Interviews/Surveys - Why and How to Measure Food Loss and Waste

Interviews and surveys (hereafter surveys) can be a cost-effective way to develop rough quantitative estimates of FLW and to gather information about its causes. Surveys can also help collect information from a wide array of individuals or entities on attitudes toward food waste.

Surveys can be grouped into two categories: those used to collate existing data and those used to generate new FLW estimates.

A summary of the strengths and limitations of the two different types of surveys is shown in Tables A10 and A11.

Table A10. Factors to Consider when Using a Survey to Collate Existing Data

Strengths

Limitations / Points to Consider

Cost-effective method of collating information Can standardize the information requested from each interviewee

Relies on third parties Can be challenging to extract the exact type of information needed, and can be difficult to ensure that collated information has the same definition and scope of FLW Questionnaire may need to be flexible to accommodate different levels of information (e.g., granularity of data) Can be limited by commercial sensitivities and confidentiality Unlikely to include information on root causes (i.e., the reasons why food is thrown away)

Source: Authors.

Table A11. Factors to Consider when Using a Survey to Generate New Data

Strengths	Limitations / Points to Consider
Relatively cost-effective to administer Can provide data by food group or preparation stage Can provide information by demographic group and/or other characteristics Can provide data on root causes of waste and help identify hotspots	Respondents tend to underestimate the amount of food waste due to aspirational biases Not yet known how this underestimation varies over time, between groups and during intervention studies

Source: Authors.

How to Conduct a Survey to Quantify FLW

This section describes seven steps to conduct a survey to gather information about FLW.

Step 1: Set hypotheses and determine the survey approach

Before starting a survey, have a hypothesis in mind for the results you expect from the survey. This hypothesis will help focus the research and establish goals. An example of a hypothesis is: “We expect that corn farmers will report that 30 percent of their crop is left in the field during harvest.” This simple hypothesis identifies the type of crop (corn), the intended respondent (farmers) and what is being measured (crop left in field during harvest).

Next, determine which type of survey to use. If the respondents are likely to have already collected data of their own, you can use a survey focused on collating existing data. If the survey asks respondents to contribute or quantify new FLW data, a survey focused on quantifying is needed.

Step 2: Identify the method by which the survey will be administered

Surveys can be administered by mail, by telephone, electronically or in-person. Each method has advantages and disadvantages, as seen in Table A12.

Table A12. Advantages and Disadvantages of Methods for Conducting Surveys

Method	Advantages	Disadvantages
By mail	Relatively low cost Allows for both visual and written prompts	Impractical if mail service is limited Low response rate
Telephone	Interviewer can administer survey directly and explain any unclear questions Reduces travel costs as compared to in-person method	No visuals can be shared Limits respondents to those with telephone access Can be difficult to schedule
Electronic	Low cost Wide reach	Limits respondents to those with technological capability
In-person	Interviewer can administer survey directly and explain any unclear questions	More costly in terms of time and expense Interviewer can unconsciously bias responses Can be difficult to schedule

Source: Authors.

Step 3: Identify respondent audience

In some cases, the participant audience for a survey-based study will be a discreet group. For surveys with a large number of target respondents, a random sample may need to be developed. If so, a professional statistician should be consulted, although simple random sampling can be conducted if a list of the members of a population is available and complete (Laerd 2012).

Step 4: Prepare questions to quantify FLW

The next step is to develop the questionnaire to be distributed for the survey.

Some common topics for questions in an FLW quantification survey are (CEC 2017):

• estimates of FLW generated;
• reasons or causes for FLW;
• how FLW is managed; and
• current strategies or suggestions on how to prevent or reduce FLW.

You might also want to collect income or livelihood data on the respondents to cross reference some of the answers.

Questions should be sequenced in a logical progression, with simpler or more important questions at the beginning, since respondents frequently fail to complete the entire surveys (Statpac 2017). If a survey is too long it may be off-putting to respondents, so each question should be evaluated for its importance to the study.

A further discussion of the benefits and drawbacks of a number of types of questions can be found in section 7.2 of the “Guidance on Surveys” developed by the FLW Protocol.

Step 5: Test the survey and revise

If possible, test the survey with a subset of the target audience to provide insight into questions that may be confusing or unclear for the respondent. The survey can then be revised to address these concerns.

Step 6: Administer the survey

Once the survey has been designed and tested, it can be distributed to the intended audience of respondents. A complete list of the survey recipients should be kept along with those who have responded in order to track response rates.

Step 7: Prepare and analyze the data

After responses are received, they must be standardized and collated. The simplest method for doing this is to enter the data into an electronic spreadsheet.

Points to highlight in a summary of an FLW survey are:

• Frequency and amount of FLW;
• Reasons for different types of FLW;
• Relationship between FLW and variables (such as income and location); and
• Strategies used and suggestions to address or reduce FLW.

Common Data Challenges in Conducting a Survey

Low response rates. Because surveys require respondents to take time from their schedules to complete, many suffer from low response rates. For example, a survey from Food and Consumer Products of Canada in 2015 to collect FLW data from companies had just a 35 percent response rate (Food and Consumer Products of Canada 2015). Although it can be difficult to boost response rates, a common strategy is to provide respondents with a benefit for participating, such as compensation (usually quite small) or a promise of sharing the survey results (Statpac 2017).

Concerns over confidentiality. Companies are understandably reluctant to share information that could affect their competitive advantage. This can be addressed by reporting information from an entire sector rather than identifying data from individual companies. This requires the company to trust the entity conducting the survey to keep the information confidential.

Underreporting. Respondents may underreport FLW because they don’t want to appear wasteful or because they lack awareness around FLW. To counteract these biases, clear instructions should be given on the importance of accurate responses and that the survey administrators are not seeking to “shame” participants over their FLW. Survey results can also be cross-referenced with the findings of other quantification methods (such as a waste composition analysis) to determine the extent of underreporting.

Additional Resources on Conducting a Survey

FLW Protocol. 2016. Chapter 7. “Guidance on surveys,” in Guidance on FLW quantification methods. < http://flwprotocol.org/wp-content/uploads/2016/06/FLW_Guidance_Chapter7_Surveys.pdf >.

David S. Walonick. 2012. “Steps in designing a survey.” StatPac. < www.statpac.com/survey-design-guidelines.htm >.