What is data collection?

3 October 2024

Lesson Progress

0% Complete

Data are facts and statistics collected together for reference or analysis.
Data collection is a systematic process of gathering observations or measurements.
Data can be quantitative or qualitative:

Quantitative data includes anything that can be counted, measured, or given a numerical value.

Qualitative data is data that is not easily reduced to numbers. Qualitative data tends to answer questions about the ‘what’, ‘how’ and ‘why’, instead of the ‘how many’ and ‘how much’.

What methods are used to collect data?

Primary and secondary data collection methods include:

Interviews
Questionnaires and surveys
Observations
Documents and records
Focus groups
Oral histories

What techniques are used for sampling?

Sampling is the process of identifying a subset of a population that provides an accurate reflection on the whole population.

Sampling methods

There are two main methods of sampling:

Probability sampling
Non-probability sampling

We will focus on probability sampling techniques in this lesson.

Probability sampling techniques include:

Simple random sampling

Simple random sampling Involves picking respondents with no design or order Pros: Eliminates bias Cons: Requires some planning Example: Using a random number generator to generate 100 households out of a total of 1000 households.

Systematic sampling

Systematic sampling Follows a set of rules to create regularity in sampling Pros: Easier than random sampling and retains some benefits of randomness Cons: Data may skew in one way or the other and is prone to bias Example: Interviewing every tenth customer. The bias may be that every 10th customer is female.

Stratified sampling

Stratified sampling Involves dividing population into subgroups that share similar characteristics Pros: Reduces bias Cons: Characteristics can sometimes be difficult to determine, which can invite bias Example: Breaking down a population by age or by gender, income groups etc.

Cluster sampling

Cluster sampling Using subgroups of populations rather than individuals, with predefined clusters Two types: Single stage cluster (all individuals in the cluster are included) and two-stage cluster (only random individuals in the cluster are chosen) Pros: Some work has already been done as a group has been defined Cons: Potential for bias if clusters do not accurately represent the whole population Example: Choosing a municipality to represent the national population.

How big should the sample size be for quantitative data?

How long does a sample survey take?

Example: 10 data collectors can take 8 days to survey 400 households in Bobo-Dioulasso.

Assumptions:

1 hour spent in a household (HH) per enumerator.
5 HHs surveyed per day per enumerator.

In addition, remember to factor in about 2 weeks for back-checks and data validation.

CoM SSA Sustainable Energy Access and Climate Action Plan (SEACAP) course

Participants 1632