We hereby attempt to detect plausible reasons for the discrepancies between the measured and modeled tropospheric temperature anomalies in the tropics. For this purpose, we calculate the trends of the upper-minus-lower tropospheric temperature anomaly differences (TAD) for both the measured and modeled time series during 1979-2010. The modeled TAD trend is significantly higher than that of the measured ones, confirming that the vertical amplification of warming is exaggerated in models. To investigate the cause of this exaggeration, we compare the intrinsic properties of the measured and modeled TAD by employing detrended fluctuation analysis (DFA). The DFA exponent obtained for the measured values reveals white noise behavior, while the exponent for the modeled ones shows that they exhibit long-range power law correlations. We suggest that the vertical amplification of warming derived from modeled simulations is weighted with a persistent signal, which should be removed in order to achieve better agreement with observations.