OFF THE WALL. The Effects of Artist Eccentricity on the Evaluation of Their Work ROUGH DRAFT

Transcription

1 OFF THE WALL The Effects of Artist Eccentricity on the Evaluation of Their Work ROUGH DRAFT Hannah Thomas AP Statistics Period 6 May 29, 2014 This study explores the relationship between perceived artist eccentricity and how observers evaluate the skill and quality of his/her artwork. Although previous studies suggested that eccentricity improves evaluations, this study did not produce significant evidence to suggest this claim.

2 2 An artist creates a print of an ordinary Campbell s soup can, but what makes it such a legendary work of art? Does the print appeal to the observer s idea of aesthetic? Do they find deep meaning in a simple image? Or is it the artist behind the work that invokes such praise? Andy Warhol fascinated the public with his unconventional behavior and avant-garde style, which may have been a large part of his success. A recent study published in the European Journal of Social Psychology revealed that people are more likely to find art appealing when they know an artist is eccentric, coining the term eccentricity effect. Researchers showed a group of students an image of Van Gogh s sunflower painting and informed the students that the artist cut off his own ear lobe. They found that these students rated the artworks higher in quality than groups who were told nothing. Other studies showed similar results. Experiments like these provide insight on the effect of stereotypes on social and cultural perceptions, and may explain why some artists and musicians do what they do. To test the eccentricity effect, students at Austin High School were asked to rate two paintings on a scale of 1 (no appeal) to 5 (high appeal). Some were given information about the artists denoting whether or not they lived eccentric lifestyles, while others received no information. The paintings were chosen to reduce bias in as many ways possible. For example, two relatively unknown pieces of art were chosen to prevent recognition. They were similar in appearance, but were painted by different artists, in different time periods. The selected artworks were both painted in a monochrome grayscale to reduce bias pertaining to color preference among participants. Both artworks were of an abstract design so that objects in the paintings did not act as confounding variables.

3 3 Painting 1: Jones, John Paul. Diversion Intaglio on paper. Walker Art Center, Minneapolis. Painting 2: Hubbard, George. Untitled. No. date. Various mediums on canvas. Participants in the experiments were selected through a simple random sample (SRS) of all students enrolled in Austin High School. Students were assigned a number between 1 and 2114 and all had an equal chance of being selected. The random number generator function on the TI-84 calculator was used to randomly select forty-five participants. The selected participants were then randomly assigned to one of three groups, each composed of fifteen participants. The first group served as a control; participants were given no information about the artists of the paintings. The second and third groups were given short, fictional artist biographies to accompany the paintings. Group 2 was told that Painting 1 (pictured above) was painted by a noneccentric artist, while Painting 2 was done by an artist who was eccentric. Group 3 was told the opposite of this. Because the sample was not blocked by grade level, gender, or artistic knowledge, it is possible that it was not truly representative of the population. For example, there were not an equal number of males and females in each group. If the paintings quality ratings are somehow influenced by gender, this could produce some bias. This is the same for grade level, as well as experience with artistic critique. Some

4 4 participants may have known what to look for to evaluate the skill and quality of the work, while others may have evaluated the art based on personal appeal. Ultimately, the pure, consistent, and thorough randomization in this experiment, along with the sampling method, made it possible to analyze data and use the tools of statistical inference. This is evidence of mastering one of the four conceptual themes of this course, anticipating patterns. As data was being collected, it became clear that there was not a significant difference between the ratings of Painting 1 and Painting 2, even when artist eccentricity was introduced. The anticipated patterns of data were nonexistent. Upon analysis, this proved to be true. Each painting in the three groups maintained a mean rating of approximately 3 on a scale of 5. The sample standard deviations were also very similar to each other. The lack of variation could be attributed to the small range of values, which also made it difficult to differentiate between what was statistically significant and what was not. To investigate the data further, it was necessary to ensure that all conditions were met in order to create confidence intervals and run hypotheses tests. To determine the true population mean difference in the ratings of Painting 1 and Painting 2 by students at this school, a one-sample hypothesis test for mean difference was conducted. Because the data for Paintings 1 and 2 were paired in the control group, it was necessary to state this under conditions for a test of mean difference. After validating that an SRS had been conducted, the next step was to observe the normal quantile plots of the difference between Painting 1 and Painting 2. Because the NQP was approximately linear, it was determined that the data was approximately normal. It

5 5 was also verified that the sample size did not exceed 10% of the population. Finally, because the population standard deviation, σ, was unknown, a t-test would be necessary, rather than a z-test. The calculated critical value was with a consequent p-value of.582. At significance level of α =.05, we failed to reject the claim that the true population mean difference in the ratings of Paintings 1 and 2 by students at this school is 0 because the p-value was greater than α. Thus, there was not sufficient evidence to support a preference of Painting 1 over Painting 2, or vice versa, in the population.