362 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS results from a clinical test (Study 4) using the four-wash protocol during a two-week period in which the average dew point for each week was nearly 50øF. These results indicate that increases in skin dryness were observed during both weeks of the study in spite of the high dew points. The results in this study were equivalent to those in Studies 2 and 3, indicating that the same results are achieved with both wash protocols. EFFECT OF TEST DESIGN In the initial phase of method development, the test design was as follows: single product design, small base sizes (25-30 per product), four washings a day, ten-second exposure to lather, and no washing on weekends. With this approach, statistically significant differences in skin condition were observed among certain products, as shown for Study 1 in Table II. In subsequent studies, the method of paired comparison was used in order to reduce the subject-to-subject variability and improve sensitivity. The initial pair tests were conducted on groups of about one hundred subjects with two daily washings and ten-second exposure to lather. This approach resulted in significant differences in skin condition between mild and irritating control products, as shown in Study 5 (Table III). Further method refinement indicated that significant differences could be determined with much smaller test populations (25-30) if the lather exposure time was extended to ninety seconds, as illustrated by a comparison of Studies 5 and 6 (Table III). FIGURE 2: Study 4 DRYNESS GRADE 3.0 2.5 2.0 1.5 1.0 0.5 Mild Control ] - 3.0 2.5 2.0 - 1.5 ß 1.0 0.5 0.0 0 1 2 WEEK Figure 2. Performance of the exaggerated-use protocol during moist ambient weather conditions. Average dryness grades for two bar soap formulations (an irritating control and a mild control) are shown from Study 4, when the average dew point was 48øF in Week 1 and 50øF in Week 2.

FOREARM WASH TEST FOR MILDNESS 363 Table III Effect of Test Design on Test Sensitivity* Study 5.' Base size = 115, pair test, IO-second exposure, two washes per day Clinical skin grades Product Redness Dryness Dew point Irritating control 1.43 1.97 Mild control 1.15 1.72 Difference 0.28 0.26 p-value 0.003 0.0001 27 ø F. Study 6.' Base size = 30, pair test, 90-second exposure, two washes per day Clinical skin grades Product Redness Dryness Dew point Irritating control 1.59 2.88 Mild control 1.27 2.35 Difference 0.31 0.53 p-value 0.054 0.004 27 ø F. Study 7.' Base size = 30, pair test, 90-second exposure, four washes per day Clinical skin grades Product Redness Dryness Dew point Irritating control 1.52 2.26 Mild control 0.96 1.56 Difference 0.56 0.70 p-value 0.0001 0.0001 56 ø F. * Irritating and mild control products are the same for all three studies. Studies 6 and 7 were conducted for the same two products under different dew point conditions. Study 6 used two daily washes and Study 7 used four. Significant differences were observed between the irritating and mild control products. EFFECT OF SEX AND SKIN COLOR (STUDY 8 AND STUDY 9) Data from Studies 8 and 9 were used to demonstrate differences between the males and females with regard to the clinical response to soap treatment. Table IV compares changes in skin grades for two products which are known from previous testing to differ in skin mildness. The final skin grades for the two products, the differences and the p-values from the statistical analysis, are given for both male and female subjects. The test products produce larger differences in females than in male subjects for redness, dryness, and smoothness. However, we continue to include male subjects in the test populations because their response is similar to that of females and their inclusion makes the studies more representative of the general population. The assignment of test product pairs is balanced with respect to the sexes. Skin color also affects the results of a product comparison clinical study. Trained graders