do
{
  t=1;
  cout << "\nEnter the encryption key (1-25): ";
  cin >> k;
  if ((k<1)||(k>25))
  {
      cout << "\nERROR: Enter a valid number moron!\n";
      t=0;
  }
}
while (!t);

do
{
  t=1;
   std::string input;
  cout << "\nEnter the encryption key (1-25): ";
  getline(cin,input);
  k = atol(input.c_str());
  if ((k<1)||(k>25))
  {
      cout << "\nERROR: Enter a valid number moron!\n";
      t=0;
  }
}
while (!t);

1)declare variable
   int k = 0; 
2)attempt input 
   std::cin > k;
3)check the state variable of the stream
   if(std::cin.fail())
4)Do something if in failed state
   {
5)Reassign a default value to k---say user entered 5Q instead of 51. 
cin would assign 5 to the value of k, which would be desired.
cin would then try to put Q in k which would cause a failed state, but wouldn't remove the 5 from k (so k would have an incorrect input, 5 vs 51, eventhough it is a valid int) and would leave the invalid input in the input stream .
      k = 0;
6)clear the state bit (aka fail flag) of the stream
    std::cin.clear();
7)clear the input stream buffer of all unwanted information.  If you want to use this format and don't know what the following means, ask.  
    std::cin.ignore(std::<numeric_limits<streamsize>::max());
8)notify user of invalid input
    cout << "invalid input" << endl;
9)move on to whatever's next
    }

#include <iostream>
#include <limits>
using namespace std;

int main(int argc, char* argv[])
{
	int a;
	cout << "enter something ...";
	cin >> a;
   std::cin.ignore(std::numeric_limits<streamsize>::max());
   cout << "\nenter something else";
   cin >> a;
   cout << "a = " << a << endl;

	return 0;
}