Tanner B A
Eval Program Plann. 1985;8(2):147-53. doi: 10.1016/0149-7189(85)90009-6.
Because of the apparently inflated scores commonly reported for consumer satisfaction with mental health services, we developed a 5-item validity scale for the SHARP questionnaire. The scale discriminated significantly (P less than .001) between those former clients who offered suggestions for improvement of services and those who did not (N = 38). In another experiment, the scale significantly improved the ability to predict the 46 clients' true scores from their obtained scores via stepwise regression (p less than .01). Weights were obtained from that sample of former clients, which significantly reduced the deviation of the obtained satisfaction scores from the true scores, with all p less than .05. In two successive experiments (N = 49 and 26), weights were applied to new samples, significantly reducing deviation from the criterion scores for 5 of the 6 scales (p less than .05), and new weights were then derived. A final set of weights was derived from the combined samples of the three experiments, (N = 121) which significantly improved prediction of the criterion and lowered the mean on all scales, with all p less than .05.
由于心理健康服务消费者满意度报告中普遍存在明显夸大的分数，我们为SHARP问卷编制了一个包含5个条目的效度量表。该量表在那些为服务改进提出建议的前客户和未提出建议的前客户之间有显著差异（P小于0.001）（N = 38）。在另一项实验中，通过逐步回归，该量表显著提高了从46位客户的所得分数预测其真实分数的能力（p小于0.01）。权重是从前客户样本中获得的，这显著降低了所得满意度分数与真实分数的偏差，所有p值均小于0.05。在连续两项实验（N = 49和26）中，将权重应用于新样本，显著降低了6个量表中5个量表与标准分数的偏差（p小于0.05），然后得出了新的权重。最后一组权重是从三项实验的合并样本（N = 121）中得出的，这显著提高了对标准的预测，并降低了所有量表的均值，所有p值均小于0.05。
