Dear mates.
i´d like to know your point of view about things that ever appear in methodology. Yin (2003) mentions that question with *How and Why* should be studied qualitatively. I´d like to know if we raise an issue with HOW and the process is terminated, can it be possible to quantify.
THANKS
QUANTIFYING QUALITATIVE DATA: Qualitative data may be quantified. In fact, they should be quantified. Reporting only in qualitative form is an an incomplete work. The questions that deal with "how" and "why" should be studied qualitatively, not qualitatively. The rationale follows.
HOW?: The question of "how" refers to methodology. The method by which the data was collected, treated, and analyzed will influence the result and conclusion made. faulty methodology (bad how) will lead to bad result, and wrong conclusion. In this respect, the issue of "how" must be governed by the accepted principle of practice in the field. Is this judgmental? No, it is very objective. To say that the "researcher" fails to follow an appropriate procedure or methodology is an objective statement.
WHY?: Both the "how" and "why" are governed by scientific standard of acceptable procedures and reasoning. Both must pass peer review under scrutiny. For instance, a brilliant brain surgeon performs a brain surgery that follows all correct procedures in the field (no problem with how)---but the surgery was conducted to treat arthritis on the patient's ankle---this would fail on the "why" prong.
HOW TO QUANTIFY QUALITATIVE DATA: For the HOW question, there must be a reference answer where the method or procedure is correct. Similarly, for the question of WHY, there must be acceptable bound of reason of why such a procedure is taken. In both cases, after acceptable standard is set, the data could be dichotomized into Yes = 1 and No =0. Hereafter, binomial distribution analysis and probability analysis may be used on both HOW and WHY.
Given a sample size n, count all YES and score it 1 and sum the YES counts as success = s. The probability of success is given by the Laplace Rule of Succession:
(1) p = (s + 1) / (n + 2)
The wrong answer for HOW (similarly wrong reasoning for WHY) is given by:
(2) q = 1 - p
if the sample is large enough and if the data manifest normality characteristic, the test statistic may thus be given as:
(3) Z = [(X / n) -p] / sqrt(pq / n)
X is the target number to be evaluated. Is it statistically significant to have this number of success: X? What is the probability that X will occur could be answered by:
(4) P(X) = [n! / ((n - X)!] pXqn-X
In sum, a qualitative research without quantitative analysis is an incomplete work. By quantifying the data, we can analyze (b y scientific basis, not personal opinion) and able to forecast events not yet happen within a specific degree of confidence interval.
Thank you Dear Paul.
I´d like to comment that we can also lose some info by doing quantitative research
- Badges
- Science method
- Similar topics
- Methodology