Data analysis is to verify our results whether it is valid, reproducible and unquestionable. Even the industry do not interested in significance or insignificance of our data, we as a scientist must do data analysis. You can simply report 'passed' or 'failed' to industry people but behind that, there is data analysis performed (which is not necessarily to be reported).

In research work, we often experience cluster of results and dispersion in the measurement of single variable. It is generally easy for readers to understand in short time how the experimental results/ values are changing, and if these are changing what is amount or percentage of variation etc. Also inter- dependance of factors and comparison of results is possible.

I think it is simply a matter of language misunderstanding and misconcept about the meaning of a test. To make a decision ("they want to know is whether their samples have passed or failed a test") everybody (scientist or not) must 'analyse' the data. This, at least, because there are several methods to test the data, so you have to decide which one looks more appropriate. In addition, in examining the samples with a given test procedure, the data you get depend on many factors. Finally, to make a decision it is not sufficient to insert your data in a 'black box', push a button and get off 'the' result. Data analysis is not only a matter for 'scientists'.

Data analysis helps to make sense of our data otherwise they will remain a pile of unwieldy information; perhaps a pile of figures. Data analysis differentiates the scientist from the general population as it guides him from making largely unsubstantiated claims or remarks resulting perhaps from chance occurrence. It is not simply for the sake of publication. Also, it is not always about statistical significance as a distinction is often made between statistical significance and practical/clinical significance. So, it is not always about statistical significance or nothing. After all, anything can achieve statistical significance if the sample size is large enough.

From a qualitative point of view, data analysis provides a basis on which to make sense of the data and to interpret it in ways that are relevant to the research purpose and research questions. It provides the building blocks from which the researcher can construct and substantiate an argument about the findings. .

Though the question is now quite old, it remains important to be crystal clear that the data are NOT "facts" in the sense that most industry people (but also many others) intend them. The data on clima are a very good example. Data can be affected by so many reasons of failure or insignificance, that ONLY a data analysis by competent people can validate it.

Should be that unnecessary, why the users of the data should need a certification by recognised organisations, certifying their meaning? They could have an instrument in their office, and use the numbers that come printed by their machine: if they are within a certain range, they could tell that the test has passed.

They need to take decisions, irrespective to the risk of false evidence, and will never understand that science (and related techniques) never provides certainty, without for that becoming useless.

Competence is a must in any field, to ensure an acceptable level of confidence in the statements provided. That applies also to the need of an analysis of the data originated by even the smartest instrument.