Levels of measurement also known as the scales of measure are the relationship between the values assigned to attributes for a variable. They are classification describing information’s nature within numbers that are assigned to variables. There are four levels of measurement, namely: interval, nominal, ratio and ordinal.
This is a level of measurement which differentiates between subjects or items based on their names together with other qualitative classifications that they belong to. Therefore, dichotomous data is involved with classification of items as well as construction of classifications. Variables may be represented by variables though they such numbers lack numerical relationship or value.
This type gives room for ranking (1st, 2nd) whereby data can be sorted. However, it does not accept any degree of difference in between. For instance, the dichotomous data has values such as “healthy” vs “sick” when measuring health. Non-dichotomous data has spectrum of values, eg “mostly agree” “completely disagree” when measuring opinion.
It allowsa degree of difference in items with no ratio between them. An example of this level of measurement is the temperature having the Celsius Scale. It has two defined points, (boiling and freezing point of water at certain conditions).
It possesses a meaningful (non-arbitrary and unique) zero value. It takes its name from the idea that measurement is an estimation the difference between unit magnitude and that of continuous quantity.
In social statistics, the application of a mathematical function to every point in a set of data is known as data transformation. It is done in order to meet the procedure’s assumptions. Information is transformed for several purposes, such as compute and record. Not all data can be transformed. Data transformation does not apply in literature citations. Only the numerical data can be transformed.
Validity is the length to which a tool for measuring measures that which it is supposed to measure. For instance, the thermometer is supposed to measure the body temperature and not the body temperature. Therefore, it is not valid. Data validation is ensuring that a program operates on correct, useful and clean data.
The concept of design validity is the length to which a research is sound while measurement validity is the length to which an instrument measures that which it intends to measure. A valid research design uses ways that lead to relatively true findings. A valid design depends on external and internal validity to bring credence results together with the inferences they make.
Three types of validity exist within the idea of measurement validity(Frankfort-Nachmias, C., & Nachmias, D, 2008). They include: content validity, empirical validity and construct validity. Concept validity is the length to which the tool used measures the salient behaviors of the measured subject. Empirical validity shows the extent to which the produced results have similarities to real relationships between variables. Construct validity is the degree to which a tool measures that which it claims to be measuring and not something else(Cronbach, L. J., & Meehl, P.E. , 1955). It is important that a researcher comes up with appropriate information about the above named types of validity in order to validate the measurement.