1. Introduction
Global fish production (both capture fisheries and aquaculture) is projected to reach 204 million tons by the year 2030, with 89% intended for human consumption [
1]. It has been estimated that up to 70% of fish biomass can be wasted during processing, which can have social, economic, and environmental implications [
2]. As such, valorization of seafood byproducts has been gaining scientific interest [
2,
3]. Keeping edible and nutritive fish components, such as skins, in human food systems (rather than waste or non-food applications) can contribute to an increasingly circular and sustainable food supply where natural resources are fully exploited [
4]. However, adoption of circular economic principles will require changes in consumption practices [
4]. Therefore, consumer insight is needed for effective waste-to-value product development.
Our previous research turned to seafood consumers for indications of product appropriateness for seafood byproducts, which is a cognitive dimension of food choice based on perceived fit or “appropriateness” of an ingredient within a final product concept or preparation [
5]. In that study, consumers identified fish products as most appropriate for seafood byproduct incorporation and snack foods as a top-five product category. It was proposed that food made from seafood byproduct could be both a snack and a seafood product unto itself, which motivated the product concept for this consumer research: catfish skin chips (CSC).
Fish skin chips are available commercially, which was verified by a simple internet search, with most resulting products being made from salmon skin. In Asian countries, fried basa fish (
Pangasius bocourti) skins are marketed as snack food [
6]. Research has been conducted on composition, instrumental hardness, and shelf-life of fried tilapia (
Oreochromis spp.) skin chips [
7] and physiochemical and sensory properties of puffed tilapia snacks by air-frying [
8] and other cooking methods [
6]. While non-traditional cooking methods may lend novelty to these snack preparations, the current research is unique in its development and testing of air-fired chips made from catfish (
Ictalurus punctatus) skins.
In addition to understanding consumer perceptions of potential new products, investigations of food made from seafood byproduct also present opportunities to explore broader consumer research metrics. Specifically, benefits of food-waste reduction and sustainability, along with nutritional value of fish skin [
9], presented the authors with an opportunity to compare formats for product-related informational cue delivery, and their effects on product acceptability and food-evoked emotions in the present research. As an unintended consequence of conducting two sensory studies (
N = 115 consumers each) at the same location approximately six months apart, we found that twenty-five consumers participated in both studies, thus allowing for a test of repeat-exposure effects on acceptance of CSC.
Previous research has shown that providing consumers with product benefit information after a blind tasting can improve ratings of product liking and purchase intent (PI). This has been demonstrated with informational cues about the fatty acid profile of steaks from grass-fed cattle [
10], risks of high sodium intake paired with a “low sodium” claim for roasted peanuts [
11], and several other product/benefit combinations. Some studies have integrated multiple pieces of information into a single message, e.g., the satiating, metabolism-boosting, and “good” HDL cholesterol properties of certain oils [
12]. Others have presented informational cues separately and sequentially, such as a safety claim for fried fish made with a bone powder breading mix, followed by a separate calcium claim [
13]. Still, it is unclear whether multiple pieces of food product benefit information are more impactful when presented in a single integrated format (one cue) or when presented separately (multiple individual cues). The present investigation compared the effects of CSC informational cues in both formats on product acceptability.
The objectives of this research were to investigate consumers’ perceptions of CSC sensory quality with different added flavorings, product-elicited emotions, and effects of informational cues on CSC acceptability and emotions. Additionally, the authors hypothesized that information delivery format (two separate product benefit messages versus one integrated message) would differentially impact consumers’ overall liking and purchase intent of CSC. This was accomplished via two consumer studies at the same location (N = 115 each, with 25 repeat participants), which permitted investigation of a repeat-exposure effect on CSC acceptance.
2. Materials and Methods
2.1. Preparation of Catfish Skin Chips
Fresh channel catfish (Ictalurus punctatus) skins (CS) were obtained from a local catfish processor (Breaux Bridge, LA, USA) and transported on ice to the research facility (Baton Rouge, LA, USA). Skins were a byproduct of filleting catfish which yielded 9 to 12 oz (55 cm to 75 cm in length) fillets. Immediately upon arrival at the research facility, the CS were cleaned by passing through water 5 times. CS were then rested on a sieve for 10 min to drain off excess water. CS were dried in an oven (OV310 G rotating rack oven, BAXTER Inc., Deerfield, IL, USA) on aluminum trays for 4 h at 50 °C. Dried CS were left at room temperature overnight to be used the next day for the consumer study.
On the days of testing, dried CS were cut into squares (5 cm × 5 cm) using only the most uniform part. Square pieces of CS were battered in egg (beaten grade A chicken eggs; Great Value, Walmart, Inc., Bentonville, AR, USA) and coated with flour (all-purpose wheat flour; Great Value). CS squares were air-fried (ten pieces per basket; NINJA
® Foodi, SharkNinja Inc., Needham, MA, USA) for 12 min at 199 °C. Preliminary air-frying parameters were considered based on Fang et al. [
6], who used 180 °C for up 12 min for tilapia skins. However after in-house evaluations of different conditions, a temperature of 199 °C was found to achieve desired CSC crispiness using the above-mentioned raw material and equipment. Each batch started with a cold unit and took approximately one minute to reach the final set temperature of 199 °C.
Three treatments (flavors) of CSC were prepared by coating air-fired CS with one of the following seasonings (all from McCormick & Company, Baltimore, MD, USA): Lemon & Pepper, LP (salt, black pepper, citric acid, onion, sugar, garlic, calcium stearate, silicon dioxide, calcium silicate, celery seed, lemon oil, and FD&C Yellow N5 Lake), paprika, PK (paprika and silicon dioxide), barbecue seasoning, BBQ (brown sugar, salt, spices including celery, seed, pepper, tomato, garlic, onion, red bell pepper, extractives of paprika, acetic acid, or natural flavor (including hickory smoke)). The plain CSC without any seasoning were considered the control flavor. These flavors were chosen from a range of seasonings based on current snack market trends and pre-screening by ten untrained panelists who ranked the top three seasoned CSC samples based on appearance and flavor. CSC were seasoned, 25 pieces at a time, by shaking them with 60 g of each respective seasoning in a sealed container.
2.2. Chemical and Physical Analyses of Catfish Skins and Catfish Skin Chips
2.2.1. Proximate Analysis of Catfish Skins
Proximate analysis and mineral composition analysis of catfish skin were conducted at the Louisiana State University Ag Center Agricultural Chemistry Laboratory (Baton Rouge, LA, USA). Moisture was analyzed following the AOAC Method 934.01, proteins by the AOAC Methods 981.10 and 990.03, fat by AOAC Methods 2003.05 and 2003.07, and ash by AOAC Method 942.05 [
14].
2.2.2. Color (L*, a*, b*)
Instrumental color was analyzed using a spectrophotometer (CM-5, Konica Minolta Inc., Osaka, Japan). Measurements were recorded as L* (darkness/lightness), a* (greenness/redness), and b* (blueness/yellowness) values. Five measurements per flavor treatment (Plain, Lemon & Pepper, Paprika, and BBQ) were taken. Mean L*, a*, and b* values were used to calculate the delta-E pairwise color differences between treatments (∆E; Equation (1); L*, a* and b* represent means of each respective index, and subscripts
1 and
2 refer to two samples of interest [
15].
2.2.3. Hardness
Hardness of the CSC was measured using a texture analyzer (TA-XTPlus, Texture Technologies, Godalming, UK) equipped with a spherical probe (TA-8, ¼” diameter, aluminum and stainless-steel) and a crisp fracture support rig (Texture Technologies), with a 30 kg load cell. Test parameters were 1 mm/s pre-test speed, 1 mm/s test speed, and 10 mm/s post-test speed. Samples were punctured to a distance of 5 mm. Hardness was expressed as the maximum force recorded during deformation. Ten CSC samples of each flavor treatment (Plain, Lemon & Pepper, Paprika, and BBQ) were measured.
2.3. Consumer Sensory Evaluation
2.3.1. Panelists and Recruitment
Two separate consumer studies were conducted on different days (6 months and 17 days apart), each with a total of 115 consumers; 25 same panelists participated in both studies. Panelists for both studies were recruited from a pool of faculty, staff, and students at the Louisiana State University campus (LSU, Baton Rouge, LA, USA). Participants were at least 18 years old (pregnant women excluded), regular seafood consumers, and had no allergy to fish, wheat, and egg. Informed consent was provided by a signed consent form, and the research protocols were approved by the Louisiana State University Agricultural Center Institutional Review Board (IRBAG-21-0063). Participation was voluntary and no compensation was provided.
The consumer sensory studies were performed at the LSU AgCenter Sensory Services laboratory (Baton Rouge, LA, USA) in partitioned booths under white light. An online questionnaire (Qualtrics software, Provo, UT, USA) was used to collect data. To avoid bias, samples were labeled with three-digit blinded codes, and unsalted crackers and water were provided for palate cleansing to avoid carryover effects. Microbial analysis was done for CSC samples before the consumer study. APC and E coli/coliforms counts were below the detection level, which was 20 CFU/g and 2 CFU/g, respectively; hence, CSC samples were safe for taste testing.
2.3.2. First Consumer Study (Study 1)
The first consumer sample (N1 = 115) consisted of 56% males and 44% females. Forty-five percent were US-born (including White, African American, Hispanic, Asian, and other races), and 32% were Hispanic/Latino—not born in the US. As would be expected from sampling on a college campus, consumers were primarily between the ages 18–25 (57%) and 26–35 (30%).
Product Acceptability
A total of N1 = 115 consumers evaluated four CSC samples (Plain, LP, PK, BBQ flavor) which were served following a balanced and randomized complete block design. Panelists were first asked to observe the sample and rate overall visual quality (OVQ) and surface color, then to smell the sample and rate aroma, and finally to taste the sample and rate texture, crispiness, flavor, and overall liking (OL) (all using a labeled 9-point hedonic scale). A 3-point Just About Right (JAR) scale was also used to rate crispiness (Not crispy enough, JAR, Too crispy). Purchase intent (PI) of each sample was reported on a binomial “yes/no” scale.
Emotions
After informational cues, OL, and PI ratings, consumers responded to the question “After visual, aroma and taste evaluations of samples, how does this catfish snack make you feel? Check all that apply.” Consumers selected applicable emotions from a list of 25 terms [
16] presented in a randomized check-all-that-apply (CATA) format. These included: active, adventurous, aggressive, bored, calm, disgusted, enthusiastic, free, good, good-natured, guilty, happy, interested, joyful, loving, mild, nostalgic, pleasant, satisfied, tame, understanding, unsafe (related to health), warm, wild, and worried.
2.3.3. Second Consumer Study (Study 2)
The second consumer sample (N2 = 115) consisted of 51% males and 49% females. The majority (63%) were US-born (including White, African American, Hispanic, Asian and other races), and 23% were Hispanic/Latino—not born in the US. Consumers were primarily between the ages of 18–25 (66%) and 26–35 (23%).
Product Acceptability
A total of N2 = 115 consumers evaluated three flavored CSC samples (Plain, LP, and BBQ) which were served following a balanced and randomized complete block design. Paprika flavor was not included in this second study due to the significantly lower acceptance scores compared to the other treatments during the first consumer test. During the second study, consumers rated only OL (on a labeled 9-point hedonic scale) and PI (yes/no scale) of samples.
Emotions
Food-evoked emotions were measured at two points during the study using a labeled 5-point Likert-type scale anchored at “strongly disagree” and “strongly agree”. The first instance was immediately after blind ratings of OL and PI: “Please rate each EMOTION term according to your current emotional state after TASTING each sample.” Consumers rated each of eight emotions, which included food-evoked sensation seeking emotions [
17] curious, daring, energetic, interested, and wild, as well as negative emotions unsafe, disgusted, and worried presented in a randomized order. These emotions were rated a second time as the last task of the sensory test, after the HPFWS message and final OL and PI ratings.
2.4. Data Analysis
Mean CSC hardness, color, hedonic scores, and consumer emotions (before HPFSW; Study 2) were compared by analysis of variance (ANOVA). Tukey’s post-hoc test was used for mean separation with physical, chemical, hedonic, and emotion data, and Bonferroni’s adjustment was used with repeated measures ANOVA comparing the effects of separate informational cues on OL during Study 1. Descriptive discriminate analysis (DDA) identified the hedonic attributes influential to overall CSC samples’ differences. A logistic regression model was fit with demographic and hedonic predictors to identify significant predictors of CSC PI for the different informed conditions (before and after HP and FWS messages). McNemar’s test for marginal homogeneity was conducted for significant changes in PI responses with increasing product information. Cochran’s Q test with Bonferroni adjustment compared CATA response frequencies for emotions across CSC treatments. Two-sample t-tests and chi-square tests were used to compare the impact of separate informational cues (HP and FSW) versus the integrated cue (HPFWS) on OL and PI, respectively. Two-sample t-tests and chi-square tests were also used to compare OL and PI, respectively, between repeat-exposure and first-time consumers during Studies 1 and 2. Paired t-tests were used to compare OL of CSC before and after HPFWS messaging and to compare mean emotions ratings (before and after HPFWS; Study 2). Data analysis was conducted using the R software version 4.0.3 (RStudio, Inc., Boston, MA, USA) and Statistical Analysis System (SAS) version 9.4 (Cary, NC, USA).