Of the following techniques for collecting data, which one is generally considered the least costly to implement: experiments, telephone surveys, mail questionnaires, direct observation, personal interview?

What will be an ideal response?

In most instances, the mail questionnaire is the least costly method for data collection. Even though response rates are notoriously low, the cost of sending surveys in the mail is very low on a per item basis. However, there are many instances where mail surveys can't be used to collect the required data. For instance, if we are interested in the impact of greater drying heat in the press when plywood is manufactured, we can't very well send a survey to a piece of plywood asking for information about its strength and quality! An experiment would be required. In other instances, we need to use observation to get accurate data since people might be tempted to reply with the desired response rather than the truth. For instance, a question like, "Do you wear your seatbelt when driving in your car?" might not provide accurate data in a written survey. Observation might provide better data on driver behavior.