(1) A double randomization feature that would instantly select test questions totally randomly from a large database for presentation onto a computer screen. This feature would not only prevent inmates from ever being able to predict the order of the test questions. An electronically administered test would also assure that there would be no tests to steal. As the software system evolved, it was discovered that the more test questions placed into the data bank for a particular test, the more secure and unpredictable testing became for the inmate.

To assure complete unpredictability of test question presentation, a second randomization feature was also developed. This sub-program randomly scrambled the position of test question answers each time a test question was presented. Between these two features, a double randomized test was generated. Between the randomization of test question presentation and the scrambling of their associated answers, the software automatically generated millions of test question permutations that were presented to prisoners in a completely unpredictable fashion. Such double randomization precluded any possibility of test predictability and completely eliminated inmate cheat sheets as a threat to program validity.

This computer controlled double-randomization assured that the same test run simultaneously on two computers side by side would result in each inmate receiving test questions and answers in a different order. Prying eyes looking at a friend's computer screen administering the same exact test would serve no practical purpose. Consequently, copying the answers from a friend's computer, cheat sheets, or any system of signals that inmates would constantly dream up to defeat valid testing for the purpose of cheating their way through program testing would also be defeated. With each test being uniquely generated, any inmate faced with taking a computerized test could expect no help from anyone or anything. His testing ordeal would be a unique and individual matter directly reflecting what he learned back to him very quickly and with brutal honesty. My fear from inception was that inmates who attempted to cheat the computer tested programs would loose their cool and destroy one of the computers. I can honestly say that this never happened. Out of all the thousands of hours these computers operated on different prison units, not a single one of them was destroyed by an inmate in a fit of anger, even after repeated failures.

Under no circumstance would the computer issue an inmate a passing score if he did not demonstrate that he had legitimately learned the program information given him to assimilate. Nor would a computer pay any attention to even the slickest silver-tongued inmate manipulator making a case that something about the test or the study material was unfair and his score should be raised. This kind of inmate harangue was a constant problem with humans who graded paper and pencil tests. Inmates would constantly be making a case that this question or that was easily misinterpreted and they should receive credit, or that all they needed was one more test question to pass and to let them slide. This kind of behavior constantly wore away at program managers nerves and was a major contributing factor in program managers dropping out of programs and applying for other duties.

(2) Since the software generates, administers scores and data banks all tests electronically there are no hard copy tests for prisoners to steal or copy. Since paper is not used, the high paper costs associated with ongoing paper and pencil testing of large inmate programs was totally eliminated. This feature alone would save the cost expended to implement one of these programs within a relatively short time, and go on saving institutions money years afterward. Computer assistance assured that tests were scored immediately upon completion and resultant feedback flashed onto the computer screen immediately upon inmates completing these tests. This was a favorite feature of inmates in general who did not like any kind of delayed gratification. They wanted feedback as to their test results immediately. Giving them such via computer was the closest thing to a video game they would find in prison. Instantaneous feedback eliminated inmates constantly nagging program managers to look up scores they had previously been given but forgotten.

Quick test feedback proved to be a very popular feature with both inmates and program managers. At the completion of any particular test, program managers quickly knew whether or not to give the next study booklet in the program to an inmate or to hand him back the one he had just failed a test over to re-study.

(3) To control inmates who did not study and then attempted to guess their way through testing, while others, more fully prepared waited to get into the lab, a timed test feature was created. It appeared that prisoners had to exhaust virtually every method of cheating they could dream up before taking programming seriously. In order to counteract the logjams caused by prisoners who deemed themselves slick enough to guess their way through computerized testing, the software was reprogrammed to give each student fifteen minutes to take a twenty-question test before the computer shut down, locked up and sounded an alarm. This feature was very successful in maintaining a brisk flow of students through the testing labs and short-circuited manipulators taking excessive amounts of time with vain attempts to guess their way through computer generated tests. Many inmates made it clear that they wouldn't settle for the word of other inmates that the program couldn't be beaten. They had to test it for themselves, often taking massive amounts of computer time to try to guess their way past tests, or try different kinds of cheating methods.

The successful automation of these features knocked out 95% of inmate cheating, maintained program validity, high program volumes, a smooth flow of inmates through the lab and valid measurability of program objectives. Computerization also preserved the programs' utility, as an inmate management tool for administration. This means that instead of running around and getting in trouble with the guards, thousands of hours of previously unstructured inmate time was now taken up in studying for programs. In prison, the term "Idle hands are the devil's workshop takes on a poignant new meaning.

Back to Chapter 8