To investigate the Halo effect.

Our experiment was about the Halo effect, a cognitive bias in which someone's positive trait or traits (typically (good) physical appearance) subconsciously influences another's perception of other unrelated traits, so when the goodness of one's traits cause someone else's perception of their other, possibly less desirable traits to be better than they really are. In simpler terms, someone's good physical appearance (or any other desirable attribute) can make others believe that they are more kind or intelligent than they really are, as it 'shines out like a halo over the other traits.

The original study which we attempted to replicate was Dion et al (1972). The aim of this study was to investigate the role of the halo effect when choosing a partner.

There was a sample size of 61 undergraduate students from a US university (30 male and 31 female). The participants were told that it was a study on "accuracy in perception" and compared with people "trained in people perception". Each participant was given 3 envelopes with an image of a person, one of which being attractive and of a similar age to the participants, another being moderately attractive, and the other not being considered attractive. The levels of attractiveness of the people used in the stimuli were determined by the results of a survey conducted by the same university before the study. 100 students participated in said survey. In Dion et al (1972), the participants were asked to rate each stimulus person based on 27 personality traits, graded on a 6-point scale. They did this for each photo from the envelope, before completing another survey in which they were asked which one was most likely to experience parental happiness, marital happiness, and general happiness. The final task was to match or assign each person to different professions, being of varying status.

The results established that they found a positive correlation between attractiveness, success and having a good personality, that is, the more attractive the person was, the happier and more successful they would be, in addition to possessing a personality of a higher quality.

Our experiment aimed to investigate if, and to what extent this effect happens, with high school students. This is very relevant in modern life as it is prevalent in various forms, such as celebrities, influencers, and social media. Participants in our study had to read descriptions with a photo of someone and then rate some of their personality traits. The photos were of people of varying levels of attractiveness: 3 being "attractive", the other 3 not being particularly so. We chose to only compare the trait of open-mindedness, as specification of the degree to which this trait was present in the person was not given in any of the descriptions, therefore meaning that the rating would be the most subjective.

The experimental hypothesis (H₁) was that the more physically attractive someone was, the higher their overall rating score would be. The null hypothesis (H₀) was that the relationship between these things was null and there would be no difference.

We used a repeated measures design to best compare how each participant rated the attractive and unattractive people. If we had used independent samples, this may have led to results that were biased and thus harder to compare. Although we were concerned about the limitations of a repeated measures design, such as order effects, we mitigated these concerns by limiting the number of descriptions to 6 and also randomly distributing the unattractive and attractive photo stimuli throughout the questionnaire.

We used an opportunity sample by approaching participants in communal areas and asking if they'd be interested in participating in the study. We chose to use this sampling method because it was the most convenient, fast, and efficient in a school setting.

Sample size - 23 participants.
Gender ratio - Male : Female : Decline to Answer = 10:13:1.
Age range - all participants were between 16 and 18 years, inclusive (16 ≤ x ≤ 18).
Sexual Orientation - See below.

Fluency in English -
All participants could speak English to a certain extent, given that this study took place in a school located in England wherein the courses are also taught in English. However, from researcher observation, the level of proficiency varied somewhat. Data regarding participants' fluency in English was not collected for this experiment, as it was not deemed to be necessary pertaining to the aims and objectives of the investigation.

In order to control any extraneous variables, we conducted the experiment in a quiet and controlled environment (an empty classroom). We also provided standardised instructions and the same questionnaire to all participants to ensure that they all received the same experience. Lastly, we asked participants to not discuss the experiment with others. We did a pilot study to operationalise attractiveness to better reduce participant variability in subjective opinion regarding attractiveness.

There were no obvious ethical concerns regarding our experiment, however the participants were informed (on the consent forms) that should they have any, they could email us and we would do our best to sort it.

Data was anonymised and any personally-identifiable information was redacted in any documentation. All participants were debriefed after they completed the procedure by means of a debriefing form to allow them to know about what they did, why they did it, and its significance.

Before we began our main experiment, we did a pilot study in order to operationalise our independent variable of attractiveness. We showed participants pictures of people and asked them to rate the attractiveness of these people based on appearance. We then split these pictures into attractive or unattractive categories and used these pictures in our experiment.

The procedure we took for our experiment was as follows (after planning and creating the materials, etc.):

1. All participants were given consent forms, which they read and signed.
2. Participants were emailed the link to the questionnaire, which had standardised instructions. They were given verbal instructions and explanations as well.
3. In a controlled environment, participants completed the questionnaire independently. Our researchers stayed with them in case they had any questions.
4. After completing the questionnaire, participants were given a debriefing form and then were permitted to leave.

For raw data collected from the questionnaire, please refer to Appendix 6.

In our rating of attractiveness, as decided by the pilot study, we created a cut-off point. Pictures rated between 3 and 5, inclusive (out of 10) were assigned to the "unattractive" category, whilst the pictures rated over 7, inclusive (out of 10) were assigned to the "attractive" category. Whilst we took the average rating for attractiveness, we decided to take the median scores for open-mindedness, since all the personality traits were ranked, ordinal data.

Here, we can see that the lower someone was rated in terms of attractiveness, the lower they were generally rated in terms of their open-mindedness, too. In comparison, the people rated as more attractive had higher open-mindedness ratings. This can be seen in the bar chart below: