News coverage of current affairs is one of the most frequently received information from the external world. The individual allocates a large amount of attention resources on these information from mass media every day. What information individual attends to process affects his/her cognitive construction of the world, emotional states of mind and behavior responses, and even affects the social ethos at the group level. However, few studies have examined the bias in the processing of news coverage. Group membership is a vital factor affecting people’s bias. In the question of whether individual attention and processing on news information are in positivity bias or negativity bias, the group attribute of news is an important factor which is often ignored. Therefore, in this study, two experiments were conducted to investigate the attentional selection processing bias in news context, and to explore whether the bias is modified by the news attributes, and whether there are differences of bias in early and late attention stages. Experiment 1 was 2 ( group attribute of news : ingroup news / outgroup news ) × 2 ( valence of news : positive / negative ) within-subjects designed. In experiment 1, a total of 209 participants were recruited and attended the experiment through WenJunXing online. There were 10 news items for each condition. The participants were asked to choose one news item which they intended to read from 2 news items randomly paired ( two news items contained one positive news item and one negative news item, and two items had the same group attribute. In short, they were both ingroup news or outgroup news ). In this way, experiment 1 could indirectly investigate the attentional selection bias toward news information from the selective rate in behavioral level. Experiment 2 was 2 ( group attribute of news : ingroup news / outgroup news ) × 2 ( congruence: positive congruence / positive incongruence ) x 2 ( SOA: 200ms / 500ms ) within-subjects designed. In this study, 30 participants were recruited and attended the experiment in a laboratory. There were also 10 news items as experimental stimuli for each condition. Dot-probe paradigm was adopted to directly investigate the attentional bias of participants. A fitting test was used to analyze the difference of news selective rate in different conditions in experiment 1. Results of experiment 1 showed that, the participants were more likely to choose positive news among the ingroup news( p < .001 ), but more likely to choose negative news among the outgroup news ( p < .05 ). A repeated measures analysis of variance was used to analyze the respond time in different conditions in experiment 2. Results of experiment 2 showed that, for SOA = 200ms, the participants responded faster in the positive congruence condition than the positive incongruence condition for both ingroup and outgroup news ( p < .001 ). However, in SOA = 500ms condition, the participants responded faster in the positive congruence condition for ingroup news( p < .001 ), while responded faster in the positive incongruence condition for outgroup news ( p < .001 ). In this study, both experiment 1 with high ecological validity and experiment 2 with high internal validity found that there were a positive processing bias toward ingroup news and a negative processing bias toward outgroup news. These findings support the theory of social identity and the theory of social comparison, which declare ingroup favoritism and outgroup derogation in individual’s processing bias. However, these biases existed only in the conscious selection behavior and the late stage of attention ( SOA = 500ms ) involving cognitive processing. In the early attention stage ( SOA = 200ms ), the participants showed positivity bias for both ingroup news and outgroup news. We can infer that positivity bias in the processing of news information can be expressed in the early stage of attention because the valence of news is involuntary processed. While in the conscious stages with cognitive resources involved, the bias can be cognitively adjusted by news group attributes according to social identity theory and social comparison theory.