Ball bearings are widely used for providing support to the rotating parts in machinery and vehicles. Any damage to a bearing element during operation leads to vibration and catastrophic failure. Therefore, it is essential to monitor the condition of bearing elements during operations to detect early the occurrence and propagation of defects in bearing elements. Thus, we are motivated to identify the best condition indicator for detecting bearing defects and tracking their progression. The objective of this paper is to study and analyze the effects of different types of bearing defects and their sizes on bearing vibration responses, using different time-domain statistical indicators, and to determine the best indicator for detecting bearing defects and the evolution of defect sizes. In this paper, vibration signals obtained from normal and defective bearings are analyzed by using six traditional time-domain statistical indicators (TDSIs); peak, root mean square, crest factor, kurtosis, impulse factor and shape factor. Also, six new indicators developed by other researchers, namely TALAF, THIKAT, “kurtosis, crest factor and root mean square (KUCR)”, engineering condition indicator (ECI), SIANA, and INTHAR, are used to analyze the vibration signals. In addition, the effects of shaft speed on vibration responses are analyzed for a normal bearing using all these indicators. Vibration signals of bearings are obtained from the bearing datasets which are made available by the bearing data center of Case Western Reserve University (CWRU). A MATLAB code is developed to obtain TDSIs and new indicators from the data sets. In the results, it is found that KUCR is the most sensitive indicator to the detection of incipient defects and evolution of defect size; however, shape factor and TALAF are less sensitive to defect size detection.