短期正念引導降低閱讀時的內隱性別刻板印象

Brief Mindfulness Induction Reduces Implicit Gender Stereotype During Reading

正念引導可以降低年齡與種族的偏見及歧視，並且正念特質能使男性更和善地對待女性主義者。本研究延伸上述發現以探討短期正念引導是否能降低性別刻板印象。本研究採三因子混合設計，參與者間獨變項為短期正念引導（實驗組與控制組），參與者內獨變項為文章段落之一致性（一致與不一致）以及時間（前測及後測）。參與者閱讀含有性別特質描述的文章，文章第一段與第三段分別置入典型性別特質詞彙，其中一半文章中兩個段落的特質詞彙的性別屬性相同，另一半則不同。研究者觀察當兩個段落中特質詞彙的性別屬性不一致時，參與者是否花費較長的時間注視與性別特質有關的詞彙或是文章段落，而將其作為性別刻板印象的指標，探討正念訓練是否能降低此效應。結果顯示兩組參與者於前測中第一段落再閱讀時間產生性別不一致效應，顯示性別刻板印象的操弄產生效果。此外，實驗組參與者於前測中第三段再閱讀時間與第一段性別特質詞彙再閱讀時間的性別不一致效應均在後測中消失，顯示其在經過正念引導後減弱了性別刻板印象。從過去研究推測正念引導造成刻板印象降低的機制可能與正念狀態中的開放與不批判態度還有提升的內在動機有關。

According to the Office of the United Nations High Commissioner for Human Rights, gender stereotype refers to“a generalized view or preconception about attributes or characteristics, or the roles that are or ought to be possessed by, or performed by, women and men”(OHCHR, n.d.). Gender stereotypes can be categorized as explicit or implicit. Explicit gender stereotypes are consciously endorsed, observable, intentional biased attitudes or behaviors (National Research Council, 2004), whereas implicit gender stereotypes are generally the biased learned associations made before conscious intentions form.
Garnham et al. (2002) used eye movement measures to estimate individuals’implicit gender stereotypes. Participants read sentences in which a person’s gender was not explicitly mentioned but was implied by the person’s role name (occupation) indicated at the beginning of the sentence and by the physical attributes and clothing mentioned at the end of the same sentence. These role names, attributes, and clothing were typically associated with either men or women. They found that participants exhibited poor reading performance when the gender-associated role name was inconsistent with the gender-associated physical attributes and clothing (i.e., gender mismatch effect). Similarly, Pyykkönen et al. (2010) found that when individuals heard descriptive sentences about role names associated with gender stereotypes, they tended to look at pictures consistent with gender role stereotypes.
Several studies have suggested that gender stereotypes influence individuals’attitudes and substantially affect their behaviors (González et al., 2019; Tilcsik, 2011). Therefore, to contribute to gender equality, the current research provided effective methods for reducing gender stereotypes. Brief mindfulness induction (MI) has been reported to reduce stereotypes, including ageist, classist, and racist stereotypes (Djikic et al., 2008; Lueke & Gibson, 2015, 2016; Parks et al., 2014).
Mindfulness is generally defined in modern psychology as deliberate, nonjudgmental awareness of the present moment. Mindfulness is considered to be a state that can be achieved through practice and a trait that indicates the extent to which individuals experience a mindful state in daily life (Bishop et al., 2004). Lueke and Gibson (2015) found that participants who received MI exhibited significantly less implicit bias toward certain groups of individuals. Lueke and Gibson (2016) used the trust game and found that brief MI reduced racial discrimination. In both studies, participants were not aware that the research intended to measure their own racial prejudice; therefore, implicit racial prejudice manifested itself in behavior.
To the best of the authors’knowledge, no study has directly examined whether MI is effective for reducing gender stereotypes, bias, or discrimination. However, Gervais and Hoffman (2013) found that mindfulness in men is associated with greater warmth toward feminists. This study applied brief MI to reduce gender stereotypes to address the gap in the literature. Because studies have identified the positive effects of MI on reducing racial and age biases, this study hypothesized that brief MI reduces implicit gender stereotypes in individuals. First, we applied Garnham et al. (2002) method and asked participants to read short texts related to gender stereotypes. In addition, we applied the results of Koenig (2018) to establish typical gender–trait words based on a sample of Taiwanese college students, and we manipulated the consistency of their application in the first and third paragraphs of the texts. For example, in the gender-inconsistent texts, if the gender associated with gender–trait words was female in the first paragraph, the gender associated with the words was male in the third paragraph.
Fifty-eight undergraduate students from Chung Yuan Christian University were recruited and randomly assigned to either receive MI or no intervention. This study adopted a three-factor mixed design. The between-subject independent variable was group (experimental group or control group). The within-subject independent variables were the congruency of the gender associated with typical gender–trait words in the first and third paragraphs of the texts and time (pre- and posttest). First dwell time and rereading time of the first and third paragraphs and gender–trait words in these paragraphs were the dependent variables. First dwell time refers to the time for which participants’eyes were directed at a target area until leaving that area. Rereading time refers to the total reading time subtracted by the first dwell time.
Participants’eye movements were recorded using an EyeLink 1000 Plus eye-tracker manufactured by SR Research. At pretest, participants read eight texts, four of which contained typical gender–trait words (two gender-consistent and two gender-inconsistent texts), and four that were fillers. Subsequently, participants in the experimental group received 30-minute MI once a day a total of three times. In each MI course, a researcher guided one to three participants. The MI consisted of warm-up exercises and formal practice. The warm-ups included radix linguae movement, eye rotation, and single nostril breathing and were based on body–mind axial awareness practice (Chang, 2018; Li, 2017; Lien et al., 2019); the formal practice, including breathing mediation and body scan, were based on mindfulness-based cognitive therapy (MBCT; Williams et al., 2014/2016). After the MI, each participant participated in the posttest.
The first hypothesis of this study was that when the gender–trait words in the first and third paragraphs differed in the pretest, the participants’reading patterns would differ between the consistent and the inconsistent texts. This gender mismatch effect may have been particularly reflected in the rereading time at the paragraph or word level in the first paragraph and in the first dwell time and rereading time at the paragraph or word level in the third paragraph. The second hypothesis was that brief MI reduces gender stereotypes. That is, after intervention, the reading patterns of the experimental group between the inconsistent and consistent texts would be similar, but the difference in the reading patterns of the control group between the consistent and inconsistent texts would remain unchanged.
The main results have two parts: The gender mismatch effect and the effect of MI. First, the gender mismatch effect was observed at both paragraph and word levels. At the paragraph level, participants spent a longer time rereading the first paragraph in the inconsistent texts, indicating that participants may have been confused when they encountered the gender–trait words in the third paragraph and returned to reread the first paragraph. In addition, at the word level, participants in the experimental group spent significantly longer rereading the first paragraph in the inconsistent texts, demonstrating the presence of the gender mismatch effect. However, the gender mismatch effect was not exhibited in the control group at the word level.
The results demonstrated that MI reduced the gender mismatch effect. At the paragraph level, participants in the experimental group exhibited significantly higher rereading time for the third paragraph in inconsistent texts in the pretest, but this gender mismatch effect was not observed in the posttest. A similar phenomenon occurred at the word level. In the pretest, the rereading time for the gender–trait words in the first paragraph in inconsistent texts was significantly higher than that in the consistent texts, but these times were equal in the posttest. This indicates that brief MI effectively reduced gender stereotypes.
Although this study did not probe the mechanisms by which MI reduced gender stereotypes, several possibilities are proposed based on research. First, because the state of mindfulness enables individuals to be nonjudgmental and exhibit acceptance, MI may reduce the automatic connection to negative constructs and biases (Lueke & Gibson, 2015, 2016). Second, it has been found that MI improves reading comprehension, which enables individuals to integrate textual messages (Clinton et al., 2018) and episodic memories (Brown et al., 2016); therefore, participants may have remembered more details regarding the gender–trait words in the texts after MI and therefore were less likely to reread them. Finally, Gervais and Hoffman (2013) suggested that mindfulness enhanced men’s intrinsic motivation to think about women in a nonsexist manner, and thus it is possible that MI exerts similar effects.
To our knowledge, this is the first study that used typical gender–trait words to elicit implicit gender stereotypes and the first study that applied MI to temporarily reduce implicit gender stereotypes. Therefore, our findings enrich the experimental literature that can be used for future gender-related research and propose a simple and practical practice that may benefit gender equality.