Effect of Frying Instructions for Food Handlers on AcrylamideConcentration in French Fries: An Explorative Study

M. SANNY,1,2 P. A. LUNING,1* S. JINAP,2 E. J. BAKKER,3 AND M. A. J. S. VAN BOEKEL1

1Product Design and Quality Management Group, Department of Agrotechnology and Food Sciences, Wageningen University, P.O. Box 8129, Bomenweg 2,

NL-6700 EV Wageningen, The Netherlands; 2Centre of Excellence for Food Safety Research, Faculty of Food Science and Technology, Universiti PutraMalaysia, 43400 Serdang, Selangor, Malaysia; and 3Biometris, Applied Statistics, Department of Plant Sciences, Wageningen University, P.O. Box 100,

6708 PD Wageningen, The Netherlands

MS 12-049: Received 27 January 2012/Accepted 10 August 2012

ABSTRACT

The objective of this study was to obtain insight into the effect of frying instructions on food handlers’ control decisions in

restaurants and to investigate the impact of control decisions on the variation and concentration of acrylamide in French fries. The

concentrations of acrylamide and reducing sugars were analyzed, the frying temperature and time were measured, and thawing

practices were observed. The results obtained before and after instructions were provided to the food handlers were compared for

restaurants as a group and for each restaurant. Frying instructions supported food handlers’ decisions to start frying when the oil

temperature reached 175uC; all handlers started frying at the correct temperature. However, the effect of the instructions on the

food handlers’ decisions for frying time differed; most handlers increased the frying time beyond 240 s to achieve crispier French

fries with a final color dictated by their preference. Providing instructions did not result in a significant difference in the mean

concentration of acrylamide in French fries for the restaurants as a group. However, data analyzed for each restaurant revealed

that when food handlers properly followed the instructions, the mean concentration of acrylamide was significantly lower (169 mg/

kg) than that before instructions were provided (1,517 mg/kg). When food handlers did not complying with the frying instructions,

mean acrylamide concentrations were even higher than those before instructions were provided. Two different strategies were

developed to overcome the noncompliant behavior of food handlers: establishing requirements for the features of commercial

fryers and strict monitoring of compliance with instructions.

Potato-based products, such as French fries, can contain

high and variable concentrations of acrylamide, a compound

that is probably carcinogenic for humans (7, 11, 20).Acrylamide concentrations in French fries as high as

3,500 mg/kg have been reported in the European Union

acrylamide monitoring database (24). French fries are a

predominant source of dietary exposure of acrylamide because

they are widely consumed (10). The high acrylamide

concentrations combined with high consumption could increase

the dietary exposure for an individual by a factor of 5 over the

average exposure of 2.8 mg/day (15). Control measures to

reduce acrylamide exposure should therefore focus on food

products that are both widely consumed and that contain high

acrylamide concentrations, such as French fries (15).Various studies performed under controlled laboratory

conditions have established the effect of lowering the

concentration of reducing sugars on the reduction of

acrylamide formation (10, 16). A study to validate the

effectiveness of lowering sugar concentrations as a control

measure in real food service establishments (FSEs) revealed

that a lower mean concentration of acrylamide was obtained

using par-fried potato strips with a low concentration of

reducing sugars compared with commonly used potato

strips (33). However, in addition to the concentration of

reducing sugars, frying conditions (time and temperature)

are major technological factors that influence acrylamide

concentration, and food handlers’ variable and inadequate

decisions concerning these conditions could also contribute

to large variations in acrylamide concentrations (35). Sanny

and coauthors (34) found that differences in frying

equipment (digitally controlled fryers versus simple frying

pans) and frying instructions (oral instructions versus no

instructions at all) resulted in considerable differences in

frying practices among food handlers in FSEs and to large

variations and high acrylamide concentrations in French

fries. Various authors have found that the actual frying

practices among FSEs can differ considerably (1, 13, 25).

For example, the setting of the frying temperature in fast-

food establishments in New Zealand ranged from 175 to

190uC but the actual frying temperature ranged from 136 to

233uC (25). In another study, the frying temperatures in

nearly 70% of restaurants and public catering facilities in

Budapest were not controlled (13). Researchers have argued

that food handlers’ variable and inadequate decisions could

have a considerable impact on food quality, but the effects

of these decisions have not been well studied (21–23).* Author for correspondence. Tel: z31 317 482087; Fax: z31 317

483669; E-mail: pieternel.luning@wur.nl.

462

Journal of Food Protection, Vol. 76, No. 3, 2013, Pages 462–472doi:10.4315/0362-028X.JFP-12-049Copyright G, International Association for Food Protection

Control decisions are made to keep product properties

and processing conditions between certain acceptable

tolerances and, when necessary, to take corrective actions

(8, 22). Food handlers in an FSE do the actual frying, and

they make decisions daily to control product properties (e.g.,

sorting out fines or brown pieces of French fries) and

processing conditions (e.g., reducing frying time when

frying a smaller portion of frozen par-fried potato strips) (5,16). Likewise, managers can control food handlers’

behaviors by making decisions concerning noncompliance

(e.g., out-of-control actions) and administrative conditions

(e.g., providing procedures, instructions, decision-support

tools, and training). Administrative conditions are tailored to

direct and harmonize decision-making behavior (22). In

many cases, food handlers in restaurants use frying pans to

fry French fries. With a frying pan, handlers can decide only

when to start the frying (by visually inspecting the oil to

estimate its temperature) and when to end the frying (by

visually checking the color of the French fries). To our

knowledge, no data have been published on the quantitative

contribution of food handlers’ decisions during frying to the

concentration and variation of acrylamide. The objective of

this study was to obtain insight into the effect of frying

instructions on food handlers’ control decisions in restau-

rant-type FSEs and to investigate the impact of control

decisions on the concentration and variation of acrylamide

in French fries. Frying instructions that specified acceptable

frying practices were implemented to influence the food

handlers’ daily decisions concerning frying conditions, and

the extent of compliance with the frying instructions was

observed. Providing information about the required settings

and acceptable tolerances in the frying instructions can

support food handlers in adequately controlling the required

product properties and processing conditions. This control

may enable a more homogenous decision-making in out-of-

control situations to result in reduced variation in acrylam-

ide concentrations.

MATERIALS AND METHODS

Characteristics of restaurants. The study focused on

restaurants located within a 25-km radius of Serdang, Selangor,

Malaysia. This city is a typical urban area. Ten restaurants that

each had more than five workers were selected. The restaurants

types were family style, vegetarian, and cafe.

Development of instructions. The frying instructions were

developed using data from the literature and were designed to

include visual representation in the form of schematic diagrams

and pictures to make it easy for the food handlers to understand the

concepts and information. Thawing for a minimum of 2 h at room

temperature was stipulated in accordance with usual practice in

fast-food establishments (30). A frying temperature of 175uC was

selected because this temperature is recommended for restaurants

(5, 9, 14, 43). A frying time of 240 s (4 min) was selected because

the acrylamide concentration begins to increase exponentially after

4 min of frying at 180uC (32). A color card was included in the

frying instructions to illustrate the target color based on the advice

in the acrylamide ‘‘toolbox’’ of the Confederation of the Food and

Drink Industries of the European Union (5) to aim for a golden-

yellow color of French fries before ending the frying. We verified

the frying instructions in a preliminary experiment in the

laboratory, and the specified frying conditions resulted in the

target color of golden yellow after a frying time of 4 min at an

initial temperature of 175uC. In the preliminary experiment, which

was conducted using a frying pan, three servings of French fries

were added to 2 liters of oil. Each serving weighed approximately

67 g, i.e., the recommended serving size (38).

Implementation of instructions. The study comprised two

stages of experiments: before instructions and after instructions.

The first experiment was considered as a reference, in which a food

handler was asked to fry three servings of French fries in the usual

manner. In the second experiment, the written frying instructions

(Fig. 1) were explained to the food handlers in a training session.

The frying instructions specified the following: thaw the frozen

par-fried potato strips for a minimum of 2 h at room temperature;

heat oil to a maximum frying temperature of 175uC; place a total of

three servings of potato strips when the oil temperature reached

175uC; fry until the golden-yellow target color of the French fries

was reached, as indicated on the color card (corresponding to a

frying time of ca. 4 min).

In addition to the frying instructions, the food handler

received a digital thermometer and a digital stopwatch to measure

the frying temperature and time. The digital stopwatch produced a

beep to alert the food handler that the frying time of 4 min had

been reached. To increase awareness, the food handlers also

received an explanation about the formation of acrylamide and the

need to follow instructions to reduce exposure to acrylamide.

The training session started with a frying demonstration (first

frying) on how to fry the French fries according to the frying

instructions. Next, the food handler had to practice frying as

demonstrated (second frying) and to reproduce the method in the

third frying session. In the training session, the food handler was

instructed to start frying at 175uC (by regularly checking the

temperature with the digital thermometer). The food handler was also

instructed to stop frying after the target color of golden yellow had

been reached, as illustrated on the color card embedded in the frying

instructions. In addition, the food handler was asked if he understood

the frying instructions to verify that the correct frying procedure had

been learned. During the next three days of French-fry production,

the food handler was instructed to reproduce the frying following the

frying instructions. In each restaurant, the same food handler was

involved in both experiments, i.e., before and after instructions.

We considered the effect of the Hawthorne principle during

the 3 days of French fry production. The Hawthorne effect refers to

the tendency of food handlers to perform more responsibly when

they are participants in an experiment (39). During each day of

French fry production, we observed food handler behavior during

frying of three batches. We verified that the food handlers did not

modify their adherence to the frying instructions in response to

researchers’ presence during the 3 days of French fry production.

The food handlers acted normally when a researcher started to

observe them during frying of the first batch. These same food

handlers continued to act normally while the researcher continued

to observe them during frying of the second and third batches.

Ten restaurants participated in the experiment before in-

structions were given. However, only 7 of the 10 restaurants

participated in the experiment after instructions were given because

3 restaurants went out of business during the course of the study.

Measurements and observations. In all experiments, the

frying temperature and time were measured, and thawing practices

were observed and recorded. The actual (initial) frying temperature

was measured with a digital thermometer 51 series II coupled with

J. Food Prot., Vol. 76, No. 3 FRYING INSTRUCTIONS AS A CONTROL MEASURE IN RESTAURANTS 463

a temperature probe (Fluke Corporation, Everett, WA) just before

par-fried potato strips were inserted into the oil. The temperature

probe was placed into the oil bath at the same location between

measurements and away from the frying pan’s wall. The actual

frying time was measured using an SW01 digital stopwatch (Alba,

Seiko Corp., Tokyo, Japan) just after the removal of French fries

from the oil.

Sampling of French fries. Because we wanted to determine

the effect of instructions on actual frying practices in restaurant

settings that resembled real-life situations, we did not strictly

standardize all parameters (such as initial reducing sugar

concentrations in frozen par-fried potato strips, dimension of the

frying pan, oil volume, and serving size) but rather set some basic

requirements for raw materials, frying oil, and frying portions. The

size of commercial frozen par-fried potato strips was standardized

at a cross-section size of 8 by 8 mm and a length of 60 to 70 mm.

The food handlers had to use raw materials from the same supplier

both before and after receiving frying instructions. A new batch of

palm oil was used for each day of French fry production to avoid

cross-contamination of oil from previous frying bouts.

We did not provide instructions for the oil volume because the

restaurants used frying pans with various dimensions. The food

handlers were allowed to use the serving size and oil volume that

they normally used to reflect the actual practice in a real-life

situation. However, the food handlers had to prepare a portion that

should be approximately the amount sufficient for three servings of

French fries to customers. We measured the actual serving size and

oil volume to record the typical product/oil ratio; 10% is the

recommended ratio for optimizing culinary quality and minimizing

acrylamide concentrations (16). The weight of three servings was

added to obtain the portion size. The portion size and oil volume

were calculated, and the results were expressed as the product/oil

ratio.

In all of the experiments, samples were collected during 3 days

of production. Each day, samples were taken from three different

frying batches. In each frying batch, three servings of French fries

were collected. Of these servings, two were used to determine

acrylamide concentration and the third serving was served to a

customer. Duplicate determinations of acrylamide were performed,

i.e., one extraction was made from each serving. Each serving was

coded and stored in a polyethylene bag at 218uC before analysis.

Sampling of frozen par-fried potato strips. In all of the

experiments, samples of frozen par-fried potato strips were

collected during 3 days of French fry production before frying.

Each day, samples were taken from three different frying batches.

From each frying batch, two serving portions of the frozen par-

fried potato strips were collected. These serving portions were used

to determine the concentration of reducing sugars. Duplicate

determinations of reducing sugars were performed, i.e., one

extraction was made for each serving portion. The weight of each

serving portion was approximately 67 g. Each serving portion was

coded and stored in a polyethylene bag at 218uC before analysis.

Chemicals and stock solutions. Acrylamide (99%) was

purchased from Fluka Chemie AG (Buchs, Switzerland). 13C3-

labeled acrylamide (99% isotopic purity) was obtained from

Cambridge Isotope Laboratories (Andover, MA). Acetonitrile and

methanol were purchased from Merck (Darmstadt, Germany).

FIGURE 1. Scheme of frying instructions.

464 SANNY ET AL. J. Food Prot., Vol. 76, No. 3

Ultrapure water was used throughout the experiments (Purelab

Classic UV, Elga Labwater, Lane End, UK). Solid-phase extraction

cartridges (3 cc; Oasis Hydrophilic-Lipophilic Balance [HLB] and

Oasis Mixed-mode Cation-eXchange [MCX]) were supplied by

Waters Corp. (Milford, MA).

Stock solutions of acrylamide (0.2 mg/ml) and 13C3-labeled

acrylamide (4 mg/ml) were prepared by dissolution of the

compound in distilled water. Working standards were prepared

by dilution of the stock solution of acrylamide to concentrations of

0.5, 5, and 10 mg/ml with water. All stock solutions and working

standards were stored in a refrigerator at 4uC for a maximum of

3 months.

Method for the analysis of reducing sugars. The concen-

trations of reducing sugars were determined by high-performance

liquid chromatography (HPLC) as described previously (34). The

sample was ground in a blender (Braun Multiquik ZK3, Frankfurt,

Germany), a subsample (2 g) was weighed into a 50-ml centrifuge

tube, and 10 ml of acetonitrile-water (80:20, vol/vol) was added.

The suspension was then centrifuged at 1,643 RCF (| g) for

10 min, and the supernatant was passed through a 0.45-mm-pore-

size nylon syringe filter (Sartorius AG, Gottingen, Germany).

Aliquots (50 ml) of the filtrate were injected into a Waters HPLC

apparatus equipped with a refractive index detector and a

mBondapack NH2 column (3.9 by 300 mm; Waters Corp.). An

isocratic mode of elution was used with a mobile phase of

acetonitrile-water (80:20, vol/vol) at a flow rate of 0.8 ml/min.

Maltose was used as an internal standard. The calibration curves

were linear (r2 . 0.999). The detection limit for fructose and

glucose was 5 mg/kg. The recoveries for fructose ranged from 91.3

to 100.3%, and those for glucose ranged from 92.9 to 96.7%. The

intraday and interday precision was expressed as the relative

standard deviations: 4.6 and 6.7% (fructose) and 11.7 and 13.2%

(glucose), respectively.

Method for the analysis of acrylamide. Acrylamide was

determined by liquid chromatography tandem mass spectrometry

(LC-MS/MS) as previously described (33). The sample was ground

in a blender (Braun Multiquik ZK3), a subsample (2 g) was

weighed into a 50-ml centrifuge tube, and 10 ml of water that

contained 500 ng of 13C3-labeled acrylamide as the internal

standard (final concentration of 50 ng/ml) was added. The mixture

was shaken at medium speed (ca. 256 pulses per min) on a vertical

shaker (RS-1, Jeio Tech, Gyeonggi-do, Korea) for 10 min and then

centrifuged in a refrigerated centrifuge (3-18K, Sigma, Gillingham

Dorset, UK) at 10,956 RCF (| g) for 30 min. An approximately 2-

ml aliquot beneath the oil layer was collected with a syringe and

filtered through a 0.22-mm-pore-size nylon syringe filter (Sartorius

AG), and the filtrate was collected. Both the HLB and MCX

columns were conditioned with 2 ml of methanol and equilibrated

with 2 ml of water. The filtrate (1.0 ml) was loaded onto an HLB

cartridge, allowed to pass through the cartridge, and discarded. The

HLB cartridge was then washed with 1.0 ml of water, and this

eluate (eluate 1) was collected and allowed to pass through an

MCX cartridge. This eluate (eluate 2) was collected and transferred

to an amber vial for HPLC-MS/MS analysis.

Sample extracts and calibration standards were injected into

a Surveyor HPLC system equipped with a Hypercarb column

(2.1 mm by 50 mm by 5 mm; ThermoFisher Scientific, San

Jose, CA). The acrylamide was detected on a Thermo Scientific

TSQ Quantum Ultra triple quadrupole mass spectrometer,

which was operated in positive atmospheric pressure chemical

ionization mode. The standards contained acrylamide at

concentrations of 1, 5, 10, 25, 50, 100, 250, 500, and

1,000 ng/ml and isotopically labeled acrylamide at 50 ng/ml.

The injection volume was 10 ml.

The acrylamide was separated under isocratic conditions

using 100% water as the mobile phase at a flow rate of 0.15 ml/

min. The ion-transfer capillary temperature was maintained at

250uC, the vaporizer temperature was 375uC, and the discharge

current was 5 mA. The argon collision gas pressure was adjusted to

1.5 mTorr for the MS/MS. The collision energy was maintained at

9 eV for each monitored transition in selective reaction monitoring

mode. The monitored MS/MS transitions were m/z 72 . 55 for

acrylamide and 75 . 58 for 13C3-acrylamide. The scan time for

each monitored transition was maintained at 0.3 s, the scan width

was maintained at 0.010 m/z, and the Q1, Q3 resolution was

maintained at 0.7 m/z full width at half-mass.

The transitions m/z 72 . 55 for acrylamide and 75 . 58 for13C3-acrylamide were used for quantification, and m/z 72 . 55, 72

. 54, and 72 . 44 were used for confirmation of the peak identity.

A calibration graph was constructed by plotting the peak area of

acrylamide relative to that of the internal standard against the

corresponding ratios of analyte amounts. Acrylamide concentra-

tions in the sample extracts were calculated from the calibration

slopes and intercept values. The calibration curves were linear

(r2 . 0.999). The limit of detection was reached at 1 mg/kg, and

the recoveries were in the range of 100 to 108%. The intraday

and interday precision was expressed as the relative standard

deviations: 4.4 and 3.9%, respectively.

Statistical analysis. The original data for the acrylamide

concentration in each serving of the French fries samples were loge

transformed to construct histograms. The loge transformation was

performed to approximate a more normal distribution of the data

while retaining information about the degree of variation in the

distribution (6, 12, 18).

Three of the 10 restaurants did not participate in the

experiment after instructions were given because they had gone

out of business during the course of the study and were therefore

excluded from data analysis. For each restaurant, the mean of the

loge-transformed acrylamide concentration for the experiments

before and after instructions was calculated. Using this pair of

means, a paired-sample t test was performed to analyze differences

in the mean acrylamide concentration for the group of restaurants

before and after instructions. The obtained means of loge-

transformed acrylamide concentration were back-transformed to

an original scale of measurement and used as the geometric means

(2, 28). The coefficient of variation (CV) was calculated using the

formula 100(eSD 2 1), where SD is the standard deviation of

the loge-transformed acrylamide concentration (6, 18). The CV

compares the degree of variation for a group of restaurants before

and after instructions. The paired-sample t test also was used to

analyze differences in the frying temperature, the frying time, the

concentrations of reducing sugars, and the product/oil ratio for a

group of restaurants before and after instructions. The CV was

calculated as the standard deviation divided by the mean and

reported as a percentage.

An analysis of variance (ANOVA) with a mixed model was

used to analyze the pair of means of acrylamide concentration for

each restaurant before and after instructions. The experiment type

(before or after instructions) was included as a fixed effect, but the

establishment, sampling day, and frying batch were used as

random variables. The mixed model was selected because the

establishments, sampling days, and frying batches were not fixed

in advance, and their effects were therefore considered random. All

random variables were nested within the sampling level. A mixed

ANOVA model with the same variables was used to analyze the

J. Food Prot., Vol. 76, No. 3 FRYING INSTRUCTIONS AS A CONTROL MEASURE IN RESTAURANTS 465

pair of means of frying temperature and frying time for each

restaurant.

Bivariate correlations were used to investigate the strength

and direction of the relationship of acrylamide formation with

respect to a set of influencing factors (frying temperature, frying

time, concentration of reducing sugars, thawing time, and

product/oil ratio). The results were expressed as Pearson

correlations. Multiple linear regression analysis was used to

investigate the predictive ability of the set of influencing factors

and to assess the relative contribution of each influencing factor to

acrylamide formation. These influencing factors, considered to be

possible predictors for acrylamide formation, were included in the

model. Multiple regression analysis was checked for possible

violations from the model assumptions during analysis. In tests, a

P value of 0.05 or less was considered significant. Statistical

analyses were performed using SPSS version 16.0 (SPSS Inc.,

Chicago, IL).

RESULTS

The food handlers indicated that the frying instructions

were clear. When they did not comply with the frying

instructions (i.e., they extended the frying time beyond

240 s), they were asked to explain the reasons for their

behavior after they finished frying. The typical answers

were to obtain crispier French fries and to achieve a more

appropriate color according to their own preference.

The concentrations of reducing sugars in the par-fried

potato strips were measured to verify that they were within

the range commonly found in restaurants. The mean con-

centration of reducing sugars in potato strips was lower in

the experiment with instructions (0.97 g/kg) than in the

experiment without instructions (1.27 g/kg); however, the

difference was not significant (P ~ 0.327) (Table 1). The

CV of the concentration of reducing sugars was higher in

the experiment with instructions (79.4%) than that without

instructions (51.2%).

The mean of the actual frying temperature for res-

taurants as a group was lower after instructions (176uC) than

before instructions (185uC); however, the difference was not

significant (P ~ 0.113) (Table 2). The CV for the actual

frying temperature was also lower after instructions (0.68%)

than before instructions (7.5%). The actual frying temper-

ature for restaurants as a group ranged from 175 to 185uCafter instructions and from 145 to 235uC before instructions.

When the data for each restaurant were analyzed, the mean

of the actual frying temperature for restaurants A, C, D, E,

and F was significantly lower (P , 0.05) after instructions

than before instructions, whereas for restaurant G, the

temperature was significantly higher. No significant differ-

ence was observed in the mean of the actual frying

temperature for restaurant B.

The mean of the actual frying time for restaurants as a

group was longer after instructions (386 s) than before

instructions (365 s); however, the difference was not

significant (P ~ 0.654) (Table 3). The CV for the actual

frying time was higher after instructions (27.2%) than

before instructions (19.9%). The actual frying time for

restaurants as a group ranged from 240 to 644 s after

instructions and from 221 to 694 s before instructions.

When the data for each restaurant were analyzed, the mean

of the actual frying time for restaurants E and F was

significantly lower (P , 0.05) after instructions than before

instructions, whereas the mean times for restaurants A, D,

and G were significantly higher. No significant difference

was observed in the mean of the actual frying time for

restaurants B and C. Before instructions, four of the seven

restaurants thawed the frozen par-fried potato strips; after

instructions, all restaurants thawed the strips for a minimum

of 2 h at room temperature.

The serving size and oil volume were measured to

obtain insight into the actual product/oil ratio used by the

TABLE 1. Concentrations of reducing sugars in par-fried potato strips from various restaurants before and after instructionswere provided

Restaurant

Reducing sugars (g/kg fresh wt)

Before instruction After instruction

Mean (SD)a CV (%)b Minimum Maximum Mean (SD) CV (%) Minimum Maximum

A 1.47 (0.21) 14.3 1.21 1.94 0.12 (0.01) 8.3 0.10 0.15

B 1.28 (0.16) 12.5 0.81 1.38 0.48 (0.06) 12.5 0.41 0.58

C 2.46 (0.30) 12.2 1.69 2.80 1.56 (0.05) 3.2 1.48 1.66

D 1.00 (0.17) 17.0 0.76 1.27 0.78 (0.11) 14.1 0.58 0.99

E 1.61 (0.15) 9.3 1.28 1.84 2.40 (0.12) 5.0 2.23 2.64

F 0.68 (0.06) 8.8 0.58 0.77 0.09 (0.03) 33.3 0.06 0.16

G 0.53 (0.15) 28.3 0.33 0.80 0.56 (0.10) 17.9 0.37 0.68

H 1.00 (0.08) 8.0 0.91 1.19 NDc

I 1.31 (0.14) 10.7 1.08 1.54 ND

J 1.49 (0.19) 12.8 1.21 1.78 ND

Totald 1.27 (0.65) 51.2 0.33 2.80 0.97 (0.77) 79.4 0.06 2.64

a Mean values based on n ~ 18.b CV (coefficient of variation) was defined as the standard deviation divided by the mean, the result of which is reported as a percentage.c ND, no data.d Based on data from seven restaurants (A through G).

466 SANNY ET AL. J. Food Prot., Vol. 76, No. 3

restaurants. The weight of three servings (data not shown)

was added to obtain portion size. The mean product/oil ratio

was significantly higher (P , 0.05) in the experiment with

instructions (15.4%) than in the experiment without instruc-

tions (14.2%) (Table 4). The CV of the product/oil ratio was

higher in the experiment with instructions (47.4%) than in the

experiment without instructions (41.5%). The product/oil

ratio ranged from 2.1 to 29.5% after instructions and from

2.38 to 26.3% before instructions. The oil volume ranged

from 1.5 to 15 liters both after and before instructions.

Figure 2 illustrates the variation profiles of acrylamide

concentrations for restaurants as a group after and before

instructions. The loge transformation was performed

because our initial data analysis indicated that the frequency

distribution of the acrylamide concentrations was slightly

skewed to the right (data not shown). The mean of the

loge-transformed acrylamide concentration of French fries

was 6.42 mg/kg after instructions and 6.39 mg/kg before

instructions and the standard deviation was lower after

instructions (i.e., 0.85 mg/kg after and 0.91 mg/kg before).

TABLE 2. Actual frying temperatures at various restaurants before and after instructions were provided

Restaurant

Actual frying temp (uC)

Before instruction After instruction

Mean (SD)a CV (%)b Minimum Maximum Mean (SD)c CV (%) Minimum Maximum

A 185 (9) 4.9 170 200 176 (1.9)* 1.1 175 180

B 179 (22) 12.3 148 210 175 (0.4) 0.2 175 176

C 185 (7) 3.8 174 195 176 (1.0)* 0.6 175 178

D 189 (24) 12.7 145 218 176 (0.7)* 0.4 175 177

E 208 (18) 8.7 175 235 179 (2.6)* 1.5 176 185

F 188 (8) 4.3 178 200 175 (0.3)* 0.2 175 176

G 161 (7) 4.3 148 170 175 (0.5)* 0.3 175 176

H 186 (10) 5.4 178 210 NDd

I 192 (13) 6.8 166 210 ND

J 187 (14) 7.5 175 225 ND

Totale 185 (13.9) 7.5 145 235 176 (1.2) 0.68 175 185

a Mean values based on n ~ 9.b CV (coefficient of variation) was defined as the standard deviation divided by the mean, the result of which is reported as a percentage.c Means with asterisks are significantly different from those before instruction at the 0.05 level.d ND, no data.e Based on data from seven restaurants (A through G).

TABLE 3. Actual frying time and thawing practices at various restaurants before and after instructions were provided

Restaurant

Before instruction After instruction

Actual frying time (s)

Thawing

Actual frying time (s)

ThawingMean (SD)a CV (%)b Minimum Maximum Mean (SD)c CV (%) Maximum Minimum

A 285 (64) 22.5 222 435 Yes 473 (89)* 18.8 372 644 Yes

B 477 (112) 23.5 322 694 Yes 478 (47) 9.8 423 569 Yes

C 403 (28) 7.0 361 456 Yes 384 (38) 9.9 327 451 Yes

D 361 (42) 11.6 305 421 No 519 (58)* 11.2 425 630 Yes

E 368 (84) 22.8 221 510 No 271 (17)* 6.3 255 312 Yes

F 397 (72) 18.1 262 480 No 262 (25)* 9.5 240 320 Yes

G 264 (10) 3.8 244 276 Yes 319 (46)* 14.4 240 370 Yes

H 244 (27) 11.1 200 271 Yes NDd ND

I 276 (30) 10.9 224 306 No ND ND

J 221 (32) 14.5 170 270 No ND ND

Totale 365 (72.6) 19.9 221 694 4 of 7

restaurants

386 (105) 27.2 240 644 All 7

restaurants

a Mean values based on n ~ 9.b CV (coefficient of variation) was defined as the standard deviation divided by the mean, the result of which is reported as a percentage.c Means with asterisks are significantly different from those before instruction at the 0.05 level.d ND, no data.e Based on data from seven restaurants (A through G).

J. Food Prot., Vol. 76, No. 3 FRYING INSTRUCTIONS AS A CONTROL MEASURE IN RESTAURANTS 467

No significant difference was found in the mean

concentration of acrylamide for restaurants as a group (P~ 0.93) (Table 5). The mean concentration of acrylamide

was higher after (614 mg/kg) than before (596 mg/kg)

instructions; however, the CV of the acrylamide concentration

was lower after instructions (134%) than before instructions

(148%). The acrylamide concentration ranged from 85 to

3,369 mg/kg after instructions and from 77 to 4,283 mg/kg

before instructions. When data for each restaurant were

analyzed, the mean concentration of acrylamide for restaurant

E was significantly lower (P , 0.05) after instructions

(169 mg/kg) than before instructions (1,517 mg/kg). The mean

concentrations of acrylamide after instructions for restaurants

B, C, and F were significantly higher (715, 961, and 239 mg/

kg, respectively) than before instructions (271, 506, and

134 mg/kg, respectively). For restaurants A, D, and G, no

significant difference in acrylamide concentration was found

before and after instructions.

A small but significant correlation was found between

the frying time and the acrylamide concentration (r ~

0.228, n ~ 360, P , 0.05) (Table 6). A small negative

correlation was found between the concentration of reducing

sugars and the acrylamide concentration (r ~ 20.161, n ~

360, P , 0.05), and a small but significant correlation was

found between thawing practice and acrylamide concentra-

tion (r ~ 0.142, n ~ 360, P , 0.05). Both frying

temperature (r ~ 0.013, n ~ 360, P ~ 0.816) and product/

oil ratio (r ~ 0.033, n ~ 360, P ~ 0.570) were not

significantly correlated with the acrylamide concentration.

The multiple linear regression model revealed that all

factors (except the product/oil ratio) significantly contribut-

ed to the prediction of acrylamide formation (r2 ~ 0.104,

n ~ 360, P , 0.05). The order of contributions, from

highest to lowest, are frying time (beta ~ 0.259), thawing

practice (beta ~ 0.179), frying temperature (beta ~ 0.173),

and reducing sugars (beta ~ 20.125).

DISCUSSION

To our knowledge, this study is one of the first to

quantify the effect of frying instructions on food handlers’

control decisions with respect to food quality. After

instructions were given, a lower mean frying temperature

(176uC) with a smaller CV (0.68%) was found compared

with before instructions were given (185uC, 7.5%),

although the differences in mean frying temperature and

time for the restaurants as a group before and after

instructions were not significant. A longer mean frying

time (386 s) with a larger CV (27.2%) was found after

instructions compared with before instructions (365 s,

19.9%). Bivariate correlations and multiple linear regression

analyses revealed that frying time was significantly (but

weakly) correlated with acrylamide concentration (r ~

0.228, P , 0.05) and contributed the most to the prediction

of acrylamide formation (beta ~ 0.221, P , 0.05). Frying

temperature was not significantly correlated with acrylamide

concentration (r ~ 0.013, P ~ 0.408); however, temper-

ature made a significant (but small) contribution to the

prediction of acrylamide formation (r ~ 0.153, P , 0.05;

Table 6). When we analyzed the data for each restaurant, we

found that the mean frying temperature in all restaurants

(except restaurant G) was significantly lower after instruc-

tions than before instructions (Table 2). Apparently, the

instructions and measuring equipment supported food

handlers’ decisions to start frying when the oil reached

175uC, but an inconsistent effect of the instructions on the

food handlers’ decisions to end frying was observed. For

TABLE 4. Actual product/oil ratio, portion size, and oil volume at various restaurants before and after instructions were provided

Restaurant

Before instruction After instruction

Oil volume

(liters)d

Product/oil ratio (%)

Mean (SD)

portion size (g)c

Product/oil ratio (%)

Mean (SD)

portion size (g)

Mean

(SD)aCV

(%)b Minimum Maximum

Mean

(SD)

CV

(%) Maximum Minimum

A 20.7 (3.1) 14.9 17.2 26.3 415 (62) 22.1 (0.9) 4.1 20.0 23.0 442 (20) 2

B 11.1 (0.8) 7.2 9.96 11.8 561 (68) 10.0 (1.2) 12.0 8.57 12.5 499 (58) 5

C 18.9 (1.4) 7.4 17.0 21.5 473 (36) 14.7 (0.9) 6.1 13.4 15.8 368 (22) 2.5

D 16.9 (1.8) 10.7 14.8 20.0 254 (27) 25.4 (2.5) 9.8 21.2 29.5 381 (37) 1.5

E 15.3 (1.4) 9.2 13.5 17.6 459 (43) 15.6 (1.5) 9.6 13.5 18 468 (45) 3

F 14.1 (2.8) 19.9 8.48 17.9 415 (64) 17.4 (2.4) 13.8 15.1 22.6 523 (73) 3

G 2.6 (0.2) 7.7 2.38 2.97 387 (28) 2.4 (0.2) 8.3 2.10 2.77 364 (34) 15

H 5.1 (0.4) 7.8 4.74 6.0 400 (38) NDe 8

I 13.6 (1.5) 11.0 10.9 15.8 271 (31) ND 2

J 7.6 (1.0) 13.2 5.78 9.04 303 (32) ND 4

Total f 14.2 (5.9) 41.5 2.38 26.3 423 (100) 15.4 (7.3)* 47.4 2.10 29.5 435 (74) 4.6

a Mean values based on n ~ 9.b CV (coefficient of variation) was defined as the standard deviation divided by the mean, the result of which is reported as a percentage.c The weight of three servings was added to obtain portion size.d Data for both before and after instruction.e ND, no data.f Based on data from seven restaurants (A through G). Asterisks indicates that means before and after instruction are significantly different

at the 0.05 level.

468 SANNY ET AL. J. Food Prot., Vol. 76, No. 3

two restaurants (E and F), the mean frying time was

significantly lower after instructions than before instruc-

tions; however, for the other three restaurants (A, D, and G),

frying time was significantly higher after instructions. No

significant differences in frying time before and after

instructions were found for restaurants B and C (Table 3).

The highest concentration of acrylamide was 3,369 mg/

kg after instructions were provided and 4,283 mg/kg before

instructions were provided. These concentrations are con-

sistent with data in the acrylamide monitoring database of

the Institute for Reference Materials and Measurements

(19), which includes acrylamide concentrations as high as

4,653 mg/kg from approximately 1,377 samples of French

fries. The variation in acrylamide concentration for the

restaurants as a group before and after instructions were

provided are illustrated in Figure 2. We found no significant

difference in the mean concentrations of acrylamide

between the two experiments (Table 5). Nevertheless, when

analyzing data for each restaurant, we found significant

differences in the means for four restaurants: in one

(restaurant E), the mean concentration after instructions

was lower, whereas in the others (restaurants B, C, and F),

the concentration was higher. The food handler in restaurant

E controlled the frying conditions as instructed and used a

significantly lower mean frying temperature (179uC) and

frying time (271 s) compared with the reference scenario

(208uC, 368 s). As expected, the mean concentration of

acrylamide for this restaurant was significantly lower after

instructions (169 mg/kg) than before instructions (1,517 mg/

kg). However, food handlers in the other restaurants (A, D,

and G) increased the frying time to significantly longer than

instructed and fried the French fries until they were the color

of their own preference to obtain crispier French fries. For

example, the mean frying time in restaurant D was 519 s

(range, 425 to 630 s), which indicates that the food handler

applied longer frying times than the prescribed 240 s. The

mean concentration of acrylamide in restaurant D was

indeed higher after instructions (1,633 mg/kg) than before

instructions (1,353 mg/kg), although the difference was not

significant (P ~ 0.711). Romani et al. (32) reported that

acrylamide concentration increased exponentially with

increasing frying time after approximately 240 s of frying

at 180uC. Although the actual frying temperature was

controlled at 175uC in the restaurants, the frying times were

longer than that prescribed in the instructions, resulting in

higher acrylamide concentrations (but lower CVs). This

result may be explained by the exponential character of the

relationship between the formation of acrylamide and frying

time. Apparently, after the critical time of 240 s, a small

increase in frying time can lead to a large increase in the

concentration of acrylamide.

In the experiments conducted before and after instruc-

tions, the mean concentrations of reducing sugars were

1.27 and 0.97 g/kg, respectively, which is consistent with

findings of Palazoglu and Gokmen (26). Statistical analysis

revealed that the concentration of reducing sugars was

negatively correlated with acrylamide concentration (r ~

20.161, P , 0.05) and significantly contributed to the

prediction of acrylamide formation (r ~ 20.139, P , 0.05;

Table 6). This finding appears to contradict results of

studies conducted under strictly controlled laboratory

conditions in which a strong (positive) correlation was

found between the concentrations of reducing sugars and

acrylamide (2, 5). In the present explorative study, the

restaurants used commercial frozen par-fried potato strips

that they usually purchased from the market and used simple

frying pans, which is the standard practice at the restaurants.

In a previous study (33), we found a strong correlation (r ~

0.758, P , 0.05) between concentrations of reducing sugars

and acrylamide. However, par-fried potato strips were

specifically selected to represent three ranges of initial

reducing sugar concentrations (low, normal, and high);

therefore, there was a large difference in the initial

concentration of reducing sugars. French fries were prepared

under standardized frying conditions (177uC for 165 s)

FIGURE 2. Histograms of loge-transformed acrylamide concen-tration. (A) Experiment before instructions; (B) experimentafter instructions.

J. Food Prot., Vol. 76, No. 3 FRYING INSTRUCTIONS AS A CONTROL MEASURE IN RESTAURANTS 469

using fryers equipped with a temperature controller and a

timer. We concluded that a decrease in the concentration

of reducing sugars can be an effective control measure;

however, other factors that affect the temperature-time

profile of frying oil need to be strictly controlled to further

reduce the variation in acrylamide production. The effect

of reducing sugars in the present study may have been

overruled by other factors, such as frying time and temperature,

in agreement with our previous observations (34). These

findings highlight the importance of validating the results of

controlled laboratory in actual food processing settings, as

previously emphasized by other authors (5, 40).The mean product/oil ratio was 15.4% for the

experiment after instructions and 14.2% for the experiment

before instructions, which indicates that restaurants did not

implement the 10% rule in daily practice. These results are

in agreement with those of our previous study (34), in which

we reported that chain fast-food restaurants, institutional

caterers, and other restaurants also did not implement the

10% rule. Statistical analysis revealed no significant

correlation between the product/oil ratio and the acrylamide

concentration (r ~ 0.033, P ~ 0.570). The product/oil ratio

also did not significantly contribute to the prediction of

acrylamide formation (beta ~ 20.084, P ~ 0.179;

Table 6). Grob et al. (16) recommended the 10% rule for

the product/oil ratio to achieve optimum culinary quality

combined with minimum acrylamide concentration. Other

authors have found that portion size affects the profile of

frying temperature over time and thus the formation of

acrylamide (9, 16). In the present study, frying temperature

was not significantly correlated with acrylamide concentra-

tion (r ~ 0.013, P ~ 0.408; Table 6). The dilution effect of

frying temperature may further weaken the effect of the

product/oil ratio on the formation of acrylamide.

The inconsistent compliance of food handlers with

instructions remains a challenge, although training was

TABLE 5. Acrylamide concentrations of French fries prepared at various restaurants before and after instructions were provided

Restaurant

Acrylamide (mg/kg fresh wt)

Before instruction After instruction

Mean (SD)a CV (%)b Medianc Minimum Maximum Mean (SD)d CV (%) Median Minimum Maximum

A 541 (0.28) 32.3 558 325 787 565 (0.22) 24.6 545 406 954

B 271 (0.38) 46.2 272 152 562 715 (0.40)* 49.2 636 364 1,346

C 506 (0.25) 28.4 513 290 749 961 (0.26)* 29.7 989 593 1,478

D 1,353 (0.69) 99.4 1,285 506 4,283 1,633 (0.53) 69.9 1,914 537 3,369

E 1,517 (0.51) 66.5 1,522 796 3,816 169 (0.49)* 63.2 160 85 372

F 134 (0.42) 52.2 132 77 329 239 (0.36)* 43.3 242 125 414

G 1,271 (0.20) 22.1 1,226 824 1,814 1,296 (0.21) 23.4 1,288 890 1,964

H 286 (0.14) 15.0 280 220 345 NDe

I 439 (0.37) 44.8 426 249 723 ND

J 181 (0.34) 40.5 187 90 311 ND

Total f 596 (0.91) 148 580 77 4,283 614 (0.85) 134 657 85 3,369

a Mean values based on n ~ 18. The mean of the loge-transformed acrylamide concentration in each restaurant was back-transformed to the

original scale using an antilogarithmic function and was expressed as a geometric mean. The standard deviation is for the loge-

transformed acrylamide concentration.b The exact CV (coefficient of variation) was calculated using the formula of 100(eSD 2 1), where SD is the standard deviation of the loge-

transformed acrylamide concentration.c The median of loge-transformed acrylamide concentration in each establishment was back-transformed to the original scale using an

antilogarithmic function and was expressed as a geometric median.d Means with asterisks are significantly different from those before instruction at the 0.05 level.e ND, no data.f Based on data from seven restaurants (A through G).

TABLE 6. Bivariate correlations and multiple linear regression analyses for the influencing factors

Influencing factor

Bivariate correlations analysis Multiple linear regression analysis

Pearson correlation coefficient (r) P Betaa P

Frying time 0.228 ,0.05 0.259 ,0.05

Frying temp 0.013 0.816 0.173 ,0.05

Reducing sugarsb 20.161 ,0.05 20.125 ,0.05

Thawing 0.142 ,0.05 0.179 ,0.05

Product/oil ratio 0.033 0.570 20.084 0.179

a Beta values are standardized coefficients.b Referring to fresh weight, n ~ 306.

470 SANNY ET AL. J. Food Prot., Vol. 76, No. 3

given to the food handlers on how to fry the French fries,

and they were made aware of the need to comply with the

instructions. Changing this behavior is difficult because the

food handlers tend to have a preference for a certain target

color, even though a color card was included in the frying

instructions to illustrate the golden-yellow target color of the

finished French fries. Food handlers appear to maintain their

normal habits and used their own target color. Our

observations are consistent with those of other authors,

who have suggested that food handlers in restaurants tend to

develop individual preparation techniques to meet their

preferred target color of French fries (17, 41). Food handlers

also may not yet have been comfortable with the frying

instructions, although they practiced frying in the training

session to create familiarity. Other authors have emphasized

the need to create familiarity because when new information

becomes available, it must first trigger attention and then

achieve comprehension; only then can it influence decision

making (3, 42).Two different strategies, i.e., emphasizing a managerial

approach and emphasizing a technological approach, may

overcome the problem of the noncompliant behavior of the

food handlers in the FSE. A tailored training program

should be developed to meet the specific training needs of

food handlers and thus have a greater effect on their

intention and actual behavior. In addition to the food

handlers, supervisors also must be trained with the frying

instructions to enable them to provide feedback and stop any

noncompliant behavior. Palmen et al. (27) emphasized that a

training program should include crucial steps such as

demonstration, practice, and feedback to be effective.

Although the training session in the present study started

with a demonstration and the food handler had an

opportunity to practice and refine the required behavior,

they did not receive active feedback when they did not stop

frying at 240 s. Various authors (27, 31, 37) have stated that

feedback from supervisors encourages safe food handling

practices among food handlers in real work situations and is

necessary to correct any noncompliant actions. Seaman and

Eves (37) found that food handlers’ perception of social

normative pressures (from on-going supervision) signifi-

cantly influenced their intention to perform food safety

actions. Future studies could investigate the effect of strict

monitoring of compliance with instructions on the concen-

tration of acrylamide in fried foods. This managerial

strategy is expected to have a strong influence on food

handlers’ control decisions concerning maintenance of

frying conditions between acceptable tolerances, which will

result in a significant reduction in the variation in

acrylamide concentration.

A technological approach would be to establish

requirements for the features of commercial fryers, e.g.,

they could be equipped with a temperature controller, a

timer, and an automatic lifting mechanism that removes

the basket from the oil after a set frying time. These

technological features support food handlers by taking over

control decisions, thus facilitating behavior change. Inade-

quate equipment (frying facility) is certainly one of the

barriers that would prevent food handlers from complying

with instructions (4, 29, 35, 36). Health authorities may

need to set rules for dedicated fryers to be used in

restaurants instead of frying pans to fry French fries, and

regular inspections may be needed to ensure that restaurants

comply with this requirement.

In the present study, frying instructions supported food

handlers’ decisions to start frying when oil reached 175uC;

however, the instructions had an inconsistent effect on the

food handlers’ decisions to end frying. The impact of frying

instructions on the mean and variation of the acrylamide

concentration in French fries was not consistent because of

the inconsistent compliance with the frying instructions with

which food handlers were trained. The majority of the food

handlers exceeded the prescribed frying time because they

fried the French fries until the color and crispiness were

consistent with their own preference. When instructions

were followed strictly, the acrylamide concentration was

significantly reduced. However, in cases of noncompliance,

the mean acrylamide concentrations were higher than those

before instructions were provided.

ACKNOWLEDGMENTS

We thank all the restaurants that were involved in sampling and data

collection. We also thank Hamezan bin Muhammad@Ahmad and Noor

Hezliza binti Muhamad Nodin (Research Laboratory of Division of

Postgraduate, Research & Innovation, Faculty of Food Science and

Technology, Universiti Putra Malaysia) for technical assistance with the

acrylamide analysis. Financial support from the International Foundation

for Science (project no. E/4710-1, ‘‘Assessing risk of acrylamide in French

fries production in Malaysian food service establishments from a techno-

managerial approach’’) also is gratefully acknowledged.

REFERENCES

1. Al-Kahtani, H. A. 1991. Survey of quality of used frying oils from

restaurants. J. Am. Oil Chem. Soc. 68:857–862.

2. Bland, J. M., and D. G. Altman. 1996. Statistics notes: transforma-

tions, means, and confidence intervals. Br. Med. J. 312:1079.

3. Breakwell, G. M. 2000. Risk communication: factors affecting

impact. Br. Med. Bull. 56:110–120.

4. Clayton, D. A., C. J. Griffith, P. Price, and A. C. Peters. 2002. Food

handlers’ beliefs and self-reported practices. Int. J. Environ. Health

Res. 12:25–39.

5. Confederation of the Food and Drink Industries of the European

Union. 2006. The CIAA acrylamide toolbox. Available at: http://

www.ciaa.be/documents/brochures/CIAA_Acrylamide_Toolbox_Oct2006.

pdf. Accessed 13 May 2008.

6. Dallal, G. E. 2009. Logarithms. Available at: http://www.jerrydallal.

com/LHSP/logs.htm. Accessed 21 November 2010.

7. DiNovi, M. 2006. The 2006 exposure assessment for acrylamide.

Available at: http://www.fda.gov/downloads/Food/FoodSafety/Food

ContaminantsAdulteration/ChemicalContaminants/Acrylamide/UCM

197239.pdf. Accessed 25 November 2007.

8. Evans, J. R., and W. M. Lindsay. 2004. The management and control

of quality. West Publishing Company, St. Paul, MN.

9. Fiselier, K., D. Bazzocco, F. Gama-Baumgartner, and K. Grob. 2006.

Influence of the frying temperature on acrylamide formation in

French fries. Eur. Food Res. Technol. 222:414–419.

10. Fiselier, K., and K. Grob. 2005. Legal limit for reducing sugars in

prefabricates targeting 50 mg/kg acrylamide in French fries. Eur.

Food Res. Technol. 220:451–458.

11. Friedman, M. 2003. Chemistry, biochemistry, and safety of

acrylamide. A review. J. Agric. Food Chem. 51:4504–4526.

12. Garner, R. 2010. The joy of stats: a short guide to introductory statistics

in the social sciences. Available at: http://books.google.com.my/

books?id~sl_mrrxEtjIC&pg~PA166&lpg~PA166&dq~ThezJoy

J. Food Prot., Vol. 76, No. 3 FRYING INSTRUCTIONS AS A CONTROL MEASURE IN RESTAURANTS 471

zofzStatszandzlogztransformation&source~bl&ots~SLSfB--r8C

&sig~b_LAkWkvqtSNetPe3p9uZF1a1bc&hl~en#v~onepage&q&

f~false. Accessed 21 November 2010.

13. Gere, A. 1985. A survey on operating conditions and quality of

commercial frying fats in Hungary. Z. Ernaehrungswiss. 24:120–132.

14. Gertz, C., and S. Klostermann. 2002. Analysis of acrylamide and

mechanisms of its formation in deep-fried products. Eur. J. Lipid Sci.Technol. 104:762–771.

15. Grob, K. 2007. Options for legal measures to reduce acrylamide

contents in the most relevant foods. Food Addit. Contam. 24:71–81.

16. Grob, K., M. Biedermann, S. Biedermann-Brem, A. Noti, D. Imhof,

T. Amrein, A. Pfefferle, and D. Bazzocco. 2003. French fries with

less than 100 mg/kg acrylamide. A collaboration between cooks and

analysts. Eur. Food Res. Technol. 217:185–194.

17. Haase, N. U. 2006. The formation of acrylamide in potato products, p. 41–

59. In K. Skog and J. Alexander (ed.), Acrylamide and other hazardous

compounds in heat-treated foods. Woodhead Publishing, Cambridge.

18. Hopkins, W. G. 2003. A new view of statistics. Available at: http://

www.sportsci.org/resource/stats/logtrans.html. Accessed 21 Novem-

ber 2010.

19. Institute for Reference Materials and Measurements. 2006. Acrylam-

ide monitoring database. Joint Research Centre, European Commis-

sion. Available at: http://irmm.jrc.ec.europa.eu/activities/acrylamide/

Pages/database.aspx. Accessed 11 July 2011.

20. International Agency for Research on Cancer. 1994. Some industrial

chemicals. IARC Monogr. Eval. Carcinog. Risk Chem. Hum. 60:435–453.

21. Luning, P. A., and W. J. Marcelis. 2006. A techno-managerial

approach in food quality management research. Trends Food Sci.

Technol. 17:378–385.

22. Luning, P. A., and W. J. Marcelis. 2007. A conceptual model of food

quality management functions based on a techno-managerial approach.

Trends Food Sci. Technol. 18:159–166.

23. Luning, P. A., and W. J. Marcelis. 2009. A food quality management

research methodology integrating technological and managerial

theories. Trends Food Sci. Technol. 20:35–44.

24. Mills, C., D. S. Mottram, and B. L. Wedzicha. 2009. Acrylamide, p.

21–50. In R. H. Stadler and D. R. Lineback (ed.), Process-induced

food toxicants: occurrence, formation, mitigation, and health risks.

John Wiley & Sons, Hoboken, NJ.

25. Morley-John, J., B. A. Swinburn, P. A. Metcalf, F. Raza, and H.

Wright. 2002. Fat content of chips, quality of frying fat and deep-

frying practices in New Zealand fast food outlets. Aust. N. Z. J. Public

Health 26:101–106.

26. Palazoglu, T. K., and V. Gokmen. 2008. Reduction of acrylamide

level in French fries by employing a temperature program during

frying. J. Agric. Food Chem. 56:6162–6166.

27. Palmen, A., R. Didden, and H. Korzilius. 2010. Effectiveness of

behavioral skills training on staff performance in a job training setting

for high-functioning adolescents with autism spectrum disorders. Res.

Autism Spectr. Disord. 4:731–740.

28. Petrie, A., and C. Sabin. 2009. Medical statistics at a glance.

Available at: http://books.google.com.my/books?id~upQ5tlFEc1sC

&pg~PA30&lpg~PA30&dq~coefficientzofzvariationzandz

transformedzdata&source~bl&ots~RM2e-0MIY0&sig~o7xU5RvY_

DZXOpkopoHXJG28QKU&hl~en#v~onepage&q&f~false. Ac-

cessed 21 November 2010.

29. Quaglia, G., J. Comendador, and E. Finotti. 1998. Optimization of

frying process in food safety. Grasas Aceites 49:275–281.

30. Rasmussen, G. O., J. W. Finkowski, R. F. Meyer, R. L. Keller, T. P.

Kempf, and R. N. Phillips. March 1993. Automated French fry

cooking apparatus. U.S. patent 5189944.

31. Rennie, D. M. 1994. Evaluation of food hygiene education. Br. Food

J. 96:20–25.

32. Romani, S., M. Bacchiocca, P. Rocculi, and M. D. Rosa. 2008. Effect

of frying time on acrylamide content and quality aspects of French

fries. Eur. Food Res. Technol. 226:555–560.

33. Sanny, M., S. Jinap, E. J. Bakker, M. A. J. S. van Boekel, and P. A.

Luning. 2012. Is lowering reducing sugars concentration in French

fries an effective measure to reduce acrylamide concentration in food

service establishments? Food Chem. 135:2012–2020.

34. Sanny, M., S. Jinap, E. J. Bakker, M. A. J. S. van Boekel, and P. A.

Luning. 2012. Possible causes of variation in acrylamide concentra-

tion in French fries prepared in food service establishments: an

observational study. Food Chem. 132:134–143.

35. Sanny, M., P. A. Luning, W. J. Marcelis, S. Jinap, and M. A. J. S. van

Boekel. 2010. Impact of control behaviour on unacceptable variation

in acrylamide in French fries. Trends Food Sci. Technol. 21:256–267.

36. Seaman, P., and A. Eves. 2006. The management of food safety—the

role of food hygiene training in the UK service sector. Int. J. Hosp.

Manag. 25:278–296.

37. Seaman, P., and A. Eves. 2008. Food hygiene training in small to

medium-sized care settings. Int. J. Environ. Health Res. 18:365–374.

38. Shahar, S. S., R. Ghazali, N. A. M. Yusoff, N. S. Safii, Z. A. Manaf,

and S. A. M. Noah. 2002. Atlas of food exchanges and portion sizes.

MDC Publishers Printers Sdn. Bhd., Kuala Lumpur.

39. Shuttleworth, M. 2009. Hawthorne effect. Available at: http://www.

experiment-resources.com/hawthorne-effect.html. Accessed 20 May

2012.

40. Stadler, R. H. 2005. Acrylamide formation in different foods and

potential strategies for reduction, p. 157–169. In M. Friedman and

D. S. Mottram (ed.), Chemistry and safety of acrylamide in food,

vol. 561. Springer, New York.

41. Vinci, R. M., F. Mestdagh, C. Van Poucke, B. Kerkaert, N. De Muer,

Q. Denon, C. Van Peteghem, and B. De Meulenaer. 2011. Imple-

mentation of acrylamide mitigation strategies on industrial production

of French fries: challenges and pitfalls. J. Agric. Food Chem. 59:898–

906.

42. Wilcock, A., M. Pun, J. Khanona, and M. Aung. 2004. Consumer

attitudes, knowledge and behaviour: a review of food safety issues.

Trends Food Sci. Technol. 15:56–66.

43. Williams, J. S. E. 2005. Influence of variety and processing

conditions on acrylamide levels in fried potato crisps. Food Chem.

90:875–881.

472 SANNY ET AL. J. Food Prot., Vol. 76, No. 3