If one has ever taken a course in statistics, or, applied statistics to come to conclude in any research setting, paired t test is an animal which one encounters.

I am not going to write about the intricate details of the paired t test as there is already ton of stuff on it. My intention would be to link the basic concept to startup life…and what should one probably keep in mind before they start their own gigs…

Let me recap paired t test from a 10,000 feet view.

Imagine you are a gas dealer and would want to stock one of the quantities, lets say gas A OR gas B. Obviously you would want to stock one of them based on the buyers perception. If the car driver gets more mileage out of gas B than gas A, the driver would be more interested in buying gas B and hence you must stock gas B. However you do not have the luxury to do a real life test and then figure out your strategy. You need to take a decision NOW whether you want to go with gas B or gas A.

Linkage :Most of the times people assume that market will provide them the right direction for them to start a firm. There will be enough customer reaction from the survey that they will analyze and start a firm..Alas!! Life takes the side of the risk taker.

Coming back to the example I started with, Lets say you decide to take a random sample of 100 drivers toss a coin, pick 50 folks and ask them to use gas A and ask the rest of the 50 folks to use gas B. Well , you collect results and you see that there is a difference in mileage of the cars between A and B. However you see that the difference is not consistent. In the words of statistician, the standard deviation of the differences is large.Can you be certain that the difference is good enough that you must go with the gas that shows higher mileage?

Link to startup life :
If you think of 100 separate things and want to find out whether a startup life gives you satisfaction or not, you might find a large standard deviation between happiness from startup life and a regular life. Imagine including stuff like having a dream house , car , spending time with family, going out parties , working on interesting projects , socializing , etc etc..and lets say you decide to start a firm. suddenly you realize that you are not able to do justice to all the things you have on plate…OBVIOUSLY..a startup is a baby, it needs to be nourished with utmost care.. If you continue all the things you were doing, and want the same level of happiness from a startup life, you might not get it.
Is there a better way ?

Going back to the example of t test.
The best way to test the difference in gas A and gas B is to remove the effects of the rest from your study…Standard deviation was aggravated for the simple reason that different drivers had different driving skills. Consider the following situations

  • Aggressive driver who uses gas A when compared to slow and over cautious driver who uses gas B
  • Aggressive driver who uses gas B when compared to slow and over cautious driver who uses gas A

The difference in mileage recorded is going to be rather different because one is comparing applies to apples to oranges. The best way to go about is to ask the same driver use gas A and gas B and then record the difference

Link to startup life
Suspend everything you are doing which gives you incremental happiness, and focus completely on your startup…Like the paired t test where rest of the influences need to be removed from life to get the correct picture, my hypotheses is that you will will see that there is a huge difference in happiness levels when you do your gig only when you suspend or control other random factors in your life..What are they? Well , you know what they are ..They are different for different folks. You need to have the guts to keep the rest of the factors to a minimum level of influence..Can you do it ? If you can’t , don’t blame your gig, because I firmly believe you are better off anyday doing your own thing …Ok, my inference is probably from 1 simulation till date :) ….I will be doing more simulations in my life to check this hypothesis:).

My spin on paired t test is this:
Control the rest of the factors in life, for a startup life is a beautiful life that everyone must experience..To sum up this random post of mine :

Ho= Null Hypothesis - Startups do not give you happiness as you are betting against unfavorable odds
Ha = Alternate Hypothesis - Startup life gives tremendous amount of happiness

Figure out whether Happiness (Ha-Ho) is significant..?

I think you will reject Ho if you control the rest of the factors in life.