This may also involve numbers - the amount participating in the task/sample. To help ensure a good cross section of results/input is obtained / data collected.
As an obscure example if you wanted to find out which vegetables rabbits liked the most an unlikely group to consider if you wanted to produce a statistically significant sample would be cat or goldfish owners.
Wikipedia article - Statistical significance
Obtaining a statistically significant sample will give weight to an argument, area of reasearch - negatively or positively. Help prove/disprove issues of contention, agreement, help others to understand the significance of the data collected and any effect on policy, procedure, opinion, reputation...
Using the question in mind - or that needs to be answered / reasearched look at who you would invite to participate and why.
Mothers
Fathers
Step Parents
Siblings
Aunts, Uncles and extended family
Professionals
Age related groups
Asking unrelated groups to take part may have an impact on the results, make them less reliable or irrelevant. However including outside participants may also open up opportunities that weren't thought of originally and provide a contrast ..
Hth, an older post reply sorry but just came across the wikipedia article
xx