A Screening Tool for Selection of Field Odor Assessors

ABSTRACT

Field odor assessors are required to rate or describe several odor parameters, such as intensity, duration, offensiveness, and character. Ideally, their assessments should reflect the average odor perception of a specific community. The authors developed a three-part screening test for recruiting odor assessors: (1) distinguishing between different odorants by means of a triangular forced-choice test; (2) evaluating odor intensity; and (3) describing hedonic tone and odor character. Grading was based on two criteria: correctly answering the relevant parts of the test, and evaluation of odor parameters relative to the entire tested population. The latter involved grading each tested individual according to the similarity of their score to the average result of 179 tested individuals, comprising 48% women and 52% men whose age and residence distributions were identical between women and men (except for the oldest group). To exclude relatively less sensitive individuals who showed poor ability to distinguish between different odorants and various odor intensities, and/or provided atypical description (or rating) of odor intensity, character, and offensiveness, it was suggested that only individuals whose final score was within the upper 75% (final score ≥80.75) would be qualified as odor assessors. According to this criterion, 73.8% of men and 78.6% of women passed the test. Among urban and rural dwellers, 77.4% and 67.4%, respectively, were qualified. Pass rate clearly diminished with increasing age: from 89.3% at 21–30 years to 54.6% at 61–70 years. This screening tool is recommended by the Israel Ministry of Environmental Protection for selection of field odor assessors to serve the general community and regulatory authorities.

IMPLICATIONS

A new screening tool for selection of field odor assessors is presented. A unique feature of this tool lies in its grading procedure, which is based on two desired qualities: (1) distinguishing between different odorants and various odor intensities; and (2) describing (or rating) odor intensity, offensiveness, and character in a way that best reflects that of the average residential population. Use of this screening tool for selection of field odor assessors to serve the general community, regulatory agencies, and other environmental authorities seems practicable, and it was proposed to Israel's Ministry of Environmental Protection as a national standard tool. The test is very simple to perform and does not need any special laboratory; however, because of expected cultural variations in odor perception, local calibration of the test is recommended.

INTRODUCTION

The nuisance and health concerns raised by odors emitted from industrial and livestock facilities are among key issues that concern neighboring communities. Odors may become an issue when a residential population is located near industrial facilities, such as food manufacturers, chemical factories, oil refineries, and rendering plants,1 or agricultural activities, such as confined animal feeding operations,2 composting facilities,3 and land application of manure or biosolids.4,5 Advanced sampling systems and chemical analyses increasingly are used to characterize odors at the source and downwind6,7 but, nevertheless, sensory evaluations, either by laboratory olfactometers or by human field assessors, remain essential because of the difficulties in determining correlations between odor strength and offensiveness, and the concentrations of specific odorants in complex odorous mixtures.

Field odor assessors are often sought by environmental regulatory authorities, residential communities, and odor researchers.5,8–11 In all such roles, odor assessors are regularly asked to rate or describe several odor parameters, such as intensity, duration, offensiveness, and character. Ideally, such assessors would reflect an “average odor perception” as related to the sensitivity of the general population to the target odors, that is, neither insensitive nor hypersensitive. Thus, the standard n-butanol detection threshold that is used to qualify odor panelists in olfactometry laboratories, in accordance with international standards such as the EN 1372512 or as determined by commercially available butanol sniffing-pen kits,13 may not be suitable to qualify field assessors, whose role extends beyond the determination of odor thresholds. At present, there are no international nor, in many cases, national standards for certification of field odor assessors, and this lack can lead to difficulties in odor regulation and in comparing odor studies.

The aim of the present study was to develop a screening tool for selection of field odor assessors. Being based on nonprofessional volunteers, this kind of test must be simple and relatively short (about 30 min). For practical reasons, it is better to reduce the need for special laboratory equipment, so that the test can be conducted in available community facilities, such as school laboratories. Data analysis was regarded as an integral function of the required tool, which should meet two grading criteria: (1) distinguishing between different odorants and various odor intensities; and (2) describing (or rating) odor intensity, character, and offensiveness in a way that best reflects that of the average residential population. These two criteria sometimes conflict, and an appropriate balance is needed.

EXPERIMENTAL

Odor Test

The odor test comprised three main parts: (1) distinguishing between different odorants by means of a triangular forced-choice test; (2) evaluating odor intensity; and (3) describing hedonic tone and odor character. The stimuli included n-butanol (Bio-Lab, Israel; AR, 99.5% purity), p-cresol (Fluka, ≥99% purity), butanoic acid (PFW Chemicals, The Netherlands), and d-limonene (Frutarom, Israel). n-Butanol was selected because of its long history as a standard in olfactometry; p-cresol and butanoic acid were selected as two odorants commonly found in agricultural (manure) odors, and d-limonene was selected because it is widely used in flavor extracts, cleaning agents, and in the cosmetics industry. Personal details, including age, gender, and residential address, were requested for statistical analysis. The actual typewritten form used in the test can be found in Laor et al.14

Part 1—Triangular Forced-Choice Test

The purpose of part 1 was to determine if the assessor could correctly differentiate between two odorants with differing odor characteristics. This part of the test used five sets of odorants, each comprising three solutions, two of which were identical. The assessor was asked to select the odd solution out of each set. The sets comprised the following solutions: set I: one p-cresol/two n-butanol; set II: one n-butanol/two p-cresol; set III: one d-limonene/two p-cresol; set IV: one n-butanol/two d-limonene; and set V: one d-limonene/two n-butanol. The liquid (aqueous) concentrations of the solutions were the following: p-cresol, 250 mg/L (the pure compound is solid at room temperature); n-butanol, 2000 ppmv (1620 mg/L); and d-limonene, 200 ppmv (170 mg/L). Stock solutions were freshly prepared with deionized water the day before the test, and 60-mL aliquots were transferred into 100-mL wide-mouth glass bottles with screw caps. Each bottle was sniffed by individual candidates after being briefly swirled and opened for a few seconds. After this part of the test, candidates were asked to rest for 10 min before they entered the second part of the test.

Part 2—Odor Intensity Evaluation

The purpose of part 2 was to determine if the assessor could correctly distinguish between various odor intensities, and how he or she describes (or rates) odor intensity relative to the entire tested population. A 10-point static logarithmic scale was prepared, comprising bottles of n-butanol solution in deionized water, at the following liquid concentrations: 125, 250, 500, 1000, 2000, 4000, 8000, 16,000, 32,000, and 64,000 ppmv (101.3–51,840 mg/L). Although the last two bottles pose relatively high air concentrations compared to the more commonly used 8- or 5-point scales (up to 15,500 or 20,250 ppmv liquid concentrations, respectively),15 health concerns were not considered a major issue in this case, because the present scale was used only for the screening test and was not used as an “odor intensity reference scale” for routine odor quantification.16

Part 2A. The test administrator removed one of the bottles (the fourth, fifth, or sixth) from the series and arranged the remaining nine bottles in order in a single row. The candidate was asked to place the removed bottle back in its correct position in the series, that is, between two bottles.

Part 2B. The candidate was asked to rate the odor intensity of each of the 10 bottles on a 0–6 scale, on which 0 = no odor, 1 = very faint, 2 = faint, 3 = moderate, 4 = strong, 5 = very strong, and 6 = intolerable. To minimize confusion, the candidate was asked to use the above descriptive words, which the administrator subsequently translated into their corresponding intensity numbers.

Part 3—Hedonic Tone and Odor Character

The purpose of part 3 was to determine if the assessor could correctly identify subjective measures of odor such as offensiveness. Hedonic tone descriptions (degree of odor pleasantness or offensiveness) were applied according to a scale from −4 (most offensive) to +4 (most pleasant).

Part 3A. The candidate was asked to evaluate the hedonic tones of two n-butanol concentrations taken from the 10-point static scale of part 2. The test administrator specifically selected for each candidate one bottle representing “weak odor” and a second representing “strong odor”, based on the descriptions allocated by the same candidate in part 2B of the test.

Part 3B(I). The candidate was asked to evaluate the hedonic tones of four aqueous solutions at the following liquid concentrations: p-cresol (500 mg/L), d-limonene (400 ppmv; 340 mg/L), butanoic acid (400 ppmv; 384 mg/L), and n-butanol (16,000 ppmv; 12,960 mg/L).

Part 3B(II). The candidate was asked to describe the odor characters of the above four solutions by selecting descriptive words from the following list: (1) chemicals or solvents; (2) pungent; (3) offensive; (4) unpleasant; (5) medicinal; (6) citrus; (7) pleasant; (8) perfume; (9) herbal; (10) other. The words listed on items 3, 4, and 7 essentially do not describe specific odor characters. These words were added to the list for those candidates who had difficulties in finding proper descriptive words. Because this part of the test was analyzed based on the result of each candidate relative to the entire tested population (as explained in detail in the section Grading Procedure), it was still useful to identify candidates of more or less typical odor perception.

The Tested Population

Over 200 individuals participated in this study, and the test results of 179 of them were included in the statistics. This group was selected as comprising similar numbers of women and men, with identical distributions of age and urban/rural residence. Individuals who reported temporary health problems (such as flu-like symptoms) did not participate in the test. The finally selected group comprised 48% women and 52% men, whose age and residence distributions were identical between women and men, except for the oldest group (aged 61–70), which comprised only 11 individuals (2 women and 9 men) (Figure 1). Overall, the tested population comprised 133 and 46 individuals of urban and rural residence, respectively. The bias toward urban candidates seems inevitable for a country like Israel, in which over 90% of the population is categorized as urban (based on the number of residents).17 Yet, because rural and suburban populations commonly face odor issues related to agricultural activities, we found it appropriate to increase the fraction of rural candidates relative to the fraction they make up the national population.

Figure 1. Gender, age, and residence distributions of the tested population. (A) Age and gender distributions. (B) Residence distribution of women. (C) Residence distribution of men. The total number of individuals was 179. Age and residence were evenly distributed between women and men except for the age group 61–70.

Grading Procedure

This screening tool was developed in order to select individuals with the ability to discriminate between different type of odors and various odor intensities, and with “average odor perception”. Thus, grading was based on two criteria: (1) providing correct answers in the relevant parts of the test (parts 1 and 2A); and (2) providing descriptions of odor intensity, hedonic tone, and character similar to the average of those provided by all members of the tested population (parts 2B, 3A, 3B(I) and 3B(II)). The latter was evaluated by first averaging the results of all 179 tested individuals and then grading each individual according to closeness to that average. The full grading procedure is detailed in Table 1 and an example of the calculations involved in part 2B is given in Table 2. Averaging the results at this stage did not include an outlier analysis. Instead, individuals who reported atypically received lower scores as detailed in Table 2.

Table 1. Grading procedure in each part of the odor test

Part 1	Forced choice: 8 points for each correct answer (5 sets). Maximum: 40 points.
Part 2A	Odor intensity: 30 points for correct answer. 8 points are deducted for one-place position error, 16 points for two-place error, etc. Maximum: 30 points.
Part 2B	Odor intensity: The average rating of each n-butanol concentration by the tested population (179 individuals) is determined first. Then, for every individual, the distances of all ratings from the average are summed over all solutions. The final grading is as follows: 12 points are given if the sum of these distances is in the range 0–2.5; 11 points for 2.5–5; 9 points for 5–7.5; 6 points for 7.5–10; and 2 points for 10-12.5. Maximum: 12 points (an example is shown in Table 2).
Part 3A	Hedonic tone: The average ratings of the two selected n-butanol solutions by the tested population (179 individuals) are determined first. Then, for every individual, the distances of ratings for these two solutions from the average are summed. The final grading is as follows: 7 points are given if the sum of the two distances is in the range 0–2.5; 6 points for 2.5–5; 4 points for 5–7.5; and 1 point for 7.5–10. Maximum: 7 points.
Part 3B(I)	Hedonic tone: The average rating of each odor type by the tested population (179 individuals) is determined first. Then, for every individual, the distances of ratings for each solution from the average are summed. The final grading is as follows: 7 points are given if the sum of distances is in the range 0–2.5; 6 points for 2.5–5; 4 points for 5–7.5; and 1 point for 7.5–10. Maximum: 7 points.
Part 3B(II)	Odor character: The description of every solution most frequently selected by the tested population (179 individuals) is determined first. For each individual, 1 point is given if he or she uses one of the most frequently used descriptions of each solution; 0.5 point if he or she uses one of the less frequently used descriptions; and no point if he or she uses a rarely used description. Maximum: 4 points (1 point for each solution).
Total	Maximum: 100 points

Table 2. An example of the calculations used in part 2B of the odor test to determine a grading based on “average odor perception.”

Bottle No.	1	2	3	4	5	6	7	8	9	10
Candidate's evaluation (0–6 scale)	0	0	1	2	3	4	5	5	6	6
Average of the population	0.62	1.07	1.59	1.98	2.68	3.20	3.85	4.42	5.01	5.47
Distance from average (absolute values)	0.62	1.07	0.59	0.02	0.32	0.80	1.15	0.58	0.99	0.53
Sum of distances = 6.66
Notes: In this example, because the sum of distances appeared to be 6.66, this candidate received 9 points out of 12, as detailed in Table 1.

The maximum points given to each part (Table 1) was generally based on the time and efforts devoted for each part of the test: parts 1 and 2A (accounted for maximum 70% of the total score) comprised a total of 25 bottles; 15 in part 1 and 10 in part 2A. On the other hand, parts 2B and 3 (accounted for maximum 30% of the total score) comprised a total of 16 bottles; 10 in part 2B (which were only sniffed briefly, as these bottles were already sniffed in part 2A), 2 in part 3A (already were sniffed in part 2 as well), and 4 bottles in parts 3B(I) and 3B(II).

RESULTS AND DISCUSSION

Test Results

The test results of all 179 individuals are presented in Figure 2A–F.

Figure 2. Test results for all 179 individuals. (A) Part 1: Triangular forced choice—percentage of correct answers given for each set (maximum: 40 points). (B) Part 2A: Odor intensity—replacing 1 of the 10 bottles previously removed from a 10-point static scale of n-butanol, scored according to the distance from the correct position (maximum: 30 points). (C) Part 2B: Odor intensity—describing intensity of n-butanol samples comprising a 10-point static scale (maximum: 12 points). (D) Part 3A: Hedonic tone descriptions of two n-butanol samples—two bottles selected for each individual, previously described by the individual as “weak” and “strong”, respectively (maximum: 7 points). (E) Part 3B(I): Hedonic tone descriptions of four different solutions (maximum: 7 points). (F) Part 3B(II): Odor character—three descriptors most frequently given to d-limonene, butanoic acid, p-cresol, and n-butanol (maximum: 4 points).

Part 1: Odor Recognition in Triangular Forced-Choice Test

(Figure 2A). Between 73% and 94% of the individuals correctly identified the odd substances in the various triangular sets in this part of the test. In general, the distinctions between d-limonene and p-cresol and between d-limonene and n-butanol (sets III, IV, and V) were more successful than that between p-cresol and n-butanol (sets I and II). The lowest success rate was reported for set II, with only 73% of individuals responding correctly. This was presumably related to odor fatigue that resulted from sniffing the bottles of set I, which contained the same odorants as those in set II. The highest success rate (93–94%) was observed with the first two sets containing d-limonene. There was some decrease in the percentage of correct answers given to set V (88%), again presumably because of odor fatigue resulting from sniffing the previous two sets, which also included d-limonene. The effect of odor fatigue is relevant for field assessors, who are exposed to suprathreshold odor intensities, in contrast to laboratory panelists, who work in a relatively clean environment and smell odorous samples at near threshold levels. With regard to the responses to this part: 51.4% of the candidates answered all five sets correctly, 34.1% made one mistake, 11.7% made two mistakes, and 2.8% made three mistakes.

Part 2: Odor Intensity

Part 2A: Replacing one removed bottle in position in a 10-point static scale (Figure 2B). The removed bottle was replaced in the correct position (designated as distance “0”) by 82.1% of the candidates, 16.8% were wrong by one position, and only 1.1% were wrong by two positions.

Part 2B: Describing odor intensity (Figure 2C). A linear relationship was obtained between odor intensity (y) and the log of n-butanol concentration in aqueous solution (x): y = 0.81 ln(x) – 3.47, where odor intensity (y) was described on a 0–6 scale: 0, no odor; 1, very faint; 2, faint; 3, moderate; 4, strong; 5, very strong; and 6, intolerable. Based on this relationship, odor intensities of 0, 1, 2, 3, 4, 5, and 6 corresponded to n-butanol liquid concentrations of 73, 250, 857, 2945, 10,120, 34,772, and 119,470 ppmv, respectively. The error bars represent the variation within the tested population. As detailed in Table 1, full grading was given to results that were closest to the average intensities reported by all 179 panelists for each of the bottles. Guo et al. 9 used a scale from 0 to 5, which they matched against n-butanol liquid concentrations up to 20,250 ppmv, and obtained a similar relationship to the present one (y = 0.9 ln(x) − 4.03; obtained by plotting the data reported in this reference).

It should be noted that the concentrations of n-butanol are reported in the water phase (liquid concentrations), although the stimulus is obviously received from butanol in the gas phase. Based on Henry's constant at room temperature (3 × 10⁻⁴ to 4 × 10⁻⁴; gas–liquid ratio; mass or molar-based concentrations; dimensionless18–20), the ratio between the volume concentrations is nearly 10; that is, 500 ppm n-butanol in liquid will produce about 50 ppm in air. Yet, under the test conditions the gas-phase concentrations can be far lower from those calculated at equilibrium, because of dilution of the headspace gas with ambient air when the bottle is opened and sniffed.21

Part 3: Hedonic Tone

Part 3A: Hedonic tones of two n-butanol concentrations (Figure 2D). For the first four bottles (described by candidates as “faint odor” in part 2B of the test), the average hedonic tone ranged between −0.05 and −0.9. The tones of the remaining bottles (described by candidates as “strong odors” in part 2B of the test) typically ranged between −1.8 and −2.3. Evidently, hedonic perception was directly related to perceived odor intensity (faint or strong) but, nevertheless, the standard deviations for each bottle were fairly large, and several of the “weak” butanol odors were characterized as both pleasant (positive hedonic value) and unpleasant (negative value).

Part 3B(I): Hedonic tones of various solutions (Figure 2E). The average hedonic tones were found to be +2.3 (d-limonene), −1.8 (butanoic acid), −2.2 (p-cresol), and −1.6 (n-butanol). Notably, the butanol solution used in this part of the test was identical to that in the eighth bottle in the static series (liquid concentration of 16,000 ppmv), whose average hedonic tone was rated as −2.3 in part 3A, and then as −1.6 in this last part of the test, that is, significantly different (P < 0.05, Student's t test). This difference could be related to odor fatigue at this stage of the test or due to comparison of other offensive odorants (butanoic acid and p-cresol). Nevertheless, this part of the test was graded based on the similarity between the results of each individual to that of the entire tested population.

Part 3B(II): Odor character of various solutions (Figure 2F). Most candidates (83.2%) described the d-limonene odor as “citrus” and an additional 4.5% found it “pleasant”. The other compounds were described less homogeneously, with typically three odor descriptions used by the majority of the population. Butanoic acid was most frequently described as “offensive” (35.8%), “unpleasant” (24.0%), or “chemicals/solvents” (16.2%). p-Cresol was most frequently described as “chemicals/solvents” (27.4%), “medicinal” (22.9%), or “offensive” (21.2%). Finally, n-butanol was most frequently described as “chemicals/solvents” (44.1%), “unpleasant” (16.2%), or “offensive” (15.1%). Grading of candidates was based on these most frequently used responses, with lowest scores given to those who used descriptive words other than these, and highest scores given to those who used the most frequently used descriptors (Table 1).

Final Grading and Criteria for Certification

A critical function of this screening tool is the determination of a minimum score that would be required to pass the test (i.e., to be certified as a field odor assessor). The average final score of all 179 individuals was 86.1 (standard deviation [SD] = 8.57) and the median was 87.5. As detailed in Table 1, a high score reflects both good ability to distinguish between different types of odors and various odor intensities (parts 1 and 2A) and typical odor perception in terms of describing (or rating) odor intensity, offensiveness, and character (parts 2B and 3). For regulatory purposes, it seems useful to determine what percentage of the population can be certified on the basis of these two criteria. It is proposed herein that candidates who are within the upper 75% would pass the test. This somewhat arbitrary but practical decision suggests that three-quarters of a certain population (represented by assessors who responded well based on the two grading criteria) essentially reflect the “normal” odor perception of that population. In other words, it means that within a community, three out of four people would be considered as “normal receptors” who can report about odor events. Figure 3 presents the final scores of all 179 individuals in ascending order, indicating the median (final score of 87.5) and the 25th (final score of 80.8) percentile. Because of the grading procedure used in this study (Table 1), the scores in the upper range do not necessarily represent the most sensitive individuals (who could be hypersensitive) but rather those who both showed good smelling capability and average odor perception, and thus are considered to well represent the “normal” population. This is in contrast to bell-shaped distributions in which one tail would represent insensitive and the other tail the hypersensitive individuals.

Figure 3. Final scores of all 179 candidates in ascending order. The median (87.50 points) and the 25th percentile (80.75 points) are marked.

With regard to the higher-scoring 75% of candidates, we observed clear effects of gender, residence, and age (Figure 4). Among women and men, 78.6% and 73.8%, respectively, passed the test; and among urban and rural dwellers, 77.4% and only 67.4%, respectively, passed the test. Age was a substantial factor: 89.3% of those in the 21–30 group passed the test, but only 54.6% of those in the 61–70 group. Differences between males and females, and also the impact of age, were reported previously.22 The reason for the higher percentage of urban than of rural residents among the higher scorers may be related to the closer proximity of the latter to odor sources, especially livestock facilities, which would affect their odor perception.

Figure 4. Percentages of candidates who passed the test (i.e., fell within the top 75%), according to gender (A), residence (B), and age (C).

Impact of Each Part of the Test on the Final Score

The impact of each part of the test on the final score is presented in Figure 5. Clearly, the impact of each part depends on its given partial weight, which, as explained in the section Grading Procedure, was based on the time and efforts devoted for each part of the test: the analysis combines parts 1 and 2A, which paired together contributed 70% to the total score (Figure 5A and B), and parts 2B and 3, which paired together contributed 30% to the total score (Figure 5C and D). Thus, the first pair represented “smelling capabilities” in those parts for which an objectively correct or incorrect answer was possible, whereas the second pair represented “average odor perception” in those parts, which were scored according to the proximity of the results to the average rating given by the population. As shown in Figure 5A, the score obtained for parts 1+2A (accounted for maximum 70% of the total score) was highly correlated with the final score (R ² = 0.84). Of the candidates who received a full score for those parts, 100% passed the test (Figure 5B), that is, fell within the upper 75%. The score obtained for parts 2B+3 (accounted for maximum 30% of the total score) was only weakly correlated with the final score (R ² = 0.28; Figure 5C) but the correlation was still significant (P < 0.01), and whether a candidate passed the test was influenced by their score in this part as well (Figure 5D).

Figure 5. Impacts of the two main parts of the test on the final score and on the percentage of candidates who passed the test (i.e., who fell within the top 75%). (A and B) Parts 1+2A (based on a candidate's ability to distinguish between different odor types and various odor intensities). (C and D) Parts 2B+3 (based on a candidate's deviation from the “average odor perception” of the tested population).

It should be noted that by using this proposed scoring system, the final score of the test is more affected by the first criterion (discriminating between different odors and various odor intensities; parts 1 and 2A) than by the criterion of “average odor perception” (parts 2B and 3). Thus, although it is clear that this procedure screens out individuals who are relatively less sensitive, because of the heavy weight given to parts 1 and 2A (70%), it remains questionable whether it effectively excludes hypersensitive people or those individuals having atypical odor perception. Potentially, hypersensitive individuals would be expected to receive a full score on parts 1 and 2A (as they are able to distinguish between different kinds of odors and various odor intensities), but lower scores on the other parts, in which the score achieved is based on the similarity of the response to the average response of the tested population (as such individuals may overrate odor intensity or offensiveness). Figure 6 presents in ascending order the differences between the scores (% of maximum) obtained in these two paired parts of the test. The marked 50th percentile (median) indicates that half of the candidates had a difference above 13.6%, whereas those in the top 5% of candidates had a difference of over 40%.

Figure 6. The differences, in ascending order, between the scores (% of maximum) obtained in parts 1+2A and parts 2B+3. The median indicates that half of the candidates had a difference above or below 13.6, whereas the upper 5% had a difference of over 40. The latter candidates are potentially hypersensitive; they performed excellently in the first part of test (based on ability to distinguish between different odor types and different odor intensities), but relatively poorly in the second part of the test (based on deviation from the “average odor perception” of the tested population).

For example, for a candidate who received 70 points out of the 70 points on parts 1+2A (100% of the maximum score given to this part) but only 17 points out of the 30 points on parts 2B+3 (57% of the maximum score given to this part), the absolute difference between those two parts of the test would be 43, which falls within the top 5%. On the other hand, for a candidate who received 62 points on parts 1+2A (89% of the maximum score given to this part) and 25 points on parts 2B+3 (83% of the maximum score given to this part), the absolute difference between the two parts would be only 5.2, which falls below the top 5%. Note that the total score of both candidates allows them to pass the test based on the proposal to certify all candidates whose total scores are within the top 75% (Figure 3). Yet, the first candidate is suspected as hypersensitive or to have atypical odor perception by showing a large difference between the excellent ability to distinguish between different types of odors and various odor intensities but atypical rating of odor intensity and offensiveness. In light of this analysis, it is suggested that these possibly hypersensitive candidates be excluded from being certified as field assessors.

Additional Regulatory Concerns

The screening tool developed in this study seems useful for regulatory purposes. For example, the Israeli Prevention of Nuisance Law, 5721-1961 (the Kanowitz Law), Clause 3, Odor Prevention, states, “No individual will cause an intense or unreasonable odor, from any source, which disturbs, or might disturb an individual nearby or disturb passersby.” Unfortunately, the terms “intense” (strong) and “unreasonable” are not defined by the law, with the result that enforcement of this law is often problematic. In the present study, of those candidates who passed the test, that is, 134 out of 179 according to the first screening, or 126 out of 179 when possibly hypersensitive individuals were excluded by the additional screening, 65% described bottle 7 in the n-butanol series (liquid concentration of 8000 ppmv) as at least “strong” odor and over 90% described the rest of the bottles in this series as at least “strong” (Figure 7). This implies that a team of four certified odor assessors that reports a “strong odor” should be considered as reflecting the “average odor perception” of the community and, therefore, their report should be taken as evidence of a violation of the Odor Prevention Law. In fact, no special training is needed for participation in these certified teams; the assessors' natural odor perception as evident by the proposed screening tool must be considered sufficient to identify violations of the Law.

Figure 7. Percentages of individuals who rated odor intensity as at least “strong” for each of the bottles in the n-butanol static scale (part 2B).

In conclusion, use of this screening tool for selection of field odor assessors to serve the general community, regulatory agencies, and other environmental authorities seems practicable, and it was proposed to Israel's Ministry of Environmental Protection as a national standard tool. The test is very simple to perform and does not need any special laboratory; however, because of expected cultural variations in odor perception, local calibration of the test is recommended. This calibration is needed to obtain odor-perception data representative of the community that can be used to calculate the average responses to parts 2B and 3, for which scoring is dependent on the proximity of individual responses to the average response of the relevant population.

ACKNOWLEDGMENTS

This work was supported by the Israel Ministry of Environmental Protection (2-03) and the Israel Ministry of Agriculture (301-0497-06). The authors would like to thank four anonymous reviewers for their valuable comments.