IF loop within a WHILE loop and cumulating info into a vector

[John Macuin]

Hi, there seems to be a problem with the while loop below, although there are no specific error messages that lead me to understand the problem. I am estimating beta with Ordinary Least Square (Betachap) and then I am doing 3 t-tests (10%, 5%, and 1%). In every iteration inside the loop, if the t-statistic is greater than a level, I will reject the null hypothesis. I want to accumulate the rejections (rejections equal 1) in the (1000 x 1) column vectors, which is why you see the code: rejetHo10[i,1] == 1. The result is that not a single hypothesis is rejected. I wonder if the loop is not working. Does anyone have an idea of my errors in the while loop code? Thanks.

clear
set obs 10
gen y2 = .
gen x2 = .
gen betachap = .
mata
n = 10
alfa = 0.5
beta = 0
mu_x = 5
sigma2_x = 4
sigma2_e = 2
//**************************
//*** GENERATE MODEL ***
//**************************
cst = J(n,1,1)
x = J(n,1,mu_x) + sqrt(sigma2_x) * invnormal(runiform(n,1))
e = sqrt(sigma2_e) * invnormal(runiform(n,1))
y = cst*alfa + x*beta + e
//**********************************
//*** ESTIMATE 1000 FOIS ***
//**********************************
X = (cst,x)
i = 1
n = 1000
betachap = J(n, 1, 1)
rejetHo10 = J(n, 1, 0)
rejetHo5 = J(n, 1, 0)
rejetHo1 = J(n, 1, 0)
while (i <= n) {
E_MCO = luinv(X'X) * X'y
betachap[i,1] = E_MCO[2,1]
k = cols(X)
error = y - X * E_MCO
s2 = (error' * error) / (n-k)
var_E_MCO = s2*luinv(X'X)
std_E_MCO = sqrt(var_E_MCO)
t = E_MCO[2,1]/std_E_MCO[2,2]

if (t>1.8331){
rejetHo10[i,1] == 1
}

if (t>2.2622){
rejetHo5[i,1] == 1
}
if (t>3.2498){
rejetHo1[i,1] == 1
}

i = i + 1
}

vecteurHo = J(1,n,1)
rejetsHo10 = vecteurHo*rejetHo10
rejetsHo10
rejetsHo5 = vecteurHo*rejetHo5
rejetsHo5
rejetsHo1 = vecteurHo*rejetHo1
rejetsHo1
end

[Sergio Correia]

rejetHo5[i,1] == 1

Should be

rejetHo5[i,1] = 1

Because "==" is checking equality and "=" is for assigning values


Cheers,
S