Accurate assessment of temperature trends in the atmosphere is an important tool in our understanding of climate change. Currently there are three databases derived from satellite based Microwave Sounding Unit (MSU) radiance measurements. The three separate databases produce different results in the middle troposphere (termed MT) temperature trends, with two of the databases producing lower troposphere (termed LT) trends with differing results, introducing uncertainty which prevents the community from deriving distinct conclusions. Comparison studies have been performed using ancillary data to discriminate which method of database construction represents the closest actual climate evolution without success. This study introduces the concept of the zero trend level (ZTL) which allows the globally averaged atmosphere to be analyzed as a dichotomous system, one cooling layer and one warming layer over a chosen time period. The ZTL concept together with shorter MSU derived trend time periods are used as a means to elucidate between trend method construction by comparing consistency of the different methods with the latest data available. Results from the ZTL analysis show an important insight into the evolution of the atmospheric temperature trends over the last four decades. Combined with shorter time period analysis the ZTL analysis provide evidence that the University of Alabama-Huntsvilee (UAH) derived database is more self consistent than the Remote Sensing Systems (RSS) database over the 1988-2003 time period within the uncertainty, validity and accuracy of the latest radiosonde data available. NOAA-11 and NOAA-14 uncorrected MSU hot plate anomaly signatures are visible in the RSS-UAH brightness temperature monthly anomaly time series in the LT channel only, indicating this to be source of the inconstancy.