3.5 Stars or 3.5 Points? Consumers Perceive the Same Rating Differently

A hotel has a score of 3.5.

One platform displays this as three full stars and one half star.

Another simply writes 3.5/5.

The mathematical value is the same. However, the mental representation for the consumer may not be.

The study published by researchers Deepak Sirwani, Manoj Thomas, and Srishti Kumar from Cornell SC Johnson College of Business in the Journal of Marketing Research shows that the way a rating score is visualized can systematically affect how consumers perceive that score.

The importance of this research for the hospitality sector is clear: hotels, restaurants, and booking platforms have been using the same 1–5 scale for years, but the same score may not produce the same quality signal when presented in different formats.

Stars pull up, numbers pull down

The main finding of the study emerges with fractional scores.

For example, when a consumer sees 3.5 stars, they tend to perceive this value as closer to 4.

However, when the same score is shown as 3.5 with just numbers, it can feel closer to 3.

The researchers propose two different cognitive mechanisms.

With stars, the visual-completion effect comes into play.

The brain, seeing three full stars followed by a half star, can mentally activate the missing visual. Therefore, the half star creates a perceptual pull upwards.

In numbers, however, the whole number on the left can serve as a stronger anchor.

The “3” in the value of 3.5 can pull the evaluation downwards.

The difference in half stars is quite pronounced

In one of the experiments reported by Cornell, participants were shown:

1.5,
2.5,
3.5
and 4.5

half-point scores.

Ratings in star format were rounded up 79% of the time.

When the same values were shown as numbers, participants showed a tendency to round down 71% of the time.

In other words:

3.5 stars → can approach 4 in the mind.

3.5 as a number → can approach 3 in the mind.

The business's review score does not change.

But the quality expectation created by the consumer can change.

There is a source discrepancy regarding the number of experiments in the academic article

Here, there is an important source detail from an editorial perspective.

Tourism Newspaper and Cornell Chronicle report the research based on six experiments.

In contrast, the current summary of the article in the Journal of Marketing Research states that the findings were obtained from 12 experiments.

This discrepancy likely arises from different counts of studies/additional experiments in the article or the news text focusing on the main experiments of the research. Since the reason for this discrepancy is not clearly explained in the sources, "six experiments as the exact number" should not be used.

The fundamental result is consistent across sources: there is a tendency for fractional stars to be perceived as high, while numbers are perceived as low.

Why is this important for hotels?

A hotel's online reputation is no longer just about the number of reviews and the average score.

The same score:

can be presented differently on:
OTA,
Google,
the brand's website,
metasearch platforms,
mobile applications

For example, if a hotel's score is 3.5/5, a platform using a half-star visual may create a more positive perception of quality.

In contrast, an interface that simply states "3.5" may make the same score feel weaker.

This becomes particularly important in search results where consumers scan numerous establishments in a matter of seconds.

Score format should not be confused with the star class of the establishment

There is another problem in the hospitality sector.

A "4-star hotel" is not the same as a "hotel that received 4 stars from customers."

The former is the official or industry classification of the establishment.

The latter is user evaluation.

When both are displayed with stars on the screen, the consumer may confuse which symbol means what.

Therefore, hotels and booking platforms should:

clearly separate
the
hotel class
and
guest review score

visually.

Otherwise, the perceptual advantage of the score format can turn into information confusion.

If both formats are used together, numbers may dominate

The summary of the research reported by Tourism Newspaper indicates that when stars and numbers are shown simultaneously, consumers may give more weight to the numerical value.

This result is significant from an interface design perspective.

For example:

★★★★½ 4.5/5

When two systems are used side by side, the visual completion effect created by the stars may diminish due to the certainty signal of the numbers.

Thus, the platform's decision to "show both, it will be more explanatory" is not always neutral.

The perception can shape itself according to which signal the user processes more strongly.

Star advantage should also be considered ethically

The result of the research may initially provide a simple suggestion to marketers:

"Use stars, the score will be perceived as higher."

However, the issue is more sensitive than that.

The researchers also point out the potential for existing rating formats to mislead consumers.

If a business's goal is to lead users to expect a higher quality than they actually provide, a short-term increase in conversion may yield negative results in the long run.

Because:

expectations rise,
experience remains the same,
disappointment grows,
new review scores may drop.

Therefore, rating design is not just about click optimization; it is about expectation management.

Can the visual completion effect be reduced?

The researchers indicate that designs that reduce the visual perception of the half star as "missing" could decrease the overestimation effect. Cornell Chronicle reports that the use of visually completed half stars is one possible solution.

The Tourism Newspaper article also states that a simple visual change can reduce the perception of excessive height by approximately 86%.

This result is significant for booking platforms.

If the goal is to guide users as close as possible to the true value of the score, attention should be paid not only to the score calculation algorithm but also to the design of the icon displaying the score.

Not a new KPI for reputation management teams, but a new control point

The conclusion hotels should draw from this research should not be to "constantly change the score format."

What can actually be done is more practical.

The business can monitor the following in areas it controls on its own site:

How is the guest score displayed?

Does it get confused with star classification?

Is the scale of the review score clear?

What maximum is the 3.5 calculated from?

Are mobile and desktop using the same format?

Does the format change throughout the booking journey?

If a consumer sees 4.3/5 on one page and 8.6/10 on another, they may evaluate the same performance differently in their mind.

Consistency here turns into a matter of trust.

The same issue applies to restaurants

The research is not only conducted for hotels; it addresses consumer product and service ratings in general.

Therefore, it raises a valid design question for restaurants as well.

For example:

If Google shows 4.4,
the booking platform shows 8.8/10,
and the restaurant's site shows 4.5 stars,

even if the consumer encounters mathematically similar results, they may not perceive them all the same way.

It is healthier for the restaurant to specify the scale as clearly as possible in the channels it can control and not to choose different systems manipulatively.

The issue is bigger for OTAs

OTAs and review platforms influence millions of user decisions.

Therefore, when the rating format is changed at the platform level, a small perceptual effect can spread to millions of booking decisions.

For this reason, the researchers' call for industry standards is important.

Rating systems should be presented with:

the same scale,
clear maximum value,
consistent graphics,
accessible design

to help consumers compare different products more healthily.

What can hotels do in Turkey?

Hotels in Turkey cannot change the rating format on OTAs.

However, they have more control over their direct booking channels.

On the hotel's site:

the source of the review score should be clearly stated,
the score scale should be visible,
the official star class should not be shown like the review score,
the same system should be used on different pages.

For example, an expression like "4.6/5 — 1,238 verified guest reviews" can create a clearer reference for the consumer than a symbol.

Stafflink's Opinion

In our view, the most valuable result of the research for the hospitality sector is not "if you use stars, you will sell more"; rather, it shows that design is not neutral.

A review score can be a mathematical data point.

But the consumer never processes that data solely mathematically.

Color, icon, half star, scale, and number format can all affect quality expectations.

Therefore, the goal of hotels and restaurants should not be to inflate expectations with the rating format, but to ensure that the customer understands the actual score as accurately as possible.

Because a small perceptual advantage at the booking stage can turn into a heavier review cost when the experience does not meet it.