Forewords‎ > ‎Reviews‎ > ‎

Hardened Concrete Sampling and Population Distribution

posted Sep 9, 2018, 7:42 AM by jeffery jim
Try to find outlier in this graph

Yesterday, i was talking about hardened concrete sampling and distribution. Hours ago, i got an opportunity to engage in a discourse talking about the importance of literature before deciding on the fundamental of any research; samples' confidence interval and the determination of either sample is categorized as parametric or non-parametric. The other big chunk of that discourse session is about the horrendous framework and the mixed up between IV and DV in social science field.

To ease understanding on concrete design and issues, many researchers ended up with a simple normal distribution population when it comes to describe hardened concrete. I beg to differ and i reckon that hardened concrete has never been appropriate to be tested, validated and justified with parametric testings.

Justifications are based on the following arguments and outliers; 
1. Concrete are batched but cast in-situ and exposure to substandard practices, 
2. Directional exposures to chemical and physical degradation and delaminations,
3. Exposure to function change and subsequently mechanical loading, and
4. Deterioration such as crack and etc. are in diffused manner.

With that being said, i recommend scholars to reconsider when they make assertions from journals and specifications. If anyone or any engineer would consider hardened concrete specimens as standard distribution, and not lognormal or weibull distribution for hardened concrete; he is in dire need to get a refresher course.

Remember, literature is the nexus and take precedence above all methods of analysis. If you are fundamentally wrong, forget about piloting your research design, discriminant validity, reliability test or whatever intended to accept or disprove null or alternative hypotheses.

If i am one of the examiners during a viva voce, the first nail i will strike is to check the suitability and validity of your methodologies before showing interest in your result.

Comments