Can you suggest how the analysis of a person-centered questionnaire might identify the discrimination value of each question?

Part I:
The concept of levels of measurement is not as difficult to understand as may seem. There are times when you want to measure what the relationship between two variables is when one is a category. Political affiliation can be a category of interest with Republican, Democrat, and Independent as possible classifications. The researcher might be interested in how political affiliation relates to views on taxes. Another level is used for when we want to rank order characteristics or responses. A standard comparison is the rank order of a group of employees based on a paper-and-pencil test compared to the rank order of the same employees based on a supervisor’s ratings. The third level involves a continuous measure where there is no zero point. The measure of intelligence is a good example. By definition, there is no such thing as zero intelligence. However, at times, my brother has come close.
Although survey items with a choice on a scale (such as from 1 to 7) for a response are used as continuous measures, they can be used for any of the four levels of measurement scale. Which do you think is the most appropriate and why?

1. Cite any sources in APA format.
2. Support reasoning with research and examples.

Part II:
Measurements Scales
Nominal Scale: It is a type of measurement where numbers are used as labels but with no quantitative value (Garger, 2010). Data is normally grouped into categories then assigned certain numbers as name tags.  For example, individuals may be classified as male and female where 1 may be used to signify male while 2 signifies female. The numbers 1 and 2, in this case, have no inherent value with respect to magnitude but are used merely as names. Parametric data analyses group means while nonparametric data analyses group median. Therefore, after data is subjected to nominal scale measurement, it becomes easy to compute its mean and median.
Ordinal Scale: This is a measurement scale where data is first grouped into categories, then assigned different numbers which have different meanings but have characteristics of either greater or lesser than (Garger, 2010). Variables are measured in terms of magnitude or ranks. For example, a researcher might ask doctors to rate five types of malaria drugs in order of preference using a scale of one to five. Where one is the lowest score while five is the highest scale. In this example, 2 will be less than three but greater than one.  Therefore since parametric and nonparametric methods rely on a ranking of observations, an ordinal scale is of great benefit to the two (parametric and nonparametric methods).
Interval Scale is a type of measurement where the difference between two values is meaningful. For example, the difference between 60 and 70 is the same as that between 40 and 50 centimeters respectively (Garger, 2010). The groups have the same variability, therefore; parametric and nonparametric analyses can provide reliable results.
A ratio scale is a type of measurement where measurements are stated rationally at zero. For example, in Kelvin temperature, zero means that there is no heat existence at that level (Garger, 2010).  Both parametric and nonparametric operations are possible since they facilitate mathematical operations.
Item Analysis
Item analysis is the process of analyzing respondent’s responses towards a particular test item to assess the quality of the test and its details. During test construction, item analysis is used to examine whether the test items are measuring what they are supposed to regulate. This is achieved through reviewing the respondent’s responses towards each test item (Rivera, 2007). Items analysis for personality tests is time-saving while that for knowledge-based tests is time-consuming. For example in a case where one is researching why most motor accidents happen during festive seasons, it will be easier to test if a scale of 1 to 5 is valid. Moreover, when examining the level understanding of a mathematical concept, it will be difficult to verify the test items since difference individuals perceive ideas differently.
References
Garger, J. (2010). 4 Levels of Measurement in Social Science Research. John Garger.
Rivera, J. E. (2007). Test Item Construction And Validation. Developing Statewide Assessment For Agricultural Science Education, 30-31.
Can you suggest how the analysis of a person-centered questionnaire might identify the discrimination value of each question?

1. Cite any sources in APA format.
2. Support reasoning with research and examples.