Heatmap Analysis Shows No Difference Between Men and Women in Visual Attention

Neurons HQ

January 18, 2023

Main findings

Men and women show no differences in visual attention to images.

Across 180 images, the correlation coefficient between male and female versions of eye-tracking heatmaps for the same static stimuli was an average of 97%.

A correlation coefficient of this magnitude means that the difference between men and women is minimal and that men and women respond extremely similarly to images from different categories.

Heatmaps on a static ad showing very high correlation between female (top) and male (bottom) participants.

Scope of the study

In today's society, there has been a growing interest in understanding the similarities and differences between men and women in terms of their behavior and visual attention.

The widespread belief that men and women perceive and respond to visual stimuli differently has led to the development of gender-specific communication strategies in advertising and marketing.

Gender-based marketing is built on the assumption that men and women require different approaches to capture their attention and that consumers are more receptive to products or communication tailored to their gender.

However, our study challenges this assumption by demonstrating that men and women show no significant differences in their visual attention to images.

Methodology

We ran the study to test whether there are gender-specific differences in eye-tracking heatmaps. The study plays a key role in having a holistic view of the Neurons AI model’s performance.

  • Participants: A total of 177 participants, 90 females and 87 males from 5 countries (i.e., Denmark, USA, Brazil, Guatemala, and Iran) were exposed to a subset of images from our internal database.
  • Image set: A total of 180 images were used and analyzed in this study. The images were divided into 6 categories (i.e., product packaging, static ads, desktop website, mobile website, SoMe ads, and other) with 30 images in each category.
  • Eye-tracking: All participants were exposed to each of the images in a pseudorandomized manner, without having any particular task. We used a stationary eye-tracking device (screen-based Tobii Pro Nano eye-tracker) to record participants' eye movements to create heatmaps overlain on each image, and as saliency maps that represent the concentration of visual attention for each image across the group.
  • Comparisons: We then compared the heatmaps of the first 5 seconds of exposure on a pixel-by-pixel level to answer the research question, using 3 different metrics (i.e., KL divergence, Similarity score, Correlation coefficient) usually used to assess the similarity of heatmaps.

Key findings

We found that across all 180 images, the correlation coefficient between male and female versions of eye-tracking heatmaps for the same static stimuli was an average of 97%. See the comparison summary across all categories in the chart below.

Correlation coefficient between men and women for different types of image categories. The box plot denotes the median (middle line), upper and lower quartile (box top and bottom), upper and lower extreme (whiskers), and outliers (data points).

A correlation coefficient of this magnitude means that the difference between men and women is extremely small, and that men and women respond extremely similarly to images from different categories.

Stimulus image with one of the highest correlation coefficients.

Interestingly, there were some examples of relatively lower correlation coefficients like 86%. Nevertheless, as the correlation coefficient is still high in these cases, the main conclusion that men and women pay attention to the same areas in an image holds.

Example of heatmaps with a lower correlation coefficient. Top: heatmap for female participants. Bottom: heatmap for male participants.

Caveats and further actions

The study did not examine specific measures of eye-tracking that are locked to specific areas of interest (AOIs) such as total fixation duration (TFD) and time to first fixation (TTFF). This lies beyond the scope of the current study, but is part of our future research.

Additionally, it is worth noting that this study did not consider participants' emotions or cognitive responses while viewing the images. It might be that men and women may look at the same areas, but have different emotional or cognitive responses to what they are viewing, which is something that future studies will look into.

Limitations and disclaimer

Initially, we attempted to use a more extensive dataset with a higher number of images but found that a higher number of participants per image was required. As a result, we ultimately decided to use data from our previous global eye-tracking study.

In this study, we aimed to use images that may not be heavily skewed toward a particular gender in order to minimize any potential bias in our results. However, it is worth noting that whether an image is perceived as gender-neutral is highly subjective.

This is a preliminary study conducted on cisgender male and female participants.

Heatmap Analysis Shows No Difference Between Men and Women in Visual Attention

Neurons Icon

Ready to drive revenue with creatives that work?

Get a free demo
ArrowRight