Recently we posted a blog about valid in research. Since then, we received many requests to explain, in simple terms, the related concept of reliability. So, here goes ….
Picture the scene. You leave your house in the morning and step into your brand-new BMW car. That unique, ‘new car’ smell surrounds you – you can almost taste the freshness. The sound system immediately starts-up, playing light, classical music that is a pleasure to your ears. You sink into the Alcantara covered seats, enjoying the support and comfort that brings. Your eyes are drawn to the soft-glow of the media-screen which politely requests your input. It asks “Destination: Work?”. You press “Accept”. It blinks: “Ready to start?”. You press “Yes”.
And nothing happens.
No engine sounds. No lights. No seat-belt warning sound. Nothing.
Welcome to the world of unreliability.
When it comes to things like cars, computers, TVs, microwave ovens in our lives, we have come to expect a very high level of what is called ‘engineering reliability. This is the ability of a system, or component, to perform its required functions under stated conditions for a specified time. Yet, when it comes to research, many of us pay scant attention to reliability.
Research is only credible if the data you collect, collate and analyse is reliable. In research, reliability refers to the consistency of your findings based on how you went about gathering it. In other words, if your study is repeated at various times or by a variety of researchers, using equivalent methods, techniques and procedures, perhaps in a different setting, it should yield the same, or similar results. In a car, when you turn the key, you expect it to start. In research, when you use the same method under the same conditions, you should expect and get the same results.
What, then, is the key to achieving reliability?
If you are carrying out quantitative research, your data collection, collation and analysis approach should be designed so that it will yield the same results on different occasions or in different settings; and other researchers using the same process should reach similar observations. In other words, you must design for reliability by ensuring that no extraneous variables exist or enter the situation.
For qualitative research, your data collection, collation and analysis process should be conducted in such a way that you will be able to verify the accuracy, credibility, dependability and trustworthiness of the raw data. It should also provide transparency regarding the interpretation, analysis and transferability of that data.
While in most cases, it is not possible to ensure the high levels of reliability we have in engineering, we should strive to make our research as reliable as possible. Effective research design can help us to achieve this.