I have a ranked data where the respondents have ranked a set of 7 variables ( 1 to 7 ). what are the techniques to ensure the reliability, validity of such ranked dataset in general if any?. what are the further procedure to move ahead for further analysis?
Then I have a Likert scale ( 5 ) based data where 6 or 7 questionnaires are asked to the respondents. What would be standard procedure for these kind of data?
Are there any books that compile all these techniques, strategies and criteria for research purposes?
The ranked data can be analzed using The Kruskal–Wallis test by ranks, which is a non-parametric method. It is equivalent to one-way ANOVA for parametric data. Likert scale data can be converted to interval data as: strongly agree = 5, agree = 4, neutral = 3, disagree = 2 and strongly disagree = 1, for examples. These likert scale interval data can be analyzed using the Kruskal–Wallis test by ranks or sometimes using ANOVA subject to fulfilling the assumptions of ANOVA (normality and homogeneity of variance). Best wishes.
@Vishnu You've not talked about your research questions/objectives at all rather you only described the nature of your data. Selection of any statistical technique (metric or non-metric) purely depends on two things, viz. research questions/objectives and the type of the data. In order to get some specific suggestions, you need to provide the information regarding your research questions/objectives. In the absence of information regarding your research questions/objectives, Mr. @Yadunath has suggested Kruskal-Wallis and ANOVA assuming that your are looking for differences across some demographic categories or any other factors. Wishing you best.
Dear Vishnu, just like Mr. @ Yadunath and @ Imran have explained, I will further say that what your intent is, about analyzing likert data is not cleared since your research questions or objectives are not clearly spelt out, however if you are probably assessing the relationship between the variables measured on a likert scale and another continuous outcome variable, then you can transform your likert data by computation to generate a continuous variable which will enable you to run a linear correlational/regressional analysis. Similarly if your are to see the variability in the level of your computed likert scale variable (now continuous) across demographic variables for instance, then one-way ANOVA will be suitable for that, above all you will need to carry out transformation for your likert data and your objective is what determines the appropriate analysis you would need to embark on. best wishes.