Data Collection From Farmers: Few Housekeeping Things

I would like to share a positive experience we had with primary data collection from farmers. We collected 47 responses from farmers in a single day during the Delaware Agriculture Week this January. The survey, conducted on computers, took approximately 35-45 minutes for individual farmers. We had set up 12 computers for the event. The data collection event was announced beforehand, and participants were encouraged to sign in. However, only a couple of farmers signed in early, and most walked in during the event.

Collecting primary data from farmers is always challenging, especially in a country like the USA. Farmers are often inundated with numerous surveys and information requests, and they have limited free time. Most of their time is occupied with operating their tractors or harvesters or planning their farming activities.

How can we make the most of the limited time windows farmers have between their farm operations for data collection? First and foremost, it's advisable not to approach farmers with surveys too frequently. Whenever possible, collaborate with others conducting surveys. This is especially crucial if the farmers are specialized, and sending them with constant surveys is not a good choice.

When we collect data from an event like farm shows or Agriculture days, early planning can result in a successful data collection experience. When approaching farmers, emphasize the best management practices employed to ensure the confidentiality of their information. Be prepared to address the questions farmers may have about the purpose of the data collection. There are always benefits from strategies such as building rapport, pre-committing, and involving a person known and trusted by the farmers.

In addition to these factors, based on my recent experiences, a few housekeeping practices contribute to successful data collection meetings:

1. Opt for presentation-style meetings over demonstrations, as many people prefer participating in field or equipment demonstrations.
2. Target events that have longer individual sessions if you plan to collect during that session. If there are back-to-back half-hour sessions, participants may be unwilling to miss any consecutive sessions, reducing their likelihood of engaging in the survey or experiment.
3. Ensure that the experiment venue is close to the main event venue. This proximity encourages greater participation.    

About the Blogger

Dr. Badri Khanal is an Agricultural Economist. Currently, he is working as a post-doctoral researcher at the Centre for Experimental and Applied Economics, Department of Applied Economics and Statistics, University of Delaware. His work is primarily driven by issues related to the sustainability of the Agri-environment. He collects primary data from surveys, interviews, and economic experiments. He examines human behavioral responses to incentives and information on the economic decisions of farmers, consumers, and the public. 

Factor Analysis of Multi collinear variables in Economics

Factor Analysis is a statistical procedure used to classify the functional form of relationships between several variables and narrow it down to a smaller set of variables. If there are many variables, and they are linearly correlated, factor analysis tries to club highly correlated variables together as a latent variable called factors. The number of factors defining all variables in most cases is less in numbers than the number of variables. Factor analysis is generally useful for linearly combining two or more variables in economic analysis.

Here is an example of factor analysis in R program:

Step I: Prepare Data

Data frame name = my_data_final

Step II: Run Factor Analysis

The analysis above showed when 13 variables were tested for five factors with varimax rotation and 0.1 cutoffs. The cutoff is taking of 0.32 to while printing to define the loadings. Factor analysis is an exploratory process in nature. If adding one more factor adds more to the cumulative variance of the model, it should keep adding. Analysis of eigenvalue also gives a basis for how many factors to use. The general principle factors having more than one eigenvalues are kept.

Step III:  Exploratory Process and Further Processing:

Further information on correlation matrix and loadings is obtained as follows. 

Use print instead of head for detail result.

Step IV: Plotting

Step V: Determining Number of Factors to Use

For this, factor having eigenvalue > 1 is considered as factor to use. Here we use 5 factors based on the eigen value from Scree Plot.

Conservation Reserve Program (CRP) Area Enrolled by County (1986, 2007, 2019)

 Conservation Reserve Program (CRP) aims to retire environmentally sensitive conventional farmland that adversely affects ecosystem services like water, soil, and biodiversity. The map shows the county-level land enrolled under the CRP for three strategic years. The first map is for 1986, the first year after the federal government started the CRP program from the Farm Bill 1985. The second map is for the year 2007, which is strategically crucial because the price of agricultural products, primarily corn, increased significantly after this because of excessive demand for bio-ethanol production. Ultimately negatively affecting the land enrollment in CRP. The third map is of 2019, the latest data available for the county level. The Farm Bill of 2018 promised to increase CRP enrollment, the effect of which has still to be analyzed.