Limits...
Does sadness impair color perception? Flawed evidence and faulty methods

View Article: PubMed Central - PubMed

ABSTRACT

In their 2015 paper, Thorstenson, Pazda, and Elliot offered evidence from two experiments that perception of colors on the blue–yellow axis was impaired if the participants had watched a sad movie clip, compared to participants who watched clips designed to induce a happy or neutral mood. Subsequently, these authors retracted their article, citing a mistake in their statistical analyses and a problem with the data in one of their experiments. Here, we discuss a number of other methodological problems with Thorstenson et al.’s experimental design, and also demonstrate that the problems with the data go beyond what these authors reported. We conclude that repeating one of the two experiments, with the minor revisions proposed by Thorstenson et al., will not be sufficient to address the problems with this work.

No MeSH data available.


Related in: MedlinePlus

Color categorization as a function of saturation in (a) Experiment 1 and (b) Experiment 2. Each panel shows the proportion of correct responses for each of the six saturation levels for a given color. Mean performance in thesadness condition is represented by triangles joined by solid lines, and mean performance in thehappiness (Experiment 1) andneutral (Experiment 2) conditions is represented by circles joined by dashed lines. Error bars show parametric 95% confidence intervals on the means. (Details of the color calibration procedure were not stated byThorstensonet al. (2015a), so it is not clear how to interpret these saturation values.)
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4979646&req=5

f4: Color categorization as a function of saturation in (a) Experiment 1 and (b) Experiment 2. Each panel shows the proportion of correct responses for each of the six saturation levels for a given color. Mean performance in thesadness condition is represented by triangles joined by solid lines, and mean performance in thehappiness (Experiment 1) andneutral (Experiment 2) conditions is represented by circles joined by dashed lines. Error bars show parametric 95% confidence intervals on the means. (Details of the color calibration procedure were not stated byThorstensonet al. (2015a), so it is not clear how to interpret these saturation values.)

Mentions: AOH coordinated the team and wrote most of the main section of the article. NJLB started the discussion (on Twitter), wrote the R code to perform the detailed analysis of the dataset, and wrote most of the sections describing this analysis. PTG contributed several points andFigure 4 as well as contributing to the writing. AE contributed to the discussion of stimuli problems and (lack of) necessary stimulus sampling. SG contributed to the analysis, contributed early versions of some of the R code, and helped revise the manuscript. All authors agreed to the final content of the article.


Does sadness impair color perception? Flawed evidence and faulty methods
Color categorization as a function of saturation in (a) Experiment 1 and (b) Experiment 2. Each panel shows the proportion of correct responses for each of the six saturation levels for a given color. Mean performance in thesadness condition is represented by triangles joined by solid lines, and mean performance in thehappiness (Experiment 1) andneutral (Experiment 2) conditions is represented by circles joined by dashed lines. Error bars show parametric 95% confidence intervals on the means. (Details of the color calibration procedure were not stated byThorstensonet al. (2015a), so it is not clear how to interpret these saturation values.)
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4979646&req=5

f4: Color categorization as a function of saturation in (a) Experiment 1 and (b) Experiment 2. Each panel shows the proportion of correct responses for each of the six saturation levels for a given color. Mean performance in thesadness condition is represented by triangles joined by solid lines, and mean performance in thehappiness (Experiment 1) andneutral (Experiment 2) conditions is represented by circles joined by dashed lines. Error bars show parametric 95% confidence intervals on the means. (Details of the color calibration procedure were not stated byThorstensonet al. (2015a), so it is not clear how to interpret these saturation values.)
Mentions: AOH coordinated the team and wrote most of the main section of the article. NJLB started the discussion (on Twitter), wrote the R code to perform the detailed analysis of the dataset, and wrote most of the sections describing this analysis. PTG contributed several points andFigure 4 as well as contributing to the writing. AE contributed to the discussion of stimuli problems and (lack of) necessary stimulus sampling. SG contributed to the analysis, contributed early versions of some of the R code, and helped revise the manuscript. All authors agreed to the final content of the article.

View Article: PubMed Central - PubMed

ABSTRACT

In their 2015 paper, Thorstenson, Pazda, and Elliot offered evidence from two experiments that perception of colors on the blue–yellow axis was impaired if the participants had watched a sad movie clip, compared to participants who watched clips designed to induce a happy or neutral mood. Subsequently, these authors retracted their article, citing a mistake in their statistical analyses and a problem with the data in one of their experiments. Here, we discuss a number of other methodological problems with Thorstenson et al.’s experimental design, and also demonstrate that the problems with the data go beyond what these authors reported. We conclude that repeating one of the two experiments, with the minor revisions proposed by Thorstenson et al., will not be sufficient to address the problems with this work.

No MeSH data available.


Related in: MedlinePlus