Effects of animal welfare on sensory acceptance scoring of yogurt

Abstract

Welfare of farm animals can be increased greatly if the animal producers are convinced that increased welfare is useful in terms of product quality. Major purpose of this study was to investigate whether animal welfare changed consumer acceptance and scoring of reduced-fat and plain yogurts (PYs) with or without knowledge of the animal welfare. Four products were used, PY made from whole milk collected from high welfare (HW) farm cows, PY from low welfare (LW) farm cows, low-fat (LF) yogurt from HW farm cows and LF yogurt from LW farm cows. Three evaluations were carried out in the Canakkale Onsekiz Mart University campus; tasting yogurt without labels (Test 1), expectation scoring based on pen and paper (Test 2) and tasting with labels and information on the welfare and fat content (Test 3). In Test 1, the subjects accepted the HW better than LW, and PY better than LF yogurt without the knowledge of the yogurts origin or fat content. The taste seems to direct the subjects scoring towards more fat without the worries related to weight loss. In Test 2, subjects scored LF yogurt higher than PY, which was the opposite of Test 1, but they scored HW higher than their Test 1 scores, which were consistent rank wise. This clearly shows that the subjects participating in the study were interested in LF yogurt consciously but liked that taste of the PY better and were also interested in HW products both in terms of taste and ethics. Some of the differences which were not significant in first two tests were highly significant in Test 3 (P<0.01), which means that the scores were more distinctive among the different products, fortified by both information and taste. The PY-HW yogurts were scored the highest, followed by the LF-HW yogurt samples, the PY-LW yogurts and finally the LF-LW yogurts. This illustrates that the taste buds of the subjects were able to affect scoring significantly (PY scored higher than LF yogurt) and information regarding the fat content did not make enough of a difference to change the ranking from the first test. Sensory perceptions regarding animal welfare were strengthened when sensory input was supported with information provided on animal welfare. In all the three tests, the subjects scored the HW yogurts higher than LW yogurts regardless of the fat content, whether they knew the labels or not. However, information provided on welfare increased the consumer scoring significantly, indicating that HW products should be advertised and labelled clearly to increase consumer sensory acceptance and willingness.

Keywords:

• yogurt
• animal welfare
• consumers
• blind tasting

Introduction

Animal welfare has been discussed in the world for a long time (Verhoog 2000) and implementation of animal welfare is urged by small and large associations such as People for the Ethical Treatment of Animals and the Farm Animal Rights Movement, and large meetings have been organised to increase awareness for the welfare or rights of pets, farm animals and other captive animals. Though many of the activists know how to protest effectively, these activities have varied success, especially in farm animals (Taylor 1995). A pre-eminent way to enforce animal welfare is by providing economic incentives and other enticements to the farmers. Farm animal welfare is mainly focused on freedom of movement, cleanliness of the environment that the animals live in and general behaviour of the producer towards the animals.

The most excellent incentive for farmers is an economic incentive. Producers would be more than willing to participate in welfare practices, if they were convinced that their income is increased through welfare. This is a more effective way compared with the method of raising awareness and education of the farmers. Therefore, a study is required to investigate the need for increased sensory acceptance in milk and dairy products through animal welfare. The general view of the animal welfare is that adjusting management practices according to welfare values create increased costs (Tweeten 1991). However, the humane practices of welfare require sensible farm management practices such as cleanliness, increased udder health and humane treatment towards livestock. These practices may increase food quality as well. Clean and healthy animals are happy and they perform better in terms of production traits. Improved welfare may also contribute to decreased diseases and mortality rate (Appleby 2005). Different management practices towards the well-being of the farm animals may increase the animals’ production quantity or quality, which needs to be studied. Scott et al. (2001) reported that welfare of farm animals is associated with the economic interests of farmers and the health of animal source food consumers. Taylor (1995) wrote that ‘animal welfare groups, livestock producers, and others concerned about animal welfare must work together to fund and support bona fide research and investigation into those matters of greatest concern’.

Napolitano et al. (2008) reported that consumers use some parameters in choosing their food, including sensory properties and price, in addition to healthiness and naturalness. Good animal welfare practices and providing this information to the consumer may increase sensory acceptance of dairy products. Haddad et al. (2007) reported that fat content and sensory properties were the major determinants in purchasing decisions of consumers for Labneh yogurt, while information on price, processing and nutritional characteristics accounted for a small percentage (Napolitano et al. 2008). Fat content in yogurt is an important factor of taste (Radi et al. 2009) and may affect and overcome taste differences due to animal welfare. Therefore, it is important to include and test whether high welfare (HW) products are preferred regardless of the fat content, a strong indication of taste and acceptability in yogurt. Major purpose of this study was to investigate whether animal welfare changed consumer acceptance and scoring of reduced-fat and plain yogurts (PYs) with or without knowledge of the animal welfare.

Materials and methods

Four different products were used, PY made from whole milk collected from HW farm cows, PY made from whole milk collected from low welfare (LW) farm cows, low-fat (LF) yogurt made from milk collected from HW farm cows and LF yogurt made from milk collected from LW farm cows (Figure 1). The tasting tests were carried out in the Canakkale Onsekiz Mart University campus with mixed consumers including faculty, students and workers. The consumers were a total of 72 subjects who were between the ages of 18 and 58 and included different income levels. Yogurt consumption is very common in Turkey and in Canakkale, and the subjects were regular yogurt consumers.

Figure 1.  The products used in the tasting experiments.

Milk samples were collected in Canakkale province of Turkey, and the farms were classified as HW and LW. Assessing animal welfare on farm is costly and time consuming, and using routine data kept in the farm and in the national databases can be a more practical solution (de Vries et al. 2011). Scott et al. (2001) reported that farm animal welfare is difficult to quantify and evident animal welfare aspects consist of animal health, behaviour and husbandry practices. Verhoog (2000) wrote that animal welfare is related to values, which are subjective. The classification in this study was based on cleanliness of the farm, freedom of movement and care takers’ general behaviour towards the animals. All three criteria were based on subjective observations and scoring. Cleanliness included two scores based on subjective observation, clean and unclean, related to faeces collection and barn cleanliness. Freedom of movement included two scores, especially related to where the animals rest and the scores were small and large. The ‘small’ score included animals that were close together in the resting and feeding area, while the ‘large’ included animals that had more space. The space was not measured but was scored subjectively through observation based on the space between the animals. Care takers’ general behaviour towards the animals was scored in two levels as well. The gentle care takers groomed the animals while tough care takers were distant and did not groom the animals. Milk fat, milk protein, non-fat dry matter, lactose levels and density were measured using the Milk Analyzer (Lactoscan 90 Bulgaria). LF yogurts were produced by removing half of the fat from the yogurts produced, decreasing the fat content to around 1.5% for the LF yogurt samples (Table 1).

Table 1. Milk components for the samples obtained from HW and LW farm animals.

Components	HW samples	LW samples
Fat (%)	3.20	4.17
Non-fat dry matter (%)	10.28	9.79
Protein (%)	4.07	3.90
Lactose (%)	5.25	4.99
Density (kg/m^3)	1037.86	1035.31

The tasting experiment included three different tests summarised in Table 2, using the products shown in Figure 1. The subjects were offered the four different products, PY-LW, PY-HW, LF-LW and LF-HW yogurts. The subjects tasted the products in the first test and were asked to score based on the taste only. The first test did not include labels. The scores ranged from 1 to 5, 1 being the lowest score and 5 being the highest score in terms of the perceived acceptability. The scores were explained to the subjects in detail and any questions were answered. The second test involved only pen and paper, and the subjects were asked to score, using the same scoring and criteria, the four products labelled in Figure 1, if they were presented for tasting. The subjects were informed about the welfare situation and the fat content of the yogurt they would score. The third test was performed with a different batch of yogurt samples, to make sure that the subjects did not remember how the first batch from the first test tasted. The model included the score as the dependent variable. The independent variables were the test (1, 2, 3), sample (1, 2, 3, 4), sample×test, age (18–22, 22<), yearly income (10,000 Turkish Lira and below; 10,000–20,000 Lira; 20,000–30,000 Lira and 30,000 Lira and above), originating city (Ankara, Canakkale, Diyarbakir, Hatay, Istanbul, Izmir and Konya) and gender (male and female). However, because it is very hard to eliminate all carryover effects, repeated analyses were carried out to account for the dependencies in errors arising from individual participants. The data in the tasting experiments were repeated in time and the error term was not independently distributed. Thus, the covariance structure of the R matrix was specified according to two fitness criteria to account for the dependencies in errors. Akaike's Information Criteria (AIC) (Akaike 1974) and Bayesian Information Criteria (BIC) (Schwarz 1978) associated with the models were calculated to detect the fittest covariance structure of the repeated model. The statistical analyses were carried out using the MIXED procedure of SAS (SAS Institute Inc. 2004).

Table 2. The experimental design of the three different tests carried out.

Test	Presentation	Type of evaluation	Type of rating
1	Yogurt	Tasting without information	Blind scoring
2	Information	Expectation	Expected scoring
3	Yogurt + information	Tasting with information	Actual scoring

Results and discussion

The Toeplitz model was used as the covariance structure. This variance–covariance model was used because it had the lowest fit statistics, which indicates a better fit (Table 3). The Toeplitz was clearly the best fit structure, followed by the Heterogeneous Compound Symmetry. The results indicated that the fixed effects test (P<0.01), sample (P<0.01) and sample within test (P<0.01) were highly significant. However, age (P=0.47), yearly income (P=0.36), originating city (P=0.60) and gender (P=0.35) were not significant.

Table 3. AIC and BIC fit statistics for different covariance structures for score in the tasting experiments.

Covariance structure	AIC	BIC
Compound symmetry	2604.4	2608.9
Auto regressive	2648.7	2653.2
First-order autoregressive moving-average	2604.4	2608.9
First-order factor analytic	2581.2	2635.9
First-order unstructured	2650.7	2678.1
Huynh-Feldt	2602.3	2631.9
Toeplitz	2556.0	2583.4
First-order banded Toeplitz	2670.4	2672.6
Heterogeneous Toeplitz	2650.7	2678.1
Variance components	2670.4	2672.6
Ante-dependence	2634.0	2686.4
Heterogeneous compound symmetry	2580.8	2610.4
Heterogeneous autoregressive	2630.3	2659.9

Differences among yogurts within Test 1, tasting only

In Test 1, the subjects tasted the products without labels or information. By tasting only, the least squares means indicated that they scored the highest for PY-HW followed by the LF-HW, PY-LW and the LF-LW yogurt samples. This pointed to the conclusion that the subjects accepted the HW yogurts better than LW yogurts, because regardless of the fat content, they picked the HW yogurts over the LW yogurts without the knowledge of the yogurts origin. In addition, it indicated that the consumers preferred the PY, which has a higher fat content compared with the LF yogurt. Because this test was based on sensory input only, the taste seems to direct the subjects scoring without the worries of the labelling of fat content and the worries related to losing weight. Differences between PY-LW and PY-HW were highly significant, indicating that HW yogurt samples were scored better and the differences were not due to chance. The reason for the HW products getting better scores even in blind tasting can be stress hormone cortisol changing the taste of the milk. In addition, the LW farm animals may not be kept as clean, causing extra microorganism activity, which lowers the milk quality due to enzymes and toxins released by bacteria. Somatic cell count is an indication of udder health in dairy animals (Pala and Koyuncu 2007; Koyuncu and Pala 2008; Pala and Sahin 2011) and may be lower in a HW farm. There are large welfare implications of toxic mastitis and may lead to increased sensitivity to pain in cows (Bradley 2002). Appleby (2005) reported that improving welfare may decrease the costs to farmers by measures to decrease diseases and mortality rate.

All other differences (Table 4) were also highly significant in Test 1 (P<0.01), except the differences of PY-HW and LF-HW, which were large enough to be significant only at the alpha level (P=0.05). This indicates that when the samples were scored based on sensory perception only, the subjects were able to differentiate and score higher fat content in yogurt, but not to the full extent to give it a highly significant difference. Another exception to the highly significant differences was the differences between PY-LW and LF-HW yogurts (P=0.16). This is because PY, as seen from the previous comparisons and the numbers in Table 4 is preferable to LF yogurt, and HW yogurt seems to be preferable to LW yogurt. In this case, preference of PY seems to offset the undesirable LW yogurt, or HW offsets the undesirable LW yogurt, making LF-HW yogurt non-significant compared with the PY-LW yogurt. In other words, each of the two yogurts has an undesirable characteristic (LW or LW) and a desirable characteristic (PY or HW) preferred by the consumers and welfare and fat content seem to counterbalance each other.

Table 4. Effects of the welfare and fat on consumer scores.

The tests	Plain yogurt, low welfare (PY-LW)	Plain yogurt, high welfare (PY-HW)	Low fat, low welfare (LF-LW)	Low fat, high welfare (LF-HW)	SE
Test 1	2.38^A,a	2.97^A,b	1.86^A,c	2.65^A,a	0.25
Test 2	1.98^B,a	3.58^B,b	2.09^A,a	4.06^B,c	0.25
Test 3	2.54^A,a	3.51^B,b	1.99^A,c	3.09^C,d	0.25
Note: Pooled standard error based on most conservative number in a row.
Column^A,B,C values or row^a,b,c,d values with different superscripts differ (P<0.05).

Differences among yogurts within Test 2, information only

In Test 2, the subjects were given labels only, and they scored based on the labels, without tasting anything. The scores were the highest for LF-HW, followed by PY-HW, LF-LW and finally PY-LW yogurt samples. This ranking clearly shows that the subjects participating in the study were interested in LF yogurt consciously, and they consistently scored the LF options higher than the higher fat PY option. The subjects were also interested in HW products and scored them higher. This scoring is only different than the blind tasting of Test 1, where the subjects scored the LF yogurts lower compared with the high fat PY samples. This indicates that the consumers want LF yogurt when asked, but they prefer the higher fat variety if they only taste the samples and do not know the labels. In Test 2, all the differences were highly significant (P<0.01), except the differences between PY-LW and LF-LW (P=0.50). The non-significant difference indicated that LW options were scored low, nearly regardless of the fat content. This is expected as people tend to score in favour of animal welfare, as shown by a survey carried out by European Commission in 2007 (Napolitano et al. 2008).

Differences among yogurts within Test 3, tasting and information

In Test 3, which was conducted by tasting with information, all differences were highly significant (P≤0.01). This indicated that the information provided regarding animal welfare and fat content strengthened and fortified the consumer preference based on taste alone. Some of the differences which were not large enough to be significant in the previous two tests were highly significant in this test, which means that the scores were more distinctive, they were separated much better among the different products, compared with the within product differences. The Test 3 confirmed the results obtained in Test 1 for fat content and confirmed the results in all the previous tests for animal welfare. PY-HW yogurt had the highest scores, followed by LF-HW yogurt samples, PY-LW yogurts and finally LF-LW yogurts. This indicated and confirmed that sensory perceptions are strengthened by information provided and natural fat in yogurt made with milk from HW animals is preferred over the other alternatives. LF yogurt made from milk from LW animals was the least preferred yogurt, especially when sensory input was supported with information provided on the fat content and animal welfare. Test 3 indicated that once tasted, the subjects preferred the PY yogurt over LF yogurt regardless of weight loss concerns, while they scored LF yogurt higher when scoring is based on pen and paper only. This is a clear indication that tasting boots in grocery stores are effective in marketing whole products by giving the customers a taste of the products that they are planning to consume.

Differences among the tests for PY-LW yogurts

Differences among the tests were shown using capital letters in Table 4 for the different products used. PY-LW yogurt samples received similar scores in first and the third tests, but had significantly lower scores in the second test. The second test was carried out using information only, so this means that the consumers do not wish to consume LW-PY, when asked. The LW yogurt was scored low, which is reasonable, because consumers wished HW of animals, especially based on information only input. The subjects also did not consider PY highly, as many people try to watch their weight and many publications (Shi et al. 2001; Zemel et al. 2005), tend to promote less consumption of fat, even dairy fat for weight loss. Differences between the first test and the third test were non-significant for the PY-LW yogurt samples, indicating that tasting without information and tasting with the information provided (which yogurt is which product regarding welfare and fat content) did not make a large enough difference for the differences to be significant. This illustrates that the taste buds of the subjects were able to differentiate what is desirable, and even after information, awareness is provided, they stick with what they liked in the first place, and information regarding LW and high fat content did not make a large difference. The subjects tended to prefer PY over LF when tasting, preferred LF yogurt based on labels only, and preferred the PY samples over the LF variety based on both taste and information.

Differences among the tests for PY-HW yogurts

PY-HW yogurt samples had similar scores in the second and the third tests, while the first test, blind tasting were significantly lower (P<0.01) than the other two tests. Though the evaluation of samples within tests showed that the subjects preferred HW over LW, these statistics showed that information on animal welfare can change the consumer sensory acceptance significantly. This signifies that though products originating from HW animals are accepted better than those from LW animals, HW products should be advertised and labelled clearly to increase consumer sensory acceptance and willingness.

Differences among the tests for LF-LW yogurts

LF-LW yogurt samples had the lowest scores in the first and third tests, which included tasting the product, and had the second lowest scores in the second test, based on labelling only. However, none of the differences among the tests were large enough to be significant, indicating that the yogurt samples were not preferred by the subjects. The only success of these types of yogurt was in the label-based test, which did not include tasting, and the higher score is probably due to weight loss concerns.

Differences among the tests for LF-HW yogurts

LF-HW yogurt samples received the highest scores in the second test, followed by the third test, and the first test. All differences were highly significant, indicating that the LF and HW was the ideal yogurt for weight loss and ethical reasons, and the ideal situations received its highest score in idea, in the second test. The scores were still high in the third and the first tests, but not as high as they were in the second test. When reality sets in the form of tasting, LF yogurt seems to lose its attractiveness, even when the subjects are informed of the situation.

In general, the subjects scored HW yogurts higher than LW yogurts. However, the scores in the second test were much higher than the first test for HW yogurt samples. This indicates that though the asubjects had better sensory feeling associated with HW samples, the conscious/ethical scoring of the HW yogurt was much higher, compared with the sensory differences. The subjects scored LF yogurts higher when based on labels, but scored them lower when tasting the product, and were consistent with their scoring even when both information and tasting were offered. The reason for that may be the concern for weight gain and obesity associated with high fat products (Shi et al. 2001).

The consumers preferred HW products over LW products. However, information provided on welfare increased the consumer scoring significantly, indicating that clear labelling and advertising the welfare aspect of a product can increase the sales for HW products. This may provide a basis for farmers to increase the welfare of their animals to obtain price premiums (Appleby 2005), and increase advertising on this feature. Increased advertising and purchase of products from HW farms may increase popularity of such management among other farmers and gain ground in the normal farming practices.