Thread: One or 2 Tail test hypothesis (cobb douglas prod. function)

1. One or 2 Tail test hypothesis (cobb douglas prod. function)

Considering the Cobb Douglas production function as the initial function to start with.

We regress that to Ln y = ln a + ln b1 k + ln b2 L
Now we state in the hypothesis that ( h0; ) b1=b2=b3 = 0 is equal to zero
And the alternative hypothesis is that it's not equal to zero.

Considering the ' not equal to zero ' u might expect a two tail test because it can be either positive or negative. But since were' talking about Capital and Labor, we expect those 2 variables to have a positive! effect on output.

Can we therefore say that it's better to use a 1 - tail test?

2. Originally Posted by Massachusetts Cowboys
Considering the Cobb Douglas production function as the initial function to start with.

We regress that to Ln y = ln a + ln b1 k + ln b2 L
Now we state in the hypothesis that ( h0; ) b1=b2=b3 = 0 is equal to zero
And the alternative hypothesis is that it's not equal to zero.

Considering the ' not equal to zero ' u might expect a two tail test because it can be either positive or negative. But since were' talking about Capital and Labor, we expect those 2 variables to have a positive! effect on output.

Can we therefore say that it's better to use a 1 - tail test?
Negative Capital is rare occurences, but it happens sometimes, Goldman Sachs is one example.
Negative Labor is an indication that the company needs to hire more labor to catch up with intensive production.

There is nothing wrong in using two-tailed test in this case. If you prefer not to use two-tailed test, you must modify your null hypothesis. Hypotheses can always be tailored to you personal preference. You may set it up to accept or to reject or to postpone decisions.