Data Based Decision Making and Importance of Reliability and Validity of the Research Methods

A number of articles you find these days focus on organisations to use big data, either structured or unstructured. It is important that data collection and has to reliable and valid at all times. A number of times data collection is always lacking reliability because it is not the role of the data collecting team to provide reliable data because they might have pressure to meet targets.

Here is a scenario that is very common during data collection. William is working in a team that must go and collect information from a door to door visitation asking potential customers about our product. The remuneration of this service is based on a number of customers who responded the survey. The details on the survey included physical address, respondent gender, respondent age range, respondent salary range. William has a target of 100 respondents per day and a total of 500 per week. This contract is for a month. William is paid $15 per hour and the agreement is to pay only 8hrs and is making $600 per week. William decided to sit on his desk and search google for addresses using a zip code to populate the survey. The information id provided but not all the required information is present on google. William therefore used his random guesses to complete other missing data. He submitted his weekly surveys and was happy and paid for a month and closed the deal with the company.

The company populate their data warehouse with that data and were happy that they populated their databases for future research. The Data scientist of that company started working on that data to find hidden insights and that data revealed that there are many people who desired their products and made recommendations to the top management to target those potential customers as the data had revealed. The company had outlets in most of the targeted areas where fake data was collected and they supplied their products based on that fake data. The sales results did not reflect the Data Scientist 's recommendations and the top management were not pleased with the sales for the season.

The question is how do we make sure that data collection is reliable and the data collection methods are valid? Should we use paid contractors to collect data or use our own employees? How do we guarantee that the data collection process is done in a reliable manner either by contractors or by our own employees? All these questions will validate the point that data is just data if not properly managed from its source to its analysis and presentation. So having data warehouse and data marts that are not properly managed and guarantee their data accumulation is just as useless data not collected.