<!-- Roch Dog research paper, full text. Canonical page: https://rochstandard.com/research/what-dog-owners-will-pay-for/
     PDF: https://rochstandard.com/research/what-dog-owners-will-pay-for/what-dog-owners-will-pay-for.pdf
     DOI: https://doi.org/10.5281/zenodo.23624855
     Cite as: Bule, G. (2026). What Dog Owners Will Pay For in a Hotel. Roch Dog. https://rochstandard.com/research/what-dog-owners-will-pay-for/ (DOI 10.5281/zenodo.23624855)
     Publisher: Roch Dog, https://rochstandard.com. Research index: https://rochstandard.com/research/ -->
# What Dog Owners Will Pay For in a Hotel

**Author:** [Guise Bule](https://www.linkedin.com/in/guisebule/), CEO of [Roch Dog](https://rochdog.com)
**Subtitle:** A Hotelier's Guide to the Kyung Hee University Study
**Document Reference:** RDWP-WTP-01
**Version:** 1.1
**Date:** 8th October 2026

---

## Executive Summary

In June 2023 a team of five researchers at Kyung Hee University in Seoul published the most useful piece of academic research a hotelier can read before spending a single dollar on guests who travel with a dog. The paper is [How much are you willing to pay when you travel with a pet? Evidence from a choice experiment](https://doi.org/10.1080/13683500.2023.2223912), written by Yeonwoo Joo, Yunkyoung Jo, Hyeona Jo, Hwi-eun Choi and Yoo-shik Yoon of the university's Smart Tourism Education Platform, and it appeared in [Current Issues in Tourism](https://www.tandfonline.com/journals/rcit20), a peer reviewed journal published by Taylor and Francis and ranked as an A journal on the Australian Business Deans Council list. The work was funded by the BK21 FOUR programme of the National Research Foundation of Korea, and the authors declare no conflict of interest: nobody who sells dog beds or hotel rooms had any hand in it.

The question the team set out to answer is the one every general manager asks when a dog programme reaches the budget: of all the things a hotel could offer a guest travelling with a dog, which ones would that guest actually pay for, and how much? They answered it with a choice experiment, the method economists use to price things that are not yet on sale. Rather than asking 491 Korean dog owners what they wanted, they showed each of them a series of complete hotel packages at different prices and asked which they would book, and from thousands of those choices worked backwards to the value of every feature.

The results are the reason this guide exists. Owners would pay about $117 a night more for a hotel with a dedicated play area for dogs, a dollar more if it came with a restaurant the dog could eat in, $66 a night more for a lounge where staff would mind the dog while they went out, and $146 a night more for the same lounge with a certified dog behaviour specialist on duty. They would pay $57 a night less for a hotel that added a dog bed and steps to the room on top of a bowl and toilet supplies, and $60 less if it added a welcome gift of treats and toys as well. When the authors weighed each factor in the booking decision, the facilities took 30 per cent, the services 27 per cent, the contents of the room 15 per cent and the price of the room less than one per cent, and they concluded that owners who find the facilities and services good enough stay regardless of the price.

Read together, those numbers say something the industry has not wanted to hear. Everything a dog owner valued was outside the bedroom, in a space built for the dog or a person trained to look after it, and everything they marked down was inside it, in the basket most hotels think of as their dog offer. The study was Korean and most of its dogs were small, and the authors say the figures should be tested elsewhere before they are relied on, but the order of the findings is clear.

---

## 1. Who Did the Research and Why

The five authors are researchers at the Smart Tourism Education Platform, known as STEP, a graduate research unit within Kyung Hee University's College of Hotel and Tourism Management in Seoul. Kyung Hee is one of the best known hospitality schools in Asia and STEP is its applied research arm, and the team acknowledge the advice of three of its senior professors, Myung Ja Kim, Insik Min and Seokho Han. The corresponding author is [Yoo-shik Yoon](mailto:ysyn@khu.ac.kr), who can be reached through the university. Two of the team can be found on LinkedIn: [Hwi-eun Choi](https://www.linkedin.com/in/sylviahwieunchoi/), now a research professor at Kyung Hee and the author who placed the free copy of the paper online, and [Yeonwoo Joo](https://www.linkedin.com/in/kyla-yeonwoo-joo-9397b1156/), who has since moved into tourism and MICE research at the Goyang Research Institute. We were not able to confirm public profiles for the other three authors and have not guessed at them.

The research was paid for by the BK21 FOUR programme of the National Research Foundation of Korea. BK21, short for Brain Korea 21, is the government's long running scheme for funding graduate schools and the research their students produce, and FOUR its fourth phase. This matters to a hotelier reading the results because it means the study was commissioned by nobody with a product to sell. It was not paid for by a hotel group, a pet retailer or a tourism board, and the authors declare no conflict of interest.

The paper was received by [Current Issues in Tourism](https://www.tandfonline.com/journals/rcit20) on 6 March 2023, accepted on 4 June after anonymous review by at least two experts, which is the journal's stated policy for every article, and published online on 13 June 2023. It was later assigned to volume 27, issue 13, at pages 2118 to 2133. The publisher charges for access, but Hwi-eun Choi has placed the full text on [ResearchGate](https://www.researchgate.net/publication/371544731_How_much_are_you_willing_to_pay_when_you_travel_with_a_pet_Evidence_from_a_choice_experiment), where it can be read for nothing. If you take one thing from this guide, read theirs.

The authors explain why they did the work in terms a hotelier will recognise. South Korea had just over three million households living with a pet when the study was designed, three quarters of them with a dog, and the Korea Rural Economic Institute expected the country's pet market to grow from 1.9 trillion won in 2015 to 6 trillion won by 2027. The Korea Tourism Organization's 2022 pet travel survey had found that 53 per cent of owners who travelled with their pet had taken an overnight trip in the previous year, that 74 per cent of all owners wanted to travel with their pet in future, and that the single biggest obstacle in their way was the shortage of places to stay, eat and visit that would take the animal. Yet when the team went looking for research on what that accommodation should actually provide, they found almost none: the academic literature had studied why owners want to travel with their pets and what holds them back, but nobody had measured what they would pay for once they arrived. The study was written to fill that gap. Its title says pets because that is the word the literature uses, but every one of the 491 people who took part owned a dog, so this guide says dog owners.

## 2. How the Experiment Worked

Ask a guest what they want from a hotel and they will tell you everything. Ask them what they would pay for it and they will guess, and usually guess high, because nothing is at stake. A choice experiment avoids both problems by never asking the question directly. It shows people a small number of complete hotel packages, each with a different combination of features and a different price, and asks which one they would book, or whether they would book none of them. When enough people make enough of those choices, the pattern reveals how much each feature is worth, because the researchers can see how much extra price people will accept in order to get it. The authors chose this method over its older cousin, contingent valuation, which asks about one feature at a time, precisely because a hotel is a bundle and a guest weighs its parts against each other.

The first job was deciding what to test. The team started from the nine things the Korea Tourism Organization's survey had found owners consider when choosing somewhere to stay, which run from whether pets are admitted at all through the dining environment, entertainment for the dog, nearby attractions, the room, accessibility, amenities and in-house pet care to cost. In the first week of April 2022 they interviewed fourteen people, ten dog owners chosen across ages and both sexes and four operators of accommodation that takes dogs, about how they travel, what they look for and how the properties are run. From those interviews three attributes survived, facilities, amenities and services, each with three levels. To set the price levels they asked a further thirty owners the least and the most they would pay for a night; the averages came back at $116 and $220, so the room price was set at $114, $171 and $228 a night. The whole design was then reviewed by tourism academics before it was used.

The levels are worth setting out exactly, because the findings only make sense against them. For facilities a hotel could offer nothing for dogs, a play area reserved for dogs such as a playground or a swimming pool, or that play area plus a restaurant inside the hotel where the dog could come too. For amenities, meaning what is left in the bedroom, a hotel could provide the basics of toilet supplies and a bowl, the basics plus a bed and a set of steps, or all of that plus a welcome gift of treats and toys. For services it could provide nothing, a pet lounge, defined as a care service and space where an owner can leave the dog and go out, or that lounge staffed by a certified and experienced dog behaviour expert. The basic amenities were the floor of the design rather than something to be tested, so every hotel an owner looked at already had a bowl and toilet supplies in the room, and the question was whether adding more was worth anything.

Four attributes with three levels each give 81 possible hotels, far too many to show anyone, so the team used a statistical procedure in SAS 9.4 called an orthogonal design to pick the nine combinations that would reveal the most with the fewest questions, a set the software rated at full efficiency. Those nine were arranged into choice sets of three, and every set included the option of booking none of them, so nobody was forced to pick a hotel they did not want. Three versions of the questionnaire were produced to stop any one ordering of the options colouring the answers, with three choice questions in each, and the whole thing was tested on four tourism experts and forty dog owners between 25 November and 2 December 2022 before it went live.

The survey itself ran online from 27 December 2022 to 2 January 2023 through Embrain, a Korean research company with a panel of 1.64 million people. Respondents had to be dog owners aged twenty or over, the age of legal majority in Korea, and the sample was drawn by quota so that its mix of sex, age and home region matched the pet owning households recorded in the 2020 national census. The questionnaire gathered basic facts about the respondent and the dog, then the choice questions, with scenarios and definitions provided so that everyone understood what each feature meant. It collected 495 responses, dropped four whose reported number of pet trips in a year was implausibly high, and analysed the remaining 491 in STATA, which across three questions and three options each gave the model 4,419 observed decisions to work from.

## 3. Who Took Part

The sample is the thing to understand before reading any of the numbers, because a willingness to pay belongs to the people who expressed it. Seven in ten respondents were women and three in ten men, which the authors say reflects the population of Korean dog owners. A little over half were single, two in five married, and 56 per cent lived in the capital region of Seoul, Incheon and Gyeonggi, with the rest spread across the provinces. In age they ran from the twenties to the fifties and beyond: 29 per cent were in their twenties, 34 per cent in their thirties, 25 per cent in their forties and 11 per cent fifty or older. Household income was spread across the range, with a third earning between two and three million won a month, which at the exchange rate the authors used of about 1,316 won to the dollar is roughly $1,500 to $2,300.

The dogs themselves were overwhelmingly small. Eighty five per cent weighed under ten kilograms, which is about 22 pounds, 12 per cent were medium sized at ten to 25 kilograms, and only 3 per cent, sixteen dogs in all, were large. Most were in their prime years, with 41 per cent aged five to nine and 30 per cent aged one to four, though more than a quarter were ten or older. Their owners were spending an average of 205,000 won a month on them, about $156, and a quarter were spending more than 210,000 won. This is a sample of people who take their dogs seriously, and the proportion of small breeds reflects urban Korean dog ownership rather than the Labradors and spaniels a British or American hotel is more likely to meet, a point we return to in section 8.

Three quarters of the respondents had travelled with their dog at least once, and a quarter never had. Of those who had, most had done it only once in the past year, about a quarter had done it two or three times and fewer than one in ten had done it four times or more. In other words this was not a sample of seasoned dog travellers with settled preferences; it was a cross section of ordinary owners, most of whom had tried travelling with the dog and found it hard enough not to do it often, which is exactly the group a hotel hoping to grow this market needs to understand.

## 4. What They Found

The model produced two kinds of result. The first is the set of coefficients, which say whether each feature made a hotel more or less likely to be chosen and by how much; every one of them was statistically significant at the strictest conventional level, and the model as a whole fitted the data well, with a McFadden pseudo R squared of 0.315, which in discrete choice work is considered a good fit. The second is the marginal willingness to pay, which converts each coefficient into dollars per night by setting it against the coefficient on price. The table below sets out the six prices the study produced. Each is the extra amount per night an owner would pay for a hotel with that feature compared with the base hotel, which had no dog facilities, no dog service, and only a bowl and toilet supplies in the room. The last column is the range within which the authors are 95 per cent confident the true figure sits.

| What the hotel added | Worth per night | Range |
|---|---|---|
| A play area for dogs | +$117 | $84 to $162 |
| A play area and a restaurant that takes dogs | +$118 | $84 to $166 |
| A lounge where the dog is looked after | +$66 | $38 to $104 |
| The same lounge, run by a trained dog specialist | +$146 | $109 to $151 |
| A bed and steps in the room | -$57 | -$62 to -$51 |
| A bed and steps, plus treats and toys | -$60 | -$67 to -$49 |

The first thing that stands out is the line between the bedroom and the rest of the hotel. Everything with a positive price is outside the room: the play area, the lounge and the person running it all added value, and together they are the whole of what these owners were prepared to pay for. Everything with a negative price is inside the room. Adding a bed and steps to the bowl and toilet kit made owners pay less, and adding treats and toys on top of that made them pay a little less again. The confidence ranges on the amenity figures are tight, which means the sample was not divided on the point; it is not that some owners loved the basket and others hated it, it is that the basket consistently lowered the value of the hotel that offered it.

The second thing is that the person was worth more than the room they stood in. A lounge on its own, somewhere staff would mind the dog while the owner went out, was worth $66 a night. The same lounge with a certified dog behaviour specialist in charge was worth $146. The subtraction is ours rather than the authors', but it is hard to miss: a trained member of staff added $80 a night to the value of the space, more than the space itself was worth. The authors' explanation is that owners worry about what can go wrong when strange dogs share a space, a bite or a fight above all, and the presence of someone competent to prevent it is what turns a room into a service they will pay for.

The third thing is how little the price of the room mattered. When the authors worked out the relative weight of each attribute in the decision, facilities carried 30.2 per cent, services 27.2 per cent, the contents of the room 15.1 per cent and the room rate 0.3 per cent. Price did behave as economics says it should, in that a higher rate made a hotel less likely to be chosen, but the effect was so small next to the others that the authors concluded owners who find the facilities and services sufficient "decide to stay regardless of the price". Within the attributes, a lounge with a specialist was the single most important feature at 18.8 per cent, followed by a play area with a restaurant at 15.2 per cent and a play area alone at 15.0 per cent. One further detail is easy to miss but tells the same story: the constant in the model that stands for booking none of the hotels on offer was positive and significant, meaning that faced with hotels that did not provide enough, these owners were quite prepared to walk away and book nothing at all.

## 5. Why More in the Room Was Worth Less

A free dog bed that lowers what a guest will pay looks like a mistake in the data, and it is the finding hoteliers push back on first, so it deserves its own section. The authors offer an explanation, and anyone who lives with a dog will recognise it at once. A hotel dog bed, bowl or toy has been used by other dogs. However thoroughly it is laundered it carries the scent of the last animal, and many dogs will not go near it, so their owners have learned to pack the dog's own bed and bring its own bowl. Food and treats raise a second problem, because a dog with an allergy or a sensitive stomach cannot be given whatever the hotel has put in a bag, and an owner who has spent years finding a food that works is not going to experiment on holiday. The welcome basket, in other words, solves a problem the guest has already solved, and a thoughtful owner may read it as a sign the hotel has not thought very hard.

It is important to be precise about what the study did and did not measure here, because the point is easily overstated. Every hotel in the experiment, including the base hotel against which everything else was priced, already provided a bowl and toilet supplies in the room. Nothing in these results is a case against the water bowl, and the [Roch Dog Friendly Standard](https://rochstandard.com) treats real food and water bowls in the room as one of its six minimum requirements for the same reason the researchers built them into every option: a hotel that cannot manage a bowl has not started. What the study measured was the value of adding more, first a bed and steps and then a gift of treats and toys, and the answer was that each addition lowered the price owners would pay. The authors' explanation of why is their reading of the result and was not itself tested; what the data shows is only that these owners were not expecting more in the room and would not pay for it.

Nor does it follow that a hotel should strip its rooms. A dog bed costs little, it keeps the dog off the hotel's own linen and furniture, and it tells the guest on arrival that the dog was expected rather than tolerated, which is the whole difference between a hotel that allows dogs and one that welcomes them. The bed, the bowl and the bag of treats are the floor of a dog programme, and the study's real lesson is that nobody pays extra for a floor. A hotel whose entire offer to a dog owner is a basket in the bedroom is spending its dog budget on the one part of the stay these guests valued at less than nothing, and it is doing so while the parts they would pay for, somewhere for the dog to go and someone to look after it, remain unbuilt.

There is a wider industry pattern behind this. When Roch Dog read 236 published hotel pet policies for [The Fine Print](https://rochstandard.com/research/the-hotel-pet-policy/), only one policy in five mentioned anything as basic as a water bowl, while most led with a list of what the dog was forbidden to do and where it was forbidden to go. The owners were asking for places and people; the policies were describing restrictions and, where they offered anything at all, kit. The Kyung Hee numbers put a price on that gap for the first time.

## 6. Three Kinds of Owner, One Answer

Having priced the features for the sample as a whole, the researchers asked whether different kinds of owner wanted different things, which is the question a marketing department would ask next. They used a latent class model, a statistical method that sorts respondents into groups according to how they chose rather than according to anything they said about themselves, and then looks at what distinguishes the groups. The method lets the data decide how many groups there are; the team tested one, two and three classes, and the standard measures of fit both pointed to three. The variables they allowed to influence which class a person fell into were whether they had travelled with a pet before, how much they spent on the dog each month, how many pets they had, how old the dog was and their income.

The first class, just under a quarter of the sample, were owners for whom travelling with the dog mattered a great deal and who spent heavily on the animal every month; the authors call them the preferred group. The second and largest class, a little under half, were owners who were sceptical about travelling with a dog at all because of the practical constraints, above all how many pets they had and how old those pets were, and whose past travel experience made no difference to their outlook; this is the non-preferred group. The third class, a little over a quarter, felt the same constraints as the second but regarded travelling with the dog as important regardless, and are named the restricted group. Income and age shaped membership of every class: the preferred group tended to be older and living in the capital region, the non-preferred group younger, unmarried and living outside it, and the restricted group older, married and better off than the sceptics, while sex made no difference anywhere.

The finding that matters to a hotel is what did not differ. Across all three groups the preferences for hotel features were essentially the same, and the only measurable gap was that the restricted group leaned a little more towards a play area than the sceptics did. The authors' explanation is sobering for the industry. There are so few places that take dogs properly, they argue, that owners are not in a position to be choosy; when accommodation is scarce an owner does not weigh the optional features of a hotel, they look for one that will have the dog at all and take what they can get. The three groups want the same things because the market has never given any of them the chance to want different things.

For a hotel the practical reading is simple. There is no need to guess which kind of dog owner you are serving, and no need to design one offer for the devoted and another for the doubtful. The enthusiast, the sceptic and the determined owner all put their money in the same place, on space and on people, and a hotel that provides both is serving the whole market at once.

## 7. What It Means for Your Hotel

This section is our reading, and we separate it from the authors' own conclusion, which is short and which we agree with entirely. A hotel that wants these guests, they write, should invest in services and in space dedicated to dogs, in the staff and the programmes that let a guest have a good time with the dog without leaving the property, and should stop putting its money into supplies for the bedroom. That is the whole of their advice. What follows is what we think it means in practice for a hotel in Britain, Europe or North America rather than Seoul, drawn from the study and from the hotels Roch Dog has assessed.

Keep the basics and stop adding to them. The bowl stays, because every hotel in the study had one and the standard requires it, and the bed stays, because it protects your room and signals a welcome. What should go is the escalation: the second toy, the branded bandana, the upgraded hamper with the artisan treats, the things a marketing team adds each year because the basket is the only part of the dog programme it controls. That money is being spent on the one part of the stay these guests marked down. Move it.

Give the dog somewhere to go. In the experiment the space was a play area set aside for dogs, which a resort with grounds can build and a city hotel without them cannot, and the study did not test what a city hotel might offer instead. Our reading, and it is a reading rather than a finding, is that the play area and the lounge were valued because they got the owner and the dog out of the bedroom and into the life of the hotel, and the authors point the same way when they describe the frustration of owners in hotels that accept dogs but give them nowhere to be, so that they either stay in the room or go looking for somewhere outside. The [Roch Dog Friendly Standard](https://rochstandard.com) makes access to at least one shared indoor space a minimum requirement for the same reason. A terrace, a corner of the garden, a lounge or a bar where the dog can lie under the table while the owner has a drink does the same job for a hotel with no room for a playground, and it does it with the owner spending money at the hotel's own tables rather than at the restaurant down the street. The revenue side of that argument is set out in [The Economics of Dog Friendly Hospitality](https://rochstandard.com/research/economics-of-dog-friendly-hospitality/).

Put a person in charge. The single largest figure in the study belongs not to a facility but to a trained member of staff, and the authors are explicit that owners valued that person because they worried about what could go wrong in a shared space and wanted someone competent on hand to prevent it. A hotel does not need a pet lounge to act on that. It needs one named person on each shift who knows dogs, who can answer a question at check in without consulting a binder, who can tell a nervous owner where the dog can and cannot go, and who can sit with or walk a dog when the guest wants an hour to themselves. That person is the difference between a hotel that has a policy and a hotel that has a programme, and the study suggests guests can tell.

Compete on what you offer, not on the rate. The price in the experiment was the price of the room, not a dog fee, so nothing here is an argument for a higher fee, and the question of what a fee should buy is dealt with in [The Canine Hospitality Playbook](https://rochstandard.com/research/canine-hospitality-playbook/). What the study does show is that owners chose on facilities and service and treated the room rate as close to irrelevant once those were right. A hotel that discounts to win dog owners is cutting the one thing they were not looking at, and a hotel that holds its rate and opens its doors is doing what the data says works.

Decide which dogs stay on behaviour, not size. The owners in this study mostly had small dogs, and a hotel might be tempted to read that as support for a weight limit. It is nothing of the kind; it is a description of Korean city life. The [Friendly Dog Policy](https://rochstandard.com/research/the-hotel-pet-policy/) that Roch Dog publishes for any hotel to use replaces breed lists and weight limits with one fair test of how the dog behaves, which is the only test that tells a hotel anything about the dog in front of it.

Make the space an event. If a hotel does open a terrace or a garden to dogs, the study's logic says the value lies in the owner being able to enjoy the hotel with the dog, and the properties that do this best turn it into something people talk about. The InterContinental Miami's [Paws Patio](https://rochsociety.com/dog-friendly-review-intercontinental-miami-hotel/), a free Wednesday evening gathering on the hotel's waterfront terrace that has run for more than five years, with drinks for the owners and cooling fountains for the dogs, is the model: a shared space that became a reason to book and a photograph people share, which no welcome basket has ever been.

## 8. What the Study Does Not Tell You

One study never settles a question, and this one has limits that its authors set out themselves and that a hotelier should hold in mind before quoting any of its figures. The most obvious is that it was carried out in South Korea. The authors say plainly that owners in other countries, with different cultures and different attitudes to companion animals, may value different things, and that the work needs repeating wherever a hotel intends to rely on it. The right way to read the dollar figures is as evidence of what owners value and in what order, not as a price list that transfers to London or Los Angeles, and that is how Roch Dog has used them in its own research.

The second limit follows from the first. Most of the dogs in the sample were small, 85 per cent of them under ten kilograms, and only sixteen of the 491 were large. An owner travelling with a retriever or a shepherd may put more value on outdoor space and less on a lounge than these figures suggest, and the study simply cannot say. That is a gap in the evidence and not, as section 7 explains, a case for a weight limit.

The third is the nature of the method. A choice experiment records what people say they would choose when shown a hypothetical hotel, not what they did when their own money was on the table. It is the accepted way to price things that do not yet exist, and the authors took the usual precautions, including the option to book nothing and a design that stops respondents being led, but a real booking with a real card can still come out differently, and the only way to know is for a hotel to try it.

The fourth is scope. The experiment tested four things and only four. It did not test dog fees or deposits, weight limits, the wording of the pet policy, or access to the bar, the lobby and the lounge alongside guests without dogs; the spaces it priced were reserved for dogs, which is a different proposition from a shared one. Nor was the restaurant ever offered on its own: it appeared only on top of a play area, where it added a single dollar, so the study cannot say what a restaurant that welcomes dogs is worth to a hotel that has nothing else, and the small difference may say as much about how the option was presented as about restaurants. The study also looked only at hotels, and the authors call for the same work to be done on transport, travel distance and other kinds of accommodation.

The fifth is that it is a study of the guest's side alone. The authors note that the cost to a hotel of providing a lounge or a specialist, and the return it earns, still need measuring from the operator's side, and that consumer preferences shift quickly enough that the survey should be repeated periodically. Roch Dog has approached that side from its own data in [The Economics of Dog Friendly Hospitality](https://rochstandard.com/research/economics-of-dog-friendly-hospitality/) and set out the practical steps in [The Canine Hospitality Playbook](https://rochstandard.com/research/canine-hospitality-playbook/), but the controlled experiment that would pair this study's demand figures with measured operating returns has not yet been done, and it is the single most useful piece of research the industry could commission.

None of these limits weakens the central finding. A result that owners will pay for places and people and will not pay for more things in the room is not the kind of result that depends on the exchange rate, and it is consistent with what Roch Dog finds in the hotels it assesses, where the properties that rank highest are the ones that open their doors rather than fill their rooms. What the limits do is tell a hotelier how to use it: as a direction, confirmed by a controlled experiment, rather than a tariff.

---

## Appendix A: How the Numbers Were Made

This appendix is for the reader who wants to understand where the dollar figures came from, or to check them. It explains the model the researchers used in plain terms, with the mathematics kept to the few lines needed to see how a coefficient becomes a price. Nothing here changes the findings; it shows how they were reached.

The model rests on two ideas from economics. The first, from Kelvin Lancaster's 1966 paper on consumer theory, is that people do not value a product as a whole but value the characteristics it is made of, so a hotel room is really a bundle of attributes and its worth is the sum of theirs. The second, from Daniel McFadden's work on choice in the early 1970s, which later won him the Nobel prize, is that when a person chooses between options they pick the one that gives them the most satisfaction, or utility, but that an outside observer can only ever see part of what drives that choice. McFadden's random utility model therefore writes a person's utility from any option as two parts: a systematic part that depends on the option's attributes and the person's own characteristics, which the researcher can model, and a random part that captures everything the researcher cannot see. The probability that someone picks one hotel over another is then the probability that the systematic advantage of the first outweighs whatever unseen factors favour the second.

If the random part is assumed to follow a particular statistical distribution, that probability takes a neat closed form known as the conditional logit: the probability of choosing a given hotel equals the exponential of its systematic utility divided by the sum of the exponentials for every hotel on offer, including the option of booking none. The researchers then find the set of coefficients, one for each attribute level, that makes the choices actually observed in the survey as likely as possible, a procedure called maximum likelihood estimation. The coefficients in the study's fourth table are those values. A positive coefficient means the feature raised the chance of a hotel being chosen; a negative one means it lowered it. The play area scored 1.417, the play area with a restaurant 1.433, the lounge 0.798 and the lounge with a specialist 1.778, while the bed and steps scored minus 0.698, the full basket minus 0.726 and each dollar of room price minus 0.009. Every coefficient was significant at the one per cent level, meaning the chance of seeing an effect that size by accident was below one in a hundred, and the model's pseudo R squared of 0.315 is a strong fit for this kind of work, where values between 0.2 and 0.4 are regarded as good.

Turning a coefficient into money is a single division. The marginal willingness to pay for a feature is its coefficient divided by the coefficient on price, with the sign flipped, because the price coefficient says how much utility a dollar costs and the feature coefficient says how much utility the feature is worth, so their ratio is the number of dollars a person would give up to have it. A reader who tries this with the printed figures will notice that 1.417 divided by 0.009 is about $157, not the $117 the authors report, and that every published price implies a price coefficient of roughly 0.012 rather than 0.009. The most likely explanation is that the price coefficient was rounded for the table and the prices were computed from the unrounded value, which is routine, but it does mean the fifth table cannot be reproduced exactly from the fourth, and it is one of the questions we would put to the authors. The paper does not say how the 95 per cent confidence ranges around each price were produced, but their width is a guide to how settled each figure is; the amenity figures are tight and the lounge figure is the loosest.

Two further numbers in the results deserve a word. The alternative specific constant, which the authors report as positive and significant, is the coefficient on the option of booking none of the hotels shown, and its sign says that respondents found the walk-away option attractive in itself, which is consistent with dog owners who are used to being disappointed. The relative importance figures, which gave facilities 30.2 per cent, services 27.2 per cent, amenities 15.1 per cent and price 0.3 per cent, were derived from the coefficients, and we report them as the authors published them.

The latent class model in the second half of the paper extends the same logic by allowing for the possibility that respondents do not all share one set of coefficients but fall into a small number of groups, each with its own. Within a group preferences are assumed to be alike, and between groups they differ. The model estimates two things at once: a set of choice coefficients for each group, and a membership equation that gives the probability of a person belonging to each group as a function of their characteristics, here their past pet travel, monthly spending on the dog, number of pets, the dog's age and their income. Because the number of groups has to be chosen by the researcher, the team fitted the model with one, two and three classes and compared them using two standard penalties for complexity, the Akaike and Bayesian information criteria, which reward a better fit but punish extra parameters; both reached their lowest value at three classes, so three was used. The class sizes of 24.85, 47.86 and 27.29 per cent are the model's estimates of how the sample divides, and the socioeconomic patterns in section 6 come from a further regression of class membership on income, age, region, marital status and sex.

One technical point in that second model matters for how the result is read. When the authors tested whether the three classes differed in their preferences for the hotel features, almost none of the differences reached significance, and the fit of that comparison was very weak. They interpret this through a concept from earlier choice research called random preference, in which a respondent facing options that all fall short on the thing they care about most simply chooses on whatever comes next, so that the measured preferences look alike across groups not because the groups are the same but because the options never let their differences show. That is the statistical basis for the authors' conclusion that scarcity, rather than taste, explains why every kind of owner wanted the same things.

## Appendix B: The Research It Builds On

The Kyung Hee study did not appear from nowhere, and a hotelier who wants to go further will find the earlier work it rests on short and readable. [Kirillova, Lee and Lehto's 2015 survey](https://doi.org/10.1080/1528008X.2015.966296) of American consumers, in the Journal of Quality Assurance in Hospitality and Tourism, established that the decision to bring the dog is shaped by what owners expect to find at the destination, and that hotels with pet programmes could use them to stand apart; the Kyung Hee team cite it for the business case. [Carr and Cohen's 2009 paper](https://doi.org/10.1057/thr.2009.10), pointedly titled Holidaying with the family pet: no dogs allowed!, documented the gap between what owners want and what accommodation provides, and [Ying and colleagues' 2021 study](https://doi.org/10.1016/j.tourman.2020.104183) in Tourism Management described how repeated refusals teach owners a kind of learned helplessness about travelling with the dog, which is the condition the Korean survey's non-preferred group displays.

[Taillon, MacLaurin and Yun's 2015 research note](https://doi.org/10.1080/13032917.2014.942327) in Anatolia was an earlier attempt to measure willingness to pay for travelling with a pet, and [Hung, Chen and Peng's 2016 work](https://doi.org/10.1177/1096348013503992) in the Journal of Hospitality and Tourism Research examined what taking a dog on holiday means to the owner. The choice experiment method itself is set out for tourism readers in [Boto-García and colleagues' 2022 paper](https://doi.org/10.1177/1354816620959901) in Tourism Economics, which the Kyung Hee team used as their model, and [Zhang, Grisolía and Lane's 2023 study](https://doi.org/10.1016/j.tourman.2022.104692) in Tourism Management applied the same method to hotel demand. The latent class treatment follows [Yoo and Ready's 2014 paper](https://doi.org/10.1016/j.eneco.2013.12.007) in Energy Economics, from which the idea of random preference is taken. The Korean demand figures come from the Korea Tourism Organization's 2022 Pet Travel Survey and the market forecast from the Korea Rural Economic Institute's 2018 strategy for the companion animal industry, neither of which is published in English.

Since this guide was first published the study has been taken up by the field. [Koufodontis (2026)](https://doi.org/10.1016/j.ijhm.2026.104929), in the International Journal of Hospitality Management, surveyed 1,041 travellers in 66 countries on what makes a hotel genuinely pet friendly and found the order the Kyung Hee team found: the right facilities first, named by 78 per cent, clear rules and trained staff next, and amenities last of the four, at 71 per cent. The author's conclusion, citing the Kyung Hee experiment directly, is that amenities matter mainly as complements once access and facilities are secured, and that a hotel should meet and disclose its baseline before it spends on extras. A result from Seoul, which section 8 cautions may not travel, has now been matched by a sample drawn from every continent.

---

## Sources

1. Joo, Y., Jo, Y., Jo, H., Choi, H. and Yoon, Y. (2023). How much are you willing to pay when you travel with a pet? Evidence from a choice experiment. Current Issues in Tourism, 27(13), 2118 to 2133. Received 6 March 2023, accepted 4 June 2023, published online 13 June 2023. Funded by the BK21 FOUR programme of the National Research Foundation of Korea. DOI: [10.1080/13683500.2023.2223912](https://doi.org/10.1080/13683500.2023.2223912). Every figure in this guide is taken from this paper unless marked as our own arithmetic or our own reading.
2. Free full text of the same paper, placed on ResearchGate by the author Hwi-eun Choi: [researchgate.net/publication/371544731](https://www.researchgate.net/publication/371544731_How_much_are_you_willing_to_pay_when_you_travel_with_a_pet_Evidence_from_a_choice_experiment)
3. [Current Issues in Tourism](https://www.tandfonline.com/journals/rcit20), Taylor and Francis. Double anonymous peer review by a minimum of two experts; ranked A on the Australian Business Deans Council journal list.
4. Author profiles: [Hwi-eun (Sylvia) Choi](https://www.linkedin.com/in/sylviahwieunchoi/), Research Professor, Kyung Hee University; [Yeonwoo (Kyla) Joo](https://www.linkedin.com/in/kyla-yeonwoo-joo-9397b1156/), Goyang Research Institute. Corresponding author Yoo-shik Yoon, ysyn@khu.ac.kr, as given in the paper.
5. Lancaster, K. J. (1966). A new approach to consumer theory. Journal of Political Economy, 74(2), 132 to 157. DOI: [10.1086/259131](https://doi.org/10.1086/259131)
6. McFadden, D. (1973). Conditional logit analysis of qualitative choice behavior. In Zarembka, P. (ed.), Frontiers in Econometrics. Academic Press.
7. Kirillova, K., Lee, S. and Lehto, X. (2015). Willingness to travel with pets: a U.S. consumer perspective. Journal of Quality Assurance in Hospitality and Tourism, 16(1), 24 to 44. DOI: [10.1080/1528008X.2015.966296](https://doi.org/10.1080/1528008X.2015.966296)
8. Carr, N. and Cohen, S. (2009). Holidaying with the family pet: no dogs allowed! Tourism and Hospitality Research, 9(4), 290 to 304. DOI: [10.1057/thr.2009.10](https://doi.org/10.1057/thr.2009.10)
9. Ying, T., Tang, J., Wen, J., Ye, S., Zhou, Y. and Li, F. S. (2021). Traveling with pets: constraints, negotiation, and learned helplessness. Tourism Management, 82, 104183. DOI: [10.1016/j.tourman.2020.104183](https://doi.org/10.1016/j.tourman.2020.104183)
10. Taillon, J., MacLaurin, T. and Yun, D. (2015). Hotel pet policies: an assessment of willingness to pay for travelling with a pet. Anatolia, 26(1), 89 to 91. DOI: [10.1080/13032917.2014.942327](https://doi.org/10.1080/13032917.2014.942327)
11. Hung, K., Chen, A. and Peng, N. (2016). Taking dogs to tourism activities. Journal of Hospitality and Tourism Research, 40(3), 364 to 395. DOI: [10.1177/1096348013503992](https://doi.org/10.1177/1096348013503992)
12. Boto-García, D., Mariel, P., Pino, J. B. and Alvarez, A. (2022). Tourists' willingness to pay for holiday trip characteristics: a discrete choice experiment. Tourism Economics, 28(2), 349 to 370. DOI: [10.1177/1354816620959901](https://doi.org/10.1177/1354816620959901)
13. Zhang, X., Grisolía, J. M. and Lane, T. D. (2023). Zero price effect on hotel demand: evidence from a discrete choice experiment. Tourism Management, 96, 104692. DOI: [10.1016/j.tourman.2022.104692](https://doi.org/10.1016/j.tourman.2022.104692)
14. Yoo, J. and Ready, R. C. (2014). Preference heterogeneity for renewable energy technology. Energy Economics, 42, 101 to 114. DOI: [10.1016/j.eneco.2013.12.007](https://doi.org/10.1016/j.eneco.2013.12.007)
15. Korea Tourism Organization (2022). 2022 Pet Travel Survey. Korea Rural Economic Institute (2018). Development strategies for the companion animal industry. Both cited by the authors; neither is published in English.
16. [RDWP-03: The Economics of Dog Friendly Hospitality](https://rochstandard.com/research/economics-of-dog-friendly-hospitality/). The revenue model for a hotel that takes dogs properly.
17. [RDWP-PB-01: The Canine Hospitality Playbook](https://rochstandard.com/research/canine-hospitality-playbook/). What to change, in what order, and what each change costs.
18. [RDWP-HPP-01: The Fine Print](https://rochstandard.com/research/the-hotel-pet-policy/). 236 published hotel pet policies, and the Friendly Dog Policy.
19. [The Roch Society: InterContinental Miami, dog friendly hotel review](https://rochsociety.com/dog-friendly-review-intercontinental-miami-hotel/). Paws Patio as described in section 7.
20. Koufodontis, N. I. (2026). Pet-friendliness in hospitality and tourism: Definition, scope, and governance. International Journal of Hospitality Management, 141, 104929. DOI: [10.1016/j.ijhm.2026.104929](https://doi.org/10.1016/j.ijhm.2026.104929). The global survey described in Appendix B. Open access.

## Published by Roch Dog
