Company, food or toys? Let the individual cat choose

Citation and access

Kristyn R. Vitale Shreve, Lindsay R. Mehrkam & Monique A.R. Udell (2017). “Social interaction, food, scent or toys? A formal assessment of domestic pet and shelter cat (Felis silvestris catus) preferences.” Behavioural Processes, 141 (Part 3), 322–328. Online March 24, 2017; August 2017 issue.

DOI · Original publisher / inspected main text · Author-lab abstract · PubMed record

Publisher-linked extraction provided substantive Introduction, Methods, Results, Discussion and Conclusions. This is a full-main-text synthesis, not an abstract-only card. Figure images and raw data were not independently audited; extraction artifacts are noted below. A separately downloadable free accepted manuscript was not successfully verified, so unrestricted browser/PDF access is not promised. No copyrighted article is reproduced here.

Why it matters — Lumi

Choosing company is worth measuring, but it is not a love-meter. This study asks which available activities cats spend time with, instead of assuming food always outranks interaction with people. It offers a practical starting point for individualized enrichment—activities that make an animal’s environment more engaging—without claiming that every cat wants the same thing.

Question and method — Lumi

Would pet and shelter cats favor human interaction, food, toys or scents when given a choice?

Researchers recruited 50 adult cats, 25 from each population. Five did not complete testing because of nervous behavior; six completed the assessment but did not interact with the stimuli. This left 39 cats for within-category comparisons. One additional shelter cat did not interact during the final comparison, leaving 38 cats, 19 from each population, for that analysis.

A free-operant preference assessment lets the animal interact freely with available options, measuring engagement time rather than requiring a trained response. Three options were tested within each of four categories. Social options were petting, human vocalization and human-led feather play, presented one at a time for one minute each to keep the person constant. Food, toy and scent options were presented simultaneously for three minutes within each category.

Each cat’s favorite available option from each category then entered a three-minute simultaneous comparison. A category with no earlier interaction was omitted for that cat. The person was the owner for pet cats and an experimenter for shelter cats. Food and social attention were withheld for 2.5 hours before testing; this was a research procedure, not a recommended home exercise.

Findings — Lumi

• In the final comparison, 19 of 38 cats favored social interaction, 14 favored food, four favored toys and one favored scent. Social interaction was the largest category, but exactly half—not a strict majority—of the analyzed cats.
• The number favoring social interaction was greater than the numbers favoring toys or scent, but social interaction versus food was not statistically significant (reported p = 0.35). The study does not establish that cats generally prefer humans over food.
• No statistically significant pet-versus-shelter differences were detected in the tested preferences. This does not establish equivalence between populations.
• Preferences varied across individuals. Measuring an individual cat’s choices is more useful than converting the group ranking into a universal prescription.

Limits and interpretation — Lumi

Selection matters: The final analysis excludes nervous cats and nonparticipants; it is not a random survey of all cats. Setting and human familiarity also varied together between pet and shelter groups.

The offered choices matter: Social interaction included human-led play. A preference for that option need not isolate the value of the human from the moving toy. Serial social trials and simultaneous object trials were not procedurally identical. Motivation and previous access to toys/scents were not fully controlled.

Preference is not reinforcement or attachment: Spending time with an option does not yet show that receiving it increases a trained behavior. A reinforcer is a consequence shown to increase behavior; the authors proposed a later training experiment to test this. Nor does the task measure an owner’s secure-base role, subjective love, or long-term welfare improvement.

Reporting caveats: The abstract/discussion use “majority” although the final count is 19/38. The extracted statistical-methods passage says data were non-normal while reporting all Shapiro–Wilk p values greater than 0.05; that wording is inconsistent with the usual interpretation. Extracted figure-caption inequality symbols also appear questionable. Without inspecting figure images or reanalyzing data, do not silently repair those passages or rely on them for exact statistical conclusions. The headline counts and social-versus-food comparison above follow Results 3.3.

Reading route — Lumi

Read the abstract → Methods 2.1 (who remained in the sample) → Methods 2.2 (what choices were offered) → Results 3.3 → Discussion paragraphs on reinforcement and motivational state. The central skill is separating “largest observed category” from “demonstrably better than food,” and preference from a tested training reward.

Guided reading — Lumi

Lumi’s explanatory synthesis of the inspected publisher main text. No prior reading is required. Figure images and raw data were not independently audited.

1. Follow the cats, not just the headline

• What to look for: Methods 2.1 starts with 50 recruited cats, then describes nervousness and lack of interaction; only 38 contribute to the final preference comparison.
• What it means: The denominator is the group actually analyzed, not everyone recruited. A cat that avoids the test is missing from the ranking, not automatically classified as disliking humans.
• Why it matters: The result describes cats that engaged under these conditions. It cannot tell us how all fearful, unsocialized or otherwise different cats would respond.

2. Reconstruct the menu before interpreting a choice

• What to look for: Methods 2.2 first identifies favored options within each category, then compares those options together. Human-led feather play counts as social interaction, and cats encounter familiar or unfamiliar people depending on population.
• What it means: “Social” is a bundle of possible experiences, not simply sitting beside a motionless person. The experiment compares offered activities, not pure emotional values.
• Why it matters: Changing the toy movement, person’s behavior or familiarity could change the menu’s attractiveness. Those are candidate follow-up controls, not conditions already isolated by this study.

3. Read the counts alongside the comparison

• What to look for: Results 3.3 reports 19 social, 14 food, four toy and one scent preference; the social-versus-food comparison has p = 0.35.
• What it means: Social interaction won the largest share of individual rankings, but the data did not establish a reliable difference from food in that comparison. A nonsignificant result is not proof that the two options have identical value.
• Why it matters: “Cats preferred humans to food” removes both the uncertainty and individual differences. “Half of the analyzed cats favored the offered human interaction” preserves the actual result.

4. Separate a promising reward from a demonstrated reinforcer

• What to look for: The Discussion proposes pairing preference assessment with an operant training task in future research.
• What it means: Free engagement measures what a cat spends time with. Reinforcement requires showing that a consequence increases the behavior it follows, such as a trained target touch.
• Why it matters: This paper helps select candidate rewards or enrichment, but it did not demonstrate training effectiveness or lasting welfare improvement. Let the individual cat choose; do not interpret avoidance as stubbornness.

Two reflection questions — Lumi

1. If a cat chooses human-led play, what additional comparison would separate interest in the person from interest in the moving toy?
2. What would you need to measure before calling that preferred interaction an effective training reinforcer rather than simply an attractive activity?