Conducting Experiments
Over the last week, I’d been thinking about what makes a good experiment and if there are consistent conditions that should be met in order to conduct a good experiment. I feel that I have come up with these conditions and I wanted to jot them down so that any time I needed to refer to them, I could.
Keep in mind that these conditions are really for bigger experiments, i.e. massive life changes. These conditions aren’t really necessary for deciding breakfast.
Also, there is the possibility that this might be updated in the future.
With that out of the way, let’s begin:
What is the root cause of the problem?
When conducting a large experiment like this, most of us are probably doing this in order to solve some problem. However, most of us are terrible at identifying what that problem is. For example, the person who wants to buy a house, but really has a bad roommate and can solve their problem by moving in with someone much nicer, or the person who wants to become an entrepreneur, but really, they just have a boss who micromanages them and can do well with finding another job elsewhere. The examples here are endless, but without being able to identify what the root cause of the problem is, you run the risk of doing a bunch of stuff that won’t get you anywhere.
What is the experiment that I’m looking to conduct to solve the root cause of the problem?
This is simply identifying what the experiment you’re looking to conduct is.
What is the hypothesis for the experiment?
By conducting this experiment, what do you theorise will happen? Keep in mind that this is a conjecture. It isn’t an expectation. It’s valuable to make this distinction because when you expect a particular outcome, you devalue yourself into that outcome and can really run the risk of hurting yourself. It’s much wiser to take this as a conjecture, where you theorise that certain events might happen, but really, you adopt the approach that you have no idea what will happen, virtually conducting the experiment like an explorer explores.
Also, another way to look at this, is that being wrong can be extremely valuable. When your theories are correct all the time, you cannot be learning anything new and, in turn, cannot be growing. It’s only when our conjectures are wrong and we can accept that they are wrong that we are able to grow.
What are the specific objectives?
Objectives are required to give yourself a target and a direction to move in. Ideally, there is an ultimate objective that you’re working towards and all the other objectives are simply goals that you’re looking to accomplish along the way to your ultimate objective.
And they should be specific, because if they’re not, then how can you measure progress towards your target? Wanting more money can mean anything. $1? There you go. But wanting $1,000,000 is specific and measurable.
What measurements are being used?
The measurement(s) should align with whatever your specific ultimate objective is, i.e. if you’re objective is to get $1,000,000, your measurement would be how much money you are making (or losing).
Does the following experiment solve the root cause of the problem?
Double check the strength of the experiment. If going through the experiment doesn’t solve the root cause of the problem, then don’t do it.
To what degree is it possible that you might not solve the root problem if all the specific objectives are met?
This one is a little odd, but there are some situations when by conducting the experiment and meeting the ultimate objective, there is still a possibility that you might not actually solve the root problem. This question is more to think about those possibilities. Of course, you cannot actually know to what degree it is possible that the worst happens, but in thinking it through beforehand, you can at least prepare for the inevitable worse or decide to not conduct the experiment at all and potentially save yourself some time.
An example of this is that you have a hypothesis that completing a degree in a specific field will land you a job and, in turn, earn you money. But completing a degree doesn’t guarantee that you will earn a job. In this case, there is a possibility that meeting all your specific objectives won’t solve the root problem, which is earning money.
On the flip side of this, if your ultimate objective is to earn $1,000,000 and the root problem is that you don’t have $1,000,000, i.e. you don’t have enough money, then there is no possibility of not solving the root problem, so long as you meet the ultimate objective.
Is the experiment falsifiable?
I.e. can your hypothesis be wrong? This one is somewhat tied in with having a specific objective, because a specific objective by its very implication is falsifiable, however, this one is more of a reminder to not have your objectives measured by opinions. When a hypothesis is based on what someone says is right or wrong, then there’s no way to gauge whether or not that person is actually correct and, in turn, highlights that the experiment isn’t falsifiable. Since they cannot be proven wrong, you can really run the risk of going nowhere.
Who are impacted by the experiment?
This one is really just a reminder to be mindful of who is impacted by this decision you are about to make, to what degree they are being impacted, and to let them know before going through with your experiment what you plan to do and what your measurements and specific objectives are.
This can also be a good way to hold yourself accountable. By having friends and family, for instance, constantly asking you about your progress, you can, somewhat, force yourself to have to keep going with your experiment so that you don’t let them down.
Cost/Benefit of the experiment?
Conduct a const benefit analysis of what it is you are doing, i.e. determine what you’ll be putting in against what you will be getting out.
Let’s also break this down into money, energy and time.
How much money are you doing to pay in order to conduct the experiment? And how much money are you potentially going to get back?
Do you hypothesise that the experiment will be energy consuming?
And with time, how long will the experiment take? And how much time during the day will the experiment take?
What’s also wise to consider with the cost/benefit of the experiment is whether or not there are caps to the cost and/or the benefit. I.e. if you were using your own money to conduct the experiment, then you have capped the amount you can lose on the experiment, but if you had to borrow the money and go into debt, then you open up the possibility of losing everything. Conversely, if you move into a dying industry and there was only so much money you could make, then the benefit would be capped, whereas, if you move into a growing industry, then the benefit would be wide open. As you can possibly guess, the ideal experiment to conduct is one where you can cap the cost, i.e. the downside, and open up the benefit, i.e. the upside.
Watching…
Tropical Malady (2004) - Dir Apichatpong Weerasethakul
The Terrorisers (1986) - Dir Edward Yang
Quote of the week…
“Don’t be a jack-of-all-trades, master of none. Be a jack-of-all-trades, master of a few” - me
Support Weekly Reflections by sharing this post and/or by buying me a coffee via: ewenmunro.com/coffee
Read my book The Flow of All Things via: ewenmunro.com/the-flow-of-all-things
Buy HTML/CSS/JS responsive web template to build your own website with via: ewenmunro.com/dev-portfolio (under Templates)
Visit My Store via: society6.com/ewenmunro/all
FOLLOW ME
YouTube: youtube.com/c/ewenmunro
Instagram: instagram.com/officialewenmunro