Certain laboratory studies show that it is necessary only to hear up to 1500, or perhaps 2000, cps for normal word‐list reception. However, many patients with normal hearing up to 2000 cps but deficiencies at higher frequencies have noticeable difficulty with everyday speech. The discrepancy may arise since ordinary speech is often distorted and/or masked, conditions not usually included in speech‐reception studies. Discrimination scores (DS) for sentence material were collected on 52 sensorineural hypoacusics, the speech being distorted by having talkers wear nose clamps, by speed, interruptions, and reverberations. When the mean DS from these four distortions was averaged with the DS for undistorted speech, to create “50% distorted speech,” a condition assumed to represent everyday listening, the important acuities were at 1000, 2000, and 3000 cps. The formula for predicting everyday speech should then utilize the simple average hearing loss at these three frequencies. The Pearsonrbetween this predictor and the DS for 50% distorted speech was 0.74. Multiple‐correlation frequency‐weighting techniques are inferior to simple averaging across audiometric frequencies for theoretical reasons and because of the almost insurmountable problem of collecting a truly representative sample of hearing‐loss patterns. These results confirm those of Kryter, Williams, and Green, who found the region 2000–4000 cps to be the important predictor for reception of loud speech heavily masked by noise.