MaxMin Divide And Conquer algorithm that is O(N) worst case

This algorithm has a T(N) of O(N^2), worst case. When its to the power of two, it's O(N). However I need to find a way for it to be O(N) in worst case. Any thoughts?

max_min(A, i, j, max, min)
{ 
  if(i==j)  -----------> 0 comparisons  
 {
     max=min=A[i];  
   return (max,min);  
  } 
   if(i=j-1)  
  { 
      if(A[i]>A[j])   -------------> 1 comparisons  
     { 
        max=A[i];  
        min=A[j]; 
      }  
     else   
    { 
        max=A[j];   
        min=A[i];
    }
  }
  else 
 {  
   mid=(i+j)/2; ----> O(1)   
  (max1,min1)=max_min(A, i, mid, max, min);  --->T(n/2)  
   // if T(n) is complexity for n elements
  (max2,min2)=max_min(A, mid+1, j, max. min); ---->T(n/2)     
    if(max1>max2)  ---> 1 comparisons   
      max=max1;
    else  
      max=max2; 
    if(min1>min2)  ---> 2 comparisons  
      min=min2;   
    else  
     min=min1; 
  }  
  return(max,min);
}