Chapter 3. Summary of Consumer Survey on Issue Importance in Beef Purchases

Chapter 3. Consumer Survey

Chapter 3. Summary of Consumer Survey on Issue Importance in Beef Purchases

Summary

This chapter reports a summary of findings from a consumer survey conducted during April 2013 assessing importance of a variety of factors in beef purchasing decisions. Key conclusions include:

• Overall, consumers indicate that the relative importance of key demand issues influencing ground beef and beef steak purchasing decisions are similar.

• Using ranking questions similar to the beef industry expert survey conducted in February 2013, Food Safety is most commonly selected as having the most impact on ground beef and beef steak purchases. The second and third most commonly selected factors influencing purchase decisions are, respectively, Product Quality and Form and Price. Social Aspects and Sustainability are typically viewed as having the least impact by consumers responding to the survey with Nutrition and Health commonly falling in the middle of these rankings.

• Overall, the order of issue importance from consumers responding to this survey is similar to that obtained from the beef industry expert survey using the same ranking question. Food Safety, Product Quality and Form, and Price are the three factors ranking highest in both consumer and expert assessments whereas Social Aspects and Sustainability are the two factors with the lowest rankings in both the consumer and beef industry expert responses. Nutrition and Health are ranked in between these two extremes.

• There appears to be more of a consensus among beef industry experts than consumers regarding the relative impact of different issues. In each case except one (Sustainability in ground beef surveys), quartile ranges of responses are equal to or larger in the consumer than in the expert responses.

• Applying best-worst scaling to the broad demand factors made possible assessment of the factors relative importance. On average, the four most important issues in ground beef purchase decisions are Safety, Freshness, Taste, and Health. The four least important issues are Convenience, Environmental Impact, Origin/Traceability, and Animal Welfare. The remaining three issues (Nutrition, Hormone/Antibiotic Free, and Price) rank in between these two extremes. While some exceptions exist as discussed below, these relative rankings are generally consistent with expectations and findings of previous research.

• Significant heterogeneity in relative importance rankings exists across consumers. For instance, 20 consumers indicated Hormone/Antibiotic Free is the most important issue in their responses to all six questions that included Hormone/Antibiotic Free whereas the most common response (295 respondents) was to never select Hormone/Antibiotic Free as the most important issue. The conclusion is that, although averages accurately capture the perspectives of most consumers, not all consumers have the same perspective.

• The issues of highest importance to beef steak buying decisions are Freshness (22%) and Safety (19%). The next two most important issues, on average, are Taste, and Health. The four least important issues are Origin/Traceability, Convenience, Environmental Impact, and Animal Welfare (each less than 4%).

• Overall, the ranking of key factors influencing beef purchasing decisions are mostly consistent with previous research and findings from the survey of beef industry experts. Finding Safety, Freshness, Taste, and Health as the four most important issues and Social issues (e.g. Environmental Impact, Origin/Traceability, and Animal Welfare) as the least important issues is supported by both the research literature and the survey of beef industry experts.

• Results from the survey of consumers are still preliminary. Additional analyses are underway to better understand the heterogeneity in issue importance as well as characteristics of consumers placing different levels of importance on the examined issues. Additional research is still needed to better understand the underlying drivers of the findings highlighted here.

Introduction and Purpose

An online survey of a representative sample of U.S. households was conducted in April 2013. The survey assessed the importance of several issues in consumer food purchasing decisions. Participants were randomly assigned different food products to be evaluated with 488 respondents completing a ground beef survey and 487 respondents completing a beef steak survey. The survey was designed in part to provide comparisons to results from the expert survey and previously published research results findings. To directly compare with the expert survey noted in chapter 4, we included a parallel ranking question to assess how the seven broad factors evaluated by experts are viewed by consumers. Realize, however, that the beef industry experts were asked to look forward 10 years whereas the consumer survey focused on current preferences. Second, a sequence of 11 questions was included to assess the relative importance of 11 specific issues using a best-worst scaling (also known as maximum difference scaling) approach.

Procedure and Results

The first question on the consumer survey was: “Please rank (1 = most impact; 7 = least impact) these 7 broad factors in terms of their impact on your purchasing of {ground beef (hamburger)/beef steak}.” where the broad factors evaluated were the same as in the expert survey (chapter 4). Tables 1 and 2 provide summary statistics on responses to this question. Given this question was asked in an ordinal manner, we focus our discussion on median and quartile range (75th percentile less the 25th percentile) statistics (but we also report mean and standard deviations) to discuss the most common response and the dispersion of responses.

Tables 1 and 2 provide insights into the relative importance of a representative consumer as well as the consensus (or lack thereof) in importance rankings across consumers. Median values for the ground beef and beef steak assessments are similar. Food Safety was most commonly selected as having the most impact on purchases. Product Quality and Form and Price followed as the second and third most important factors, respectively. Social Aspects and Sustainability were typically viewed as having the least impact with Nutrition and Health commonly falling in the middle of these rankings. Overall, the order of median values from the consumer survey is similar to that obtained in the expert survey. Although their respective order varies across the two surveys, Food Safety, Product Quality and Form, and Price are the three factors receiving the highest rankings from both consumers and beef industry experts. Importantly, Nutrition and Health both received medium median rankings and Social Aspects and Sustainability received the lowest median rankings in both surveys.

Comparing relative importance of key factors influencing beef purchase decisions using median scores helps characterize a representative consumer’s perspective, but it’s important to recognize that all consumers are not alike. Examining the dispersion in responses among consumers is one way to look at some of the differences that exist among consumers. As signaled by quartile range values, considerable differences exist among consumer responses. For example, an examination of individual response data reveals that at least 3% of the sample selected each of the seven broad factors as being most impactful and another subset of at least 3% selected each of the seven broad factors as being least important.

Comparing the quartile ranges (QR) of the consumer and expert responses reveals several important differences. There was much greater consensus among experts regarding the impact of price (QR=1) than among consumers (QR=3 for ground beef, QR=4 for beef steak). In fact, in each case except one (Sustainability in ground beef versions) the quartile ranges are equal to or larger in the consumer than in the expert assessment. We are unable to explain this difference definitively, but it likely reflects different levels of familiarity with the evaluated factors and timetables of assessment (current for the consumer survey vs. 10 years ahead for the expert survey). Furthermore, experts were likely projecting impacts for typical consumers and not those on the extremes of the distribution and as such expert responses would be expected to exhibit smaller variation than consumer responses.

a Values are from survey question asking respondents to rank (1= most impact; 7=least impact) the same broad factors used in the expert survey (see chapter 4) in terms of impact on their purchasing of ground beef.

a Values are from survey question asking respondents to rank (1= most impact; 7=least impact) the same broad factors used in the expert survey (see chapter 4) in terms of impact on their purchasing of beef steak.

Our consumer survey also included a sequence of 11 questions to assess the relative importance of 11 specific issues. Our assessment was motivated by recent applications of best-worst scaling in the agricultural economics literature (Lusk and Briggeman, 2009; Lusk, 2011) as well as a desire to gain insights into the relative importance (in greater depth than just ordinal ranking) of issues influencing beef purchases. Importantly, this approach provides estimates that are significantly related to actual purchasing behavior (e.g. they have external validity) and provides insights into a wider range of issues than feasible with other methods (Lusk, 2011).

Table 3 shows the specific list of issues examined, and descriptions provided to respondents. Across the 11 presented questions, each issue or product attribute appeared a total of six times, appeared in a question with the other 10 issues exactly three times, and appeared first, second,.., sixth in presentation order exactly once. This reflects our use of a balanced, incomplete block design. An example question as presented to respondents is provided in figure 1.

Figure 1. Example Beef Steak Most Important/Least Important Question

Which of the following issues is most important and which is least important when you purchase beef steak? (check only one issue as the most important and one as the least important)

Our analysis of the most/least important responses first followed the count-based approach outlined in the appendix of Lusk and Briggeman (2009). This procedure required that we count the number of times respondents chose an issue as most important and subtracted that from the number of times respondents chose the issue as least important. Since each issue appears a total of six times across the 11 completed questions, this provides individual specific scores for each issue ranging from a possible smallest value of -6 to a possible maximum of +6. Moreover, across the 11 factors the importance scores sum to zero by design, implying each score reflects importance relative to the mean level of importance (e.g. the scores are effects-coded). The mean values and correlations of these count estimates are provided in Tables 4 and 5.

On average, Safety, Freshness, Taste, and Health received the most positive average scores reflecting the fact they were selected as most important more frequently than least important. Conversely, Convenience, Origin/Traceability, and Environmental Impact have the most negative average scores reflecting the fact they were selected as least important more frequently than most important. The other four issues evaluated fall into an intermediate group with scores (near zero) reflecting average importance.

None of the importance count estimates have correlations exceeding 0.5 suggesting they reflect unique constructs (Lusk and Briggeman, 2009). A deeper assessment of these correlations sheds additional light on what these importance scores are capturing. For instance, the Price importance score in the ground beef sample has a negative correlation exceeding 0.30 in absolute value with scores for Animal Welfare, Environmental Impact, Health, Hormone/Antibiotic Free, and Origin/Traceability. This suggests that individuals placing more importance on price in their ground beef purchasing decisions put less emphasis on these issues, which mainly fall into the Social Aspects factor discussed above. It is also worth highlighting that these importance scores derived from best-worst scaling questions are generally correlated with the 1-7 rankings discussed earlier. As an example, the forced ranking of Price and the effects coded Price importance score from our best-worst questioning have correlations of -0.64 and -0.58 in the ground beef and beef steak samples (-1.0 would imply a perfect correlation since a rank of 1, the most important rank, would correspond to an importance value of 6, the maximum importance value in the best/worst scaling analysis). This confirms there was some consistency regarding how survey respondents completed both the ranking and the best/worst scaling assessments.

The main purpose of asking a series of most/least important questions was to identify the relative importance of the 11 examined issues. Accordingly, we estimated multinomial (MNL) and random parameters logit (RPL) models to identify the share of importance of each issue (Lusk and Briggeman, 2009). Model estimates along with importance shares are presented in Tables 6 and 7. Log likelihood ratio tests suggest that the RPL model, which captures heterogeneity in importance as expressed by survey respondents, is preferred to the MNL model so we focus on RPL results. In fact, significant heterogeneity is identified by the RPL for five of the eleven issues examined in both the ground beef and beef steak samples.

Consistent with the individual scores discussion above, the importance shares must sum to 100% by design. The greater the estimated share of importance, the more important the factor is. If all factors had equal importance, the share for each factor would be 9% (100%/11 factors). In the ground beef survey (Table 6) the issue of highest importance is Safety (24%). The Health importance share of 12% implies that Safety is twice as important as Health to the typical respondent in ground beef purchasing decisions. On average, the four most important issues in ground beef assessment are Safety, Freshness, Taste, and Health. Combined these four issues account for over 70% of overall importance for the typical consumer.

The four least important issues to consumers responding to the survey are Convenience, Environmental Impact, Origin/Traceability, and Animal Welfare. The remaining three issues (Nutrition, Hormone/Antibiotic Free, and Price) importance falls in between these two groups. It is also important to recognize significant heterogeneity highlighted by the RPL. For instance, while the average importance share for Hormone/Antibiotic Free is 6% in the ground beef assessment, the RPL estimated a statistically significant standard deviation reflecting substantial heterogeneity. One way to clearly see this in the underlying data is to note that there were 20 respondents who indicated Hormone/Antibiotic Free was the most important issue in all six questions including Hormone/Antibiotic Free, however the most common response (295 respondents) was to never select Hormone/Antibiotic Free as the most important issue. This highlights that fact that low importance share values in Tables 6 and 7 do not imply an issue is necessarily irrelevant to all consumers.

In the beef steak survey (Table 7) the issue of highest importance is Freshness (22%) which is closely followed by Safety (19%). The next two most important issues on average are Taste, and Health. The four least important issues are Convenience, Origin/Traceability, Environmental Impact, and Animal Welfare (each with shares less than 4%). Overall steak importance shares are similar to those in the ground beef assessment. Comparatively, Price is more important to consumers in beef steak purchasing decisions whereas Safety and Health are more important to consumers in ground beef decisions. The fact ground beef has been of more focus in USDA FSIS recalls and the subject of discussion with respect to fat content compared to steak may help explain these differences.