Verified by 
There are many stereotypes that people have about dog owners and their dogs. For example, I remember being in a movie theater watching Arnold Schwarzenegger's action comedy film True Lies. When Schwarzenegger appeared on screen walking a tiny Chihuahua it caused an audible chuckle in the room. This was because what was being shown clashed with the popular stereotype that it is women who have small dogs and men are supposed to have large dogs. So here we had a man walking what was clearly a tiny "woman's dog" and many people watching found the clash with the stereotype amusing.
The memory of that incident returned to me when I got a note from Harold Herzog, Professor Emeritus in the Department of Psychology, Western Carolina University. Hal is a highly respected behavioral researcher who has studied many aspects of human-animal interactions. He wanted to know if I knew of any empirical evidence which indicated that women are actually more likely to own smaller breeds than men.
The popular belief that women tend to own smaller dogs is certainly strong, but Hal was asking "What do the data actually say?" To try to answer this question I spent a few hours combing through my files but I didn't find any studies specifically addressing that issue. There were a few studies in which there were secondary analyses, but these were usually in reports looking at who owns the so-called "dangerous breeds," with the consensus that men are more likely to prefer Rottweilers and the bully breeds.
In any event his question stuck with me and I began to think of different ways in which one might be able to obtain the necessary data. Some of the methods that could be used were obvious — if one had access to the data sources. For example, if you had a list of dog owners broken down by breed (such as from an American Kennel Club membership list) one could do a simple count looking at male versus female names. Alternatively one could contact breed clubs (selecting some clubs for large breeds and some small) and ask for member information by name or gender. However, my guess is that in these litigious times the various clubs would not release the data based on "privacy considerations."
Nonetheless this unanswered question was now bothering me and I wanted to be able to give Hal some kind of meaningful response. To get an answer I decided to do a version of "citizen science": I went onto the image database Flickr (which posts millions of submitted photos) and did two separate searches. The first one used the search terms "woman" and "dog" and the second used the terms "man" and "dog." If the stereotype is correct we should find a lot more images of women with small dogs and men with large dogs.
For the purpose of this study I was looking for images which showed one single human being and enough of the image of a dog to allow its size to be estimated or its breed to be identified. Images with multiple dogs were not scored (unless they were the same breed) and images with multiple human beings were not scored. Images in which the dog appeared to be a puppy were also eliminated. I decided to be a bit fussy, so I also eliminated all images in which the dogs were engaged in some sort of service activity, or hunting, or were obviously part of military or police functions. I also eliminated pictures of homeless people with dogs since those were apt to be chance connections rather than based on a considered choice. The dogs in the photos also had to look as though they actually belonged to the human in the photo, rather than part of the background. I also eliminated any images that obviously involved studio models being photographed for commercial reasons. Finally the human being had to be an adult male or female.
The dogs' sizes were estimated by their obvious size in the photo or by breed considerations. I used four size categories: small (Pugs and Pomeranians are typical), medium-small (Shetland Sheepdogs and spaniels are typical), medium-large (Golden retrievers and Labrador retrievers fit here), and large (Saint Bernards and Newfoundlands would go here).
In total I scored 600 photos (300 for each sex). The photos were selected randomly based on the order in which they appeared in the image database, eliminating all duplications in which the same recognizable person is simply appearing in different poses with the same dog.
The crude tabulation of dog size by sex is, for females: 33% small, 43% medium-small, 19% medium-large and 4% large. For males the results were: 16% small, 42% medium-small, 36% medium-large and 7% large. Looking at unprocessed data is seldom all that informative. When I did an overall global analysis it showed that there was a statistically different pattern in the preferred size of dogs chosen by men and women*. However, to make sense of this finding we have to break down the results a bit more.
To answer the overall question of whether there are sex differences in the preference for large versus small dogs, I created a new grouping which included small dogs plus medium-small dogs, and contrasted that to a grouping of medium-large and large dogs combined. Using this we find that nearly half again more women prefer smaller dogs compared to men (76.3% versus 57.7%) while men were nearly twice as likely to own larger dogs (42.3% versus 23.7%) and these differences are statistically significant**. This certainly seems to confirm the stereotype that women prefer smaller dogs than do men.
However there is a little bit more to the story. According to this data set, women are twice as likely as men to prefer the smallest grouping of dogs compared to all other sizes (33.3% versus 15.6%***) while men were twice as likely as women to prefer medium-large dogs compared to women (35.6% versus 19.3%****). There were no significant sex differences for medium-small sized dogs or for large dogs.
So on the basis of these data we can say that the stereotype that women prefer small dogs while men prefer a large dogs is mostly true. Certainly women prefer the smallest size category of dogs to a greater extent, while men prefer medium-large dogs to a greater extent than women. However these data also show that, overall, the most popular size dog, regardless of the owner's sex, are the medium-small dogs.
Copyright SC Psychological Enterprises Ltd. May not be reprinted or reposted without permission
* Chi Square for the 4 x 2 contingency table = 33.67 (4 df), p <0.001
** Chi-Square = 23.6 (1 df), p <0.001
*** Z=5.04, p<0.001
**** Z=4.48, P<0.001
