Introduction
Participants
A total of 134 participants including Florida International University students, family members and personal acquittances of the investigator were recruited for the study. Of these 134 participants, 45.5% (n = 61) were male and 54.5% (n = 73) were female. Ages ranged between a minimum of 17 and a maximum of 59 with a mean of 25.4 years old (SD = 8.03). The sample included 43.3% (n = 58) of Hispanic Americans, 27.6 % (n = 37) Caucasians, 15.7 (n = 21) African Americans, 5.2% (n = 7) MENA, 4.5% (n = 6) Asians, and 3.7% (n = 5) Indigenous.
Material and Procedure
Before data collection, all the participants were informed about the possible risks associated with the study. The participants were approached either in person or through electronic means, such as email, social media, and online messengers. After receiving a verbal or written approval, depending on the means of approach, the participants were randomly assigned into one of three groups and asked to complete a questionnaire. The questionnaire consisted of three parts. For the first part, all the participants were asked to imagine that they were approached by a best friend to respond to a social media post. The social media post was as follows:
“Hello there, my best friend! I know you see a lot of surveys and games on social media, but I need you to take this one seriously. Below is a list of words. I need you to look at each word and rate the extent to which you think that word describes _____. Be as honest as possible! Thanks!”
The blank in the social media post differed for three groups of participants. The first group, the blank was filled with “you,” for the second group, it was filled with “me, your best friend,” and the for the third group, the blank was filled with “the United States President.” Thus, the participants were asked to complete the questionnaire with different conditions. After the participants rated the words, they were asked to complete a distraction task, which asked to replace letters in the words with the numeral order. After the distraction part was over, the participants were asked to recall the twelve adjectives they rated from a list of 42 words. The number of correct recollections was registered as the score for the first part of the questionnaire.
The second part of the questionnaire asked the participants to rate how much they agreed with three statements on a scale from 1 to 7, where “1” stood for “Strongly Disagree” and “7” stood for “Strongly Agree.” The first statement measured how confident the participants were with the recall part, the second statement measured if they thought that they did better than average with the recall part, and the third statement measured if the participants believed that they remember things better if they relate it to themselves. Even though the questionnaire measured several variables, this study focused on the analysis of differences in confidence levels of participants from different groups.
The final section of the questionnaire included four demographic questions, which were used to collect the demographic data from the participants. The questions asked about the participants’ age, gender, ethnicity, and first language. Additionally, the last section asked the participants about how to encode the word list during the Social Media Post One.
Results
Using survey conditions (self-rating, friend-rating, and President-rating) as the independent variable and the attention check question as the dependent variable, a Chi-square analysis was conducted. The results of the analysis demonstrated that the manipulation had a significant effect Χ2(4) = 96.41, p < 0.001. The majority of participants from all three groups remembered correctly about the instructions of the recall part. However, the percentage of participants that were correct from the President-rating group (82.2%) was higher than the number of correct participants from the self-rating and friend-rating groups (69.6% and 60.5% correspondingly). The results suggest that the participants paid attention to the instructions.
The first one-way ANOVA test focused on the differences of mean recollection scores depending on the survey condition (self-rating, friend-rating, and President-rating). The results demonstrated that there was a significant difference in mean recollection scores with F(2, 131) = 10.62, p < 0.001. Post-hoc analysis demonstrated that that the mean scores were the highest in the self-rated group (M = 9.91; SD = 1.07), while the score in the friend-rated (M = 8.77; SD = 0.81) and President-rated groups were lower (M = 9.2; SD = 1.55). The differences between friend-rated groups and the President-rated group were statistically insignificant (p = 0.21).
The second one-way ANOVA test focused on the differences of mean confidence scores depending on the survey condition (self-rating, friend-rating, and President-rating). The results demonstrated that there was a significant difference in mean confidence scores with F(2, 131) = 8.94, p < 0.001. Post-hoc analysis demonstrated that that the mean confidence scores were the highest in the self-rated group (M = 5.65; SD = 1.14), while the score in the friend-rated (M = 4,84; SD = 0.82) and President-rated groups were lower (M = 4.84; SD = 0.9). The differences between friend-rated groups and the President-rated group were statistically insignificant (p = 0.73).
Discussion
The initial supposition was that the participants from the self-rated would be able to recollect more words in comparison with participants with other groups due to the self-reference effect. The results of this study supported the hypothesis, as the mean number of words was significantly higher in the self-rated group in comparison with friend-rated and President-rated groups, which implies that the self-reference factor was significant. Similarly, this study hypothesized that the recall confidence scores would be higher for participants from the self-rated group in comparison with other groups due to the self-reference effect. The ANOVA analysis provided empirical evidence for the hypothesis, confirming that the participants were more likely to be confident about their answered, if they referred to themselves during the assessment. However, it should be mentioned that the attention check demonstrated that while the majority of the participants were attentive to the instructions, 29% of the participants were not attentive to the instructions (n = 39). This brings the question of whether the results would be different if more participants were attentive to the instructions.