Selecting a Data Collection Method
As we have seen, there are many ways to collect data. For each of your
indicators, you need to choose an appropriate data collection method. So, how
do you decide which methods to use? Well, there are a few questions to consider
when you are choosing a data collection method:
What?
Who?
How often?
By whom?
Can we do it?
What?
What kind of data will you need to collect to measure your indicator? Detailed
qualitative data will be collected very differently than quantitative data. For
example, surveys, laboratory measurements, and document reviews work well
for collecting large amounts of quantitative data, while interviews and focus
groups are often the best choices for collecting detailed qualitative data.
It is also worth considering whether the data you are collecting is personal or
sensitive. A group setting, such as a focus group, might be too public if you are
interested in collecting personally sensitive information.
Who?
Who are the people who can provide the data for your indicator? To accurately
measure your indicator, do you need to collect data from a large group of people
spread across lots of different locations? Or, can you collect enough data by
reaching a few key people?
What is the best method for reaching these people? Will they be able to fill out
complex, written tools? Or would they prefer in-person conversation?
Additionally, consider the amount of work that you are asking your participants
to do. For example, would it be reasonable to ask your participants to spend two
hours in an interview?
How Often?
How often do you need to collect these data? Some methods, such as very large
surveys, require a lot of time to do well. These methods might be worth using
once a year, or only when reports are about to be written. Other methods, such
as observations, are inexpensive and quick to deploy. These can be used daily or
weekly.
A simple rule is that the lower an indicator is on your logframe, the more often it
should be measured. Output indicators should be measured often: daily or
weekly, for example. Outcome indicators do not need to be measured as often.
These might be measured quarterly or annually. In fact, since outcomes often
take a long time to appear, it might be necessary to measure your outcomes
several months or years after the activities are over. Note that this is general
guidance, not a rule that will apply to all situations.
Additionally, when you are deciding how often to collect data for an indicator,
you will want to consider the schedule of your team as well as other events like
school schedules, agricultural seasons and holidays.
By Whom?
Who will be collecting the data? Do you have access to skilled observers,
interviewers or moderators? What is the educational level of those who will be
collecting the data? Do they speak the same languages as the people that they
will be communicating with, or will you need to translate tools and guides into
different languages? Do they understand how to use any required technology?
Additionally, who will analyze the data? Some data collection methods like
surveys and focus groups generate huge amounts of data. If your project does
not have the expertise to analyze this data, the data may be wasted.
Can We Do It?
Do you have enough time, money, and expertise? Some data collection methods
are fairly cheap and simple, while other methods are expensive and complex. For
example, while laboratory measurement creates fantastic, reliable data, very few
teams have the expertise, time, or money to collect data this way.