M&E staff interprets data within its level of representation

Each sampling methodology has a certain level of representation and the data
collected should be interpreted within the boundaries of this representation.
Random sampling methods (discussed in Random Sampling), used for quantitative
data collection, allow data to represent the larger population from which the sample
was selected. Conversely, purposeful sampling methods (discussed in Purposeful
Sampling), used for qualitative data collection, collect data that cannot be
generalized to a larger population but that can be used to better understand the
specific context or situation of the participants.

  Generalizing either quantitative or qualitative data outside its level of representation
is likely to result in incorrect conclusions or assumptions.
For quantitative data, consider the population for which the sample was designed to
represent. If your analysis plan includes comparisons between subgroups, refer to
your sampling methodology to ensure that the sample was designed to include
stratification (statistical comparison of subgroups within the data). If your sample
was not designed to include stratification, any comparisons between subgroups
within the data are not considered statistically sound and can be viewed as
suggested differences only. Also consider the level of standard error used in
determining the sample size when interpreting quantitative results.
  The level of standard error determines the range in which the actual value in the
population falls. For example, when using a 7 percent standard error, a value of 48
percent (e.g., of households that report boiling their water before drinking) from the
sample data actually means that the value in the population is between 41 percent and
55 percent.
For qualitative data, interpret the data as only representing the contexts or
characteristics of the participants in each qualitative exercise. Refer to the purposeful
sampling methodology used to determine which types of comparisons the data will
allow. For example, if you collected data from males and from females (with other
characteristics staying relatively similar) then the data will allow for a gender
  Qualitative data can only represent the types of individuals, households and
communities that participated in the data collection activity. Refer to your analysis
plan, which should provide the specific perspectives or insights needed from
qualitative data.
Recognize any limitations or biases in the data collection methods when interpreting
the results. Note these limitations or possible biases in the monitoring or evaluation
  Limitations are nothing to hide! The majority of data collection exercises experience
one type of limitation or another due to logistics constraints or other factors. The best
approach is to be up front about limitations and to consider these limitations when
interpreting the data.

M&E staff interprets qualitative and quantitative results together

After you analyze qualitative and quantitative data separately, interpret the results
together. When interpreted together, qualitative and quantitative results will
complement each other and enhance your understanding of both the prevalence and
reasoning behind the practices, knowledge and attitudes of the surveyed population.

Md. Kaysar Kabir

Click Here to Leave a Comment Below 0 comments