i would like to collect awareness data using questionnaire scale, calculate the scores and separate them based on their high, medium and low scores. I want to show difference of awareness level have different impacts on usage. Please guide.

Hi Sultana, depending on on how awareness is distributed, it might be easiest to look at the interaction of awareness and the demographic moderator. Say the awareness variable is A and let's say gender (G) is the demographic, you are interested in their "effects" on Y, then

Y = beta0 + beta1 A + beta2 G + beta3 A*G + e

would be a sensible model (depending on if the relationship looks linear). In R (which is the package I am most familiar with) it would be:

lm(Y~A*G)

and if you wanted to test if the association of A and Y depended on G, you could test the interaction:

anova(lm(Y~A+G),lm(Y~A*G))

R is fine is either A or G is defined as categorical, but if Y is categorical a different procedure would be used. pm me if you want more details.

Hi Sultana,

This is so simple! Please find my answer in the steps stated below:

Step #1:

Compute the overall mean score for the items of that variable of interest

Step #2:

Make cutoff points using a calculator (Maximum- Minimum / n). How? In case of a five-point Likert scale, you may, for example, have assigned the value (1) for Strongly Disagree, (2) Slightly Disagree, (3) Neutral, (4) Slightly Agree, (5) Strongly Agree. So in order to make cutoff points, you've to do this simple math, as per the formula stated above: (5-1/3). Upon calculation you will get this value (1.33), right! This is the interval value.

* "Highest" refers to the highest score of the given Likert scale (5, in our example)

* "Lowest" refers to the lowest (1)

* "n" refers to the number of CATEGORIES you intend to create

Step #3:

Do the math for the three categories (Low, Mid & High). How? Just add up the score of your interval value to the three category, as in the following:

Low (1 - 2.339),

Mid (2.34 - 3.669)

High (3.67 - 5)

Step #4:

If you are using SPSS, enter these category-related values by navigating through "Transform", and "Recode into Different Variable". Here you will need to recode the Mean score we've obtained in Step# 1. After you give it a new Name and Label, click on "Old and New Values". A new window will pop up. Select "Range", and then insert the values of the categories given above. You should give a value for each range (i.e. 1 for Range no 1), in the New Value empty field. Then click "Add". Follow the same routine for the other range values of your categories, 2 & 3. Click "Continue", and then "Ok".

Step #5:

Go to the Data Editor page and scroll down to find your new created variable. You'll need to give it a Value Label. Please add in the Value field number 1, and in the Label field "Low", and thus 2 for (Mid), 3 for (High). Click "Add" each time you make a new try, and then "Ok".

Here we should be done, and you should find my answer to your question if you run your analyses , Sultana Ashiq . If not, please let me know to try further.

Cheers,

Ahmed

Hi Ahmed, that is very useful answer. i have done the same way but for the categories i used split file function, i split it into three groups and just noted the value where it splits on each stage. does that sound right?

While exactly not sure about the data you are mentioning and seeing is more believing, still I interpret this to be ORDINAL DATA as you mentioned:

Low

High

Moderate

So i believe depending upon your objective and using SPSS as it is more commonly in use.....u should utilize chi square test for these three categories for seeing if there is any significant difference the three categories.

Rest i feel if you mention your objective the situation can become more clear

Regards

Hi,

The Split Files is completely a different procedure. It is used when one sometimes wants to repeat analyses for groups (e.g., males and females) separately. So, it is not meant for making categories as in your question.

As given in my previous response, you should make ranges for your categories ---Here I attach to you some screenshots to facilitate understanding.

Ahmed

PS In my screenshot *3.7 thru 5 should be 3.67 thru 5 instead.