The first research study I did in graduate school was on why people identified with individuals they only knew from the media. It would become my first ICA conference presentation — before PowerPoints were common!

I first interviewed people to understand who they were identifying with and why they were doing so. Based on those semistructured interviews, I created a questionnaire that I distributed as part of a class project. This was back in 2004, and I never did much with the work.

But I was always fascinated by the reasons people give for why they identify with media personalities in these parasocial relationships. Whether the media personality was real or fictional, individuals can develop strong feelings of identification, based on a sense of similarity to that media personality, due to their appearance, beliefs, or actions, or even the situation the personality finds themself in. They can even identify based on a wishfulness, or a desire to be more similar to that media personality.

I created a basic questionnaire to measure who people identified with and why. What I report here is on the 122 people who responded to the questionnaire.

Now, I have reported on this study elsewhere on this site, such as:

• https://playingwithresearch.com/2016/09/13/identification-with-favorite-media-personae 
• https://playingwithresearch.com/2011/08/02/media-personae-identification-early-thoughts 

But I never did much with this work other than use it for the past seven years in my communication research methods to teach my students some basic statistics using the SPSS program. In that class, I use a version of the original dataset to teach them how to run some basic inferential statistics and to look for patterns in the answers: correlations, chi-squares, and ANOVAs. I figure, now, why not share it with everyone else?

Correlation results:

“I identify with this person because I see similarities between myself and he/she” (similarity) positively correlates with…

• “The person I identify with inspires me.” (inspiration): r=0.333, p<0.000
• “I see a lot of myself in this person.” (similarity): r=0.528, p<0.000
• “I like to do the things the person does.” (similarity): r=0.260, p=0.004
• “I would like to be the person I identify with.” (wishfulness): r=0.244, p=0.007

“The person I identify with inspires me” (inspiration) positively correlates with…

• “I see a lot of myself in this person.” (similarity): r=0.382, p<0.000
• “I like to do the things the person does” (similarity): r=0.353, p<0.000
• “I can learn how to handle things in my life by watching the person.” (inspiration): r=0.385, p<0.000
• “I would like to be the person I identify with.” (wishfulness): r=0.184, p=0.042

“I see a lot of myself in this person” (similarity) positively correlates with…

• “I like to do the things the person does” (similarity): r=0.233, p=0.01
• “I can learn how to handle things in my life by watching the person.” (inspiration): r=0.257, p=0.004
• “I would like to be the person I identify with.” (wishfulness): r=0.186, p=0.040

“I like to do the things the person does” (similarity) positively correlates with…

• “I can learn how to handle things in my life by watching the person.” (inspiration): r=0.216, p=0.017
• “I would like to be the person I identify with” (wishful): r=0.294, p=0.001

Thus, the 3 similarity items measure similarity reliably as they all positively correlate with one another, as intended. In this dataset, I needed more items to measure wishfulness, but all three similarity items also correlate with the wishfulness item, suggesting that the more someone sees themselves as similar to the media personality, the more they wish they could be more like that person — or the more they wish to be like that person, the more see such similarities. Of course, I cannot infer causality from these correlations — and it may be that this connection between similarity and wishfulness has no such linear relationship.

The 2 inspiration items correlate with each other, suggesting reliability. The inspiration item that directly addresses inspiration correlates with all similarity and wishfulness items, which makes sense given the close relationship between those identification type items. Interestingly, saying they can learn from the identification does not correlate as strongly with the identification type items: not at all with wishfulness item, and only with items measuring doing the same things and seeing self in media personality. It may be that the person needs to see the media personality as similar in appearance and actions to believe that person can be a role model from which to learn how to live.

Chi-Square results:

If the participant was male, then they tended to report less identification with a female media personality: of 49 male participants, only 1 chose a female media personality. Similarly if the participant was female, then they tended to report less identification with a male media personality. Thus, according to this result, the participants indicated less cross-gender identification ( = 75.342, p<0.000).

If the participant was a teenager (18 or 19), then they were more likely to indicate that they first identified with the media personality as a teenager. Similarly, if the participant was a young adult, their first identification happened as a young adult; and if the participant was an adult, so was their first identification. Thus, the participants were more likely to discuss a more recent media personality identification (= 98.183, p<0.000).

If the participant was a teenager, then they tended to report a real media personality more than expected: indeed, no teenagers discussed identifying with a fictional character. If the participant was a young adult, then they did discuss more fictional media personalities than expected ( = 6.342, p=0.042).

If the participant said their ethnic heritage was Caucasian, they indicated more identification with Caucasian media personalities than expected, whereas participants with African heritages discussed more African media personalities than expected — and  both discussed less of the other ethnicity than expected. Thus, in this sample, there was less cross-ethnicity identification ( = 151.684, p<0.000).

ANOVA results:

For Age Group: when answering “I identify with this person because I see similarities between myself and he/she,” Teenagers (m=4.17, sd=0.529) and Young Adults (m=4.14, sd=0.67) were significantly different from Adults (m=3.75, sd=0.88) at F=3.598, p=0.03; thus, Adult were less likely to agree that they saw similarities between themselves and the media personality they identified with.

For Age First Identify: when answering “I identify with this person because I see similarities between myself and he/she,” Teenagers (m=4.04, sd=0.52), Young Adults (m=4.22, sd=0.79), and Adults (m=3.65, sd=0.93) differed significantly from one another at F=3.289 p=0.023; thus, the participants who were Young Adult when they first identified with the media personality tended to agree the most that they saw similarities with the media personality.

For Media Personality Type: when answering “The person I identify with inspires me,” people who identified with Real Media Personalities (m=3.96, sd=0.92) differed from people who identified with Fictional Media Personalities (m=3.23, sd=0.86) at F = 10.83 p<0.00; thus, people who identified with a real person tended to find more inspiration from that media personality.

Main Conclusions:

These analyses found evidence to support a theoretical relationship between similarity and wishfulness identifications, as I have discussed elsewhere on this site (see above).

If the participant was a Teenager, then they indicated more identification with real media personalities, with more similarity identification and more inspiration. This series of findings suggests a possible causal pathway: seeing a real person who they perceived as similar to them (physical, psychological, behavioral or situational) led them to identify with that person, wish to be that person, and thus be inspired by that person.

The participants indicated more identification based on the similarity of physical appearances (e.g. gender, ethnicity), suggesting the primacy in such visual cues to determine identification. Furthermore, participants indicated identifying with people who look like them, seeing that relate to doing things the same, and then also be inspired to be more like that person. The relationship between these variables demonstrates the importance of representation for similarity identification and to be inspired.

People want to see real people who look and act like them in the media, and they want to see this to help them make sense of themselves and their lives, as well as gain information and ideas for how to be themselves and live their lives. The more diversity there is in the media, then the more likely people will be able to experience this type of identification and be inspired to be their best selves and live their best lives.