*2.4. Data Analyses*

The data analyses were divided into three different main steps. At first, the qualitative analyses were performed, followed by the quantitative analyses, and lastly both qualitative and quantitative analyses were conducted. Analyses were made in an iterative process; moving back and forth between different data, a schematic figure of the different steps in the analysis process is presented in Figure 1.

The qualitative data consisted of memory notes of interviews, observations and telephone conversations as well as photographs. Initially, the memory notes were examined in detail on several occasions to create familiarity with the content. Subsequently, the analytical coding and categorisation were inspired by the analysis in grounded theory [49]. A direct content analysis approach was applied as a preliminary coding scheme [50]. The themes were organised in two different tracks; the first track followed various activities and factors related to food waste management at the department, and the second track followed the different FV categories and their causes of waste. The qualitative data were reviewed several times before reaching a stage of saturation, at which no further themes emerged from the processes [51].

The quantitative data were processed and analysed using Excel spreadsheets. In the data processing, the same type of fruit or vegetable constituted one category; for example, all different varieties of apples became an apple category. The same procedure was performed with all types of fruits and vegetables. For each category, data about waste weight and waste quota were calculated. The waste quota was defined as the waste in store in relation to the sold quantity and was calculated with the equation: *Waste quota = Waste weight/(Waste weight + Sold weight)*. Before detailed analysis of the waste data was conducted, the information was checked manually, and any errors and inconsistencies were followed up. Whenever possible, data were corrected and in other cases, the inaccurate data were excluded from further analysis. The waste weight of the excluded products corresponded to 0.01% of the total waste weight and none of the FV categories were overrepresented. When calculations were completed for all FV categories, the top categories that together corresponded to 80% of the total amount of waste were identified. For each FV category at the top list, special attention was paid to the differences of waste between packaged/unpackaged and organic/conventional products. If the share of packaged or organic products was less than 5% for any of the FV categories, no figures were reported since the proportion was considered to be too small to warrant an analysis. followed up. Whenever possible, data were corrected and in other cases, the inaccurate data were excluded from further analysis. The waste weight of the excluded products corresponded to 0.01% of the total waste weight and none of the FV categories were overrepresented. When calculations were completed for all FV categories, the top categories that together corresponded to 80% of the total amount of waste were identified. For each FV category at the top list, special attention was paid to the differences of waste between packaged/unpackaged and organic/conventional products. If the share of packaged or organic products was less than 5% for any of the FV categories, no figures were reported since the proportion was considered to be too small to warrant an analysis. In the final phase, both qualitative and quantitative results were analysed in order to identify product-specific causes of waste. For every FV category on the top list, data pertaining to food waste were examined in relation to qualitative data about the specific product. The process was iterative and new themes related to general and product-specific causes of waste emerged until saturation was reached.

in detail on several occasions to create familiarity with the content. Subsequently, the analytical coding and categorisation were inspired by the analysis in grounded theory [49]. A direct content analysis approach was applied as a preliminary coding scheme [50]. The themes were organised in two different tracks; the first track followed various activities and factors related to food waste management at the department, and the second track followed the different FV categories and their causes of waste. The qualitative data were reviewed several times before reaching a stage of saturation, at which no further themes

The quantitative data were processed and analysed using Excel spreadsheets. In the data processing, the same type of fruit or vegetable constituted one category; for example, all different varieties of apples became an apple category. The same procedure was performed with all types of fruits and vegetables. For each category, data about waste weight and waste quota were calculated. The waste quota was defined as the waste in store in relation to the sold quantity and was calculated with the equation: *Waste quota = Waste weight/(Waste weight + Sold weight)*. Before detailed analysis of the waste data was conducted, the information was checked manually, and any errors and inconsistencies were

*Sustainability* **2022**, *14*, x FOR PEER REVIEW 6 of 20

emerged from the processes [51].

*2.5. Delimitations*  Only fresh fruits and vegetables were included in the study. Preserved, dried or frozen products were excluded. No distinction was made between edible or non-edible food waste since the products, including the peel and haulm, were sold as whole products and the weight of the whole products was reported as waste by the stores. Only in-store waste In the final phase, both qualitative and quantitative results were analysed in order to identify product-specific causes of waste. For every FV category on the top list, data pertaining to food waste were examined in relation to qualitative data about the specific product. The process was iterative and new themes related to general and product-specific causes of waste emerged until saturation was reached.
