Weighting test scores by appropriateness of confidence, has almost without exception raised the reliability of test scores. Greater gains appear to occur for the less reliable tests, but that is at least partly because the more reliable a test is to begin with, the more difficult it is to improve it. If confidence-testing allows us to weight heavily on well-settled knowledge, then the weighted score might be more valid. In the research cited, retention was the aspect to be measured. High school physics students were tested upon completion of four chapters in their text. Their scores were used to predict semester grades. Weighted scores had higher reliabilities and correlated significantly higher with semester grades. An inference is made that conventional scores become less relevant to retention with the passage of time and confidence scores become more relevant. Unquestionably, personality has an effect on confidence-marking. However, the fairness of personal bias is summed up in stating that the bias for conventional scoring was equal or greater than that for weighted scores. Suggestions for research include more investigation of personality bias and a greater concern with validity. (Author/KJ)