Extending the KAP framework: Food safety knowledge, attitudes, practices, and pathogen awareness among hotel service personnel

2.1. KAP model

The KAP model posits that knowledge influences an individual’s attitude in a positive way, and that attitude, in turn, shapes behavioural practices (Kwol et al., 2020). Frequently preferred by researchers due to its simple design, capacity to generate measurable data, and clear interpretability of results (Launiala, 2009), this model has also become a widely used framework for examining food safety issues at various stages of the food production and value chain (Elsahoryi et al., 2024). In this regard, Baş et al. (2006) demonstrated that employees in Turkish food businesses have significant gaps in their knowledge of basic food hygiene and that training these employees is an urgent necessity. Ahmed et al. (2021), on the other hand, reported that food safety knowledge has a significant effect on employees’ attitudes and practices, and that this relationship varies between moderate and strong levels. However, the translation of knowledge into behaviour does not always occur directly. Ko (2013) found that knowledge alone is insufficient to change either attitudes or food processing practices; however, a significant correlation exists between knowledge, attitudes, and hazard analysis and critical control point (HACCP) practices, with attitude acting as a mediator in this relationship. Meher et al. (2022) also found that an acceptable level of knowledge did not translate into good hygiene practices in food processing and sales processes; Nawawi et al. (2022), meanwhile, emphasised the decisive role of training in improving practices and stated that the KAP assessment serves as a fundamental starting point for identifying areas requiring intervention. Food handlers are identified as the most critical factor in the prevention of foodborne illnesses in accommodation establishments. Baş et al. (2006) noted that staff can become carriers of pathogens due to poor personal hygiene and cross-contamination, and that inappropriate staff behaviour contributes to approximately 97% of foodborne illnesses. Ahmed et al. (2021) also emphasised that food handlers can pose a serious public health risk if appropriate practices are not adopted. In addition to general food safety knowledge, awareness regarding pathogenic microorganisms constitutes an important dimension of food safety knowledge among food handlers, since insufficient microbiological awareness may increase contamination risks and unsafe food handling practices. In line with this literature, the present study aims to assess the food safety knowledge, attitudes, and practices of hotel service staff through the KAP framework. The knowledge and hygiene practices of staff directly involved in processes such as food preparation, service, storage, and the prevention of crosscontamination are of critical importance to public health. In addition to examining classical KAP relationships, the present study also aims to determine whether awareness regarding pathogenic microorganisms influences food safety practices among hotel service personnel. Based on the literature, the following hypotheses were developed:

H₁: Food safety knowledge positively affects food safety attitudes.

H₂: Food safety knowledge positively affects food safety practices.

H₃: Food safety attitudes positively affect food safety practices.

H₄: Pathogen microorganism knowledge positively affects food safety practices.

H₅: Food safety attitudes mediate the relationship between food safety knowledge and food safety practices.

3.1. Research design

The purposes of this study were to determine the level of food safety knowledge of service personnel working in hotel businesses and to reveal the importance of food safety for hotel businesses. For these purposes, a quantitative approach was used in this study. Quantitative research is a research approach that aims to answer research questions through the collection of numerical data and statistical analysis. This approach was also used to test the accuracy of hypotheses, examine the relationships between variables, and reach generalisations (Creswell, 2013).

3.2. Study area: Ankara, Turkey

Ankara, located in the northwest of Central Anatolia, was designated the capital city of the Republic of Turkey in a law enacted by Mustafa Kemal Ataturk on October 13, 1923 (Ankara Provincial Ministry of Culture and Tourism, 2024a). It is the second-most populous city of Turkey, with a population of 5,803,482 (TUIK, 2024). It has become a cultural, congress, and health tourism destination for domestic and foreign tourists (Ministry of Culture and Tourism, 2024). It hosted 2,980,263 domestic tourists and 511,486 foreign tourists in 2023 (Ankara Provincial Ministry of Culture and Tourism, 2024b).

3.3. Sample and participants

The research population consisted of service personnel working in 4- and 5-star hotel businesses operating in Ankara. The reason for choosing 4- and 5-star hotel businesses was that these businesses serve food to their customers in accordance with International Customer Service Standards. This issue was considered important in terms of obtaining reliable data and increasing the generalisability of the research findings. According to the data of the Ankara Provincial Directorate of Culture and Tourism, there were 91 hotel establishments with 4 or 5 stars in Ankara at the time of the study (Ankara Provincial Directorate of Culture and Tourism, 2024b). During the data collection process, 57 hotel establishments were contacted and invited to participate in the study. Survey could only be conducted in 18 hotel establishments, as some properties either declined to participate or did not respond. Therefore, the hotel participation rate was 31.6%. Convenience sampling was employed due to accessibility constraints and voluntary participation in the study. A total of 267 surveys were collected from service personnel working in the participating hotels. Prior to the analyses, the data were screened for missing values, inconsistent responses, and improperly completed surveys. Following the data cleaning process, 18 surveys were excluded, and the analyses were conducted with 249 valid surveys. Hotel managers were interviewed by having an appointment and/or obtaining permission to interview each participant. The confidentiality of the surveyed hotels and interviewees was ensured. Ethical approval for this study was obtained from the Scientific Ethics and Evaluation Committee of Selçuk University Faculty of Tourism (Decision No: 105, Date: 04/05/2022). The study was approved as appropriate for the use of survey instruments in research. All participants were informed about the purpose of the study and provided voluntary informed consent prior to participation.

3.4. Instrument development and data collection

The survey instrument used in this study was adapted and refined based on the food safety scale developed by Manes et al. (2014) Expert opinions from academics specialising in food safety and hospitality management were incorporated to improve the clarity, content relevance, and contextual suitability of the survey for hotel service personnel. In addition, cognitive interviews were conducted with Turkish-speaking service employees to evaluate item comprehensibility and wording appropriateness. Following the completion of the expert review and cognitive pre-test phases, a pilot study was conducted to assess the comprehensibility and feasibility of the survey items. The final survey was structured under 3 main dimensions based on the KAP framework: food safety knowledge (10 questions), food safety attitude (5 questions), and food safety practices (33 questions). In addition, 5 questions on knowledge of pathogenic microorganisms were included in the survey. Since the questions regarding pathogenic microorganisms were categorical, a pathogenic microorganism knowledge score was calculated. Furthermore, negative statements were included in the survey to increase consistency and limit response bias. Negative statements were reverse coded during the analysis process.

3.5. Validity and reliability analysis

Exploratory Factor Analysis (EFA), reliability analysis, and normality tests were conducted to assess the validity and reliability of the measurement model. Kaiser-Meyer-Olkin (KMO) values were calculated as .791 for the knowledge dimension, .810 for attitude, and .888 for application; all three values were well above the recommended threshold of .60 for factor analysis. This result indicates that the sample is adequate for factor analysis. Looking at the factor loadings, all items exceeded the acceptable lower limit of 0.40. The loadings range from 0.611 to 0.973 for the knowledge dimension, 0.611 to 0.850 for the attitude dimension, and 0.611 to 0.871 for the application dimension. This distribution provides strong evidence that the items adequately represent the relevant latent structures. Internal consistency was assessed using Cronbach’s alpha coefficients. The coefficients obtained—0.818 for the knowledge dimension, 0.824 for the attitude dimension, and 0.868 for the application dimension—all exceed the 0.70 threshold; this demonstrates that the scales are reliable at a satisfactory level. Assumptions of normality were examined using skewness and kurtosis values. When using the ±1.5 criterion, it was observed that all dimensions exhibited acceptable characteristics of a normal distribution. Skewness values ranged from –0.036 to 0.196, whilst kurtosis values ranged from 0.764 to -1.017. Considering these findings, it was accepted that the data followed a normal distribution, and parametric analyses were conducted (Table 1).

3.6. Statistical analysis

The data were analysed using IBM SPSS Statistics 26.0. During this process, descriptive statistics, frequency analyses, Pearson’s correlation analysis, one-way ANOVA, regression analysis, and mediation analysis were applied. One-way ANOVA was chosen to identify differences between the educational groups; eta-squared (η²) values were used to determine effect sizes. Eta-squared values of .01, .06, and .14 were interpreted as small, medium, and large effect sizes, respectively. Mediation analysis was conducted using PROCESS Macro v4.2 for SPSS (Model 4) developed by Hayes (2022). Bootstrap confidence intervals were estimated using 5,000 bootstrap samples with a 95% confidence interval.

4.1. Demographic characteristics of participants

Table 2 presents the demographic characteristics of the participants. Most participants were between 20 and 40 years of age (66.7%), followed by those aged 40 and above (17.7%) and those under 20 years of age (15.7%). In terms of gender, male participants constituted most of the sample (58.6%), while female participants accounted for 41.4%. Regarding marital status, more than half of the participants were single (58.6%), followed by married (32.5%) and divorced (8.8%) individuals. In terms of educational background, the largest proportion held a high school degree (44.6%), followed by secondary school graduates (17.7%), associate degree holders (17.3%), bachelor’s degree holders (15.3%), primary school graduates (3.6%), and graduate degree holders (1.6%). With respect to occupational role, most participants were employed as waiters (58.2%), while 28.1% were classified under other positions; captain waiters and head waiters accounted for 7.6% and 6.0% of the sample, respectively. Concerning professional experience, the largest proportion reported between 1 and 5 years of experience (34.5%), followed by those with 5 to 10 years (33.7%), more than 10 years (17.3%), and less than 1 year (14.5%). Regarding food safety education, 43.8% of the participants reported having received no food safety training, while 41.4% indicated that they had. Additionally, only 34.9% of the participants reported holding a food safety certificate, whereas 65.1% did not. Finally, 47.8% of the participants reported having received service training, while 38.2% indicated that they had not.

4.2. Relationships and differences among food safety knowledge, attitude, and practice dimensions

Table 3 displays the correlation analysis results among the knowledge, attitude, and practice dimensions. Statistically significant positive relationships were identified among all dimensions (p < .001). A strong positive relationship was observed between knowledge and attitude (r = .761), indicating that participants with higher food safety knowledge tended to exhibit more positive attitudes toward food safety. Knowledge was similarly strongly associated with practice (r = .671), suggesting that higher knowledge levels were related to better food safety practices. A moderate positive relationship was additionally identified between attitude and practice (r = .473). These findings indicate that food safety knowledge and attitudes play an important role in shaping food safety practices among hotel service personnel.

Table 4 reports the ANOVA results examining differences in knowledge, attitude, and practice dimensions according to food safety training status. Statistically significant differences were identified among the groups across all dimensions (p < .05). For the knowledge dimension, participants who had received food safety training demonstrated significantly higher knowledge scores than both those who had received partial training and those who had received none. Participants who had received partial training also demonstrated significantly higher knowledge levels than untrained participants. The effect size for the knowledge dimension was large (η² = .45). Regarding the attitude dimension, statistically significant differences were observed between participants who had received training and those who had not, as well as between partially trained and untrained participants; however, no significant difference was found between the trained and partially trained groups. The effect size was similarly large (η² = .29). Significant differences were also observed in the practice dimension, with trained participants exhibiting significantly higher practice scores than the other groups, although no statistically significant difference was found between the partially trained and untrained groups. The effect size for the practice dimension was likewise large (η² = .16).

4.3. Knowledge of pathogenic microorganisms

Table 5 shows the findings regarding participants’ knowledge of pathogenic microorganisms and microbiological food safety issues. Overall, the results indicated relatively low levels of pathogen-related knowledge among hotel service personnel. The highest correct response rate was observed for the item examining how bacteria causing food poisoning can enter the kitchen environment (43.8%), while the lowest was recorded for the item concerning bacterial behaviour at body temperature (16.5%). Similarly, only 16.9% of the participants correctly identified Staphylococcus aureus as the bacterium commonly found around the nose and mouth, and only 22.9% correctly identified Salmonella and Campylobacter as the most common bacteria associated with poultry products. Furthermore, only 28.9% correctly recognised all listed microorganisms as foodborne pathogens. These findings indicate substantial deficiencies in participants’ microbiological food safety knowledge, particularly with respect to pathogenic microorganisms and contamination mechanisms.

Table 6 illustrates the differences in pathogen microorganism knowledge levels according to food safety training status. A statistically significant difference was identified among the groups (F = 97.810, p = .000). Participants who had received food safety training demonstrated the highest pathogen knowledge scores ($\overline{\text{X}}$ = 2.43), followed by those who had received partial training ($\overline{\text{X}}$ = 1.08), while participants who had received no training recorded the lowest scores ($\overline{\text{X}}$ = 0.28). Post-hoc analysis indicated statistically significant differences between all groups. The eta-squared value (η² = .443) reflected a very large effect size, indicating that food safety training accounted for approximately 44.3% of the variance in pathogen microorganism knowledge levels.

4.4. Mediation analysis results

Table 7 presents the mediation analysis examining the relationships among food safety knowledge, attitudes, and food safety practices. Food safety knowledge significantly and positively affected attitudes (β = .761, p = .000), supporting H1. Food safety knowledge also exerted a significant positive effect on food safety practices (β = .741, p = .000), supporting H2. These findings indicate that employees with higher levels of food safety knowledge tend to demonstrate better food safety practices. However, attitudes did not significantly predict food safety practices (β = –.091, p = .210), and H3 was not supported. Similarly, the indirect effect of food safety knowledge on food safety practices through attitudes was not statistically significant, as the bootstrap confidence interval included zero (BootLLCI = –.098; BootULCI = .021). Accordingly, attitudes did not mediate the relationship between food safety knowledge and food safety practices, and H5 was not supported. Pathogen microorganism knowledge, by contrast, significantly and positively predicted food safety practices (β = .348, p = .000), supporting H4. This finding suggests that employees with greater awareness of pathogenic microorganisms are more likely to demonstrate appropriate food safety practices.

6.1. Theoretical implications

This study extends the classical KAP framework by incorporating knowledge of pathogenic microorganisms as a separate explanatory dimension, thereby making an original contribution to the food safety literature. Traditional KAP studies are largely based on a linear relationship between knowledge, attitudes, and practices. However, current findings indicate that microbiological awareness regarding pathogens can directly influence hygienic behaviour by bypassing attitudinal mechanisms. Another significant theoretical contribution of the study stems from the finding that attitudes do not play a mediating role between knowledge and practices. This result lends support to the growing criticism in recent years that the classical KAP model oversimplifies the complexity underlying food safety behaviour. The findings reveal that food safety behaviour cannot be explained solely by cognitive and attitudinal processes; operational, organisational, and contextual factors also carry decisive weight in determining employees’ hygienic practices. Finally, this study is also of particular value in that it contributes to a rather limited body of literature focusing on hotel service staff. The vast majority of previous studies have treated kitchen staff or food handlers as a general category. This study, however, aims to fill this gap in the literature by specifically examining hotel service staff who come into direct contact with food during the service process.

6.2. Practical implications

The findings of this study offer a number of concrete implications for hotel managers and food safety specialists. First and foremost, the significant impact of food safety knowledge and pathogen awareness on hygiene practices highlights the indispensability of continuous, specialist-led training programmes in hotel operations. These programmes should not merely convey general hygiene rules; they must also cover microbiological content such as pathogenic microorganisms, bacterial growth conditions, cross-contamination mechanisms, and critical temperature control.

Secondly, the limited impact of attitudes on practices demonstrates that a positive attitude alone is insufficient to guarantee safe food handling behaviour. Therefore, hotel management must establish an infrastructure that supports staff practices through operational monitoring systems, regular inspections, corporate food safety policies, and practical hygiene control mechanisms.

Thirdly, the low pathogen knowledge scores identified in the study clearly highlight the need for more practical and scientifically grounded microbiological training programmes for hotel service staff. Tools such as visual hygiene reminders, infection prevention guidelines, interactive training modules, and temperature control charts have the potential to enhance staff awareness of food safety and behavioural compliance.

6.3. Limitations and future research

Despite all these contributions, certain limitations of the study should be borne in mind. Firstly, the research was conducted exclusively with staff working in 4- and 5-star hotels in Ankara. This makes it difficult to generalise the findings regarding different tourist destinations, hotel categories, or the catering sector. Secondly, the study is limited to self-reported survey data. This method carries the risk of social desirability and response bias; it suggests that participants may have reported more positive attitudes and practices than their actual behaviour under operational conditions. Third, the cross-sectional design of the study limits the ability to establish causal relationships between variables over time. Future studies may employ longitudinal or experimental research designs to better examine the stability of food safety behaviours and training effects. Future research may also expand the KAP framework by incorporating additional variables such as food safety culture, organisational support, managerial supervision, work pressure, risk perception, and behavioural control. Furthermore, observational studies examining actual hygiene behaviours in operational environments may provide more objective insights into the relationship between food safety knowledge and practices.