Archived Research and Statistical Tips

Descriptive or Inferential: What Do I Want to Accomplish?
Before conducting any research or running any statistical analysis a planner must ask the essential question: what do I want to accomplish? While there are many ways to answer the question, it usually comes down to two options -- description or inference.

Descriptive research describes the total population under study. This means that the whole population can be reached and surveyed. A diocesan planner may wish, for example, to describe the opinions of a particular parish staff about a sharing a pastor with another parish. In this case the planner uses some instrument (a survey, interview, etc.) to reach each and every staff member.

Inferential research infers a characteristic of the total population from a sample of the population. Inferential research is usually used when descriptive research is impossible. The planner may wish to know the opinions of all parishioners about sharing their pastor but realizes it is impossible to reach each and every parishioner. If a survey is distributed at all Masses, for example, the opinions of parishioners who are not at the parish that weekend will not be counted. If surveys are mailed to all parishioners, some will forget to return them.

Descriptive research is considerably easier than inferential research but it is often impractical in the real world situations that planners find themselves.

Choosing Random Samples
The elementary principle about choosing a random sample is that each and every person within a target population must have an equal chance of being selected. Only when this is true can one generalize from the sample to the target population. So the first task is to clarify the target population. If a pastor want the opinions of "the people" from St. Anne's, what does a pastoral planner or researcher do? He or she needs to ask the pastor to define "the people" to which he is referring. Depending upon the situation, he might be thinking of: registered parishioners; registered parishioners who contribute financially to the parish; Mass attenders; or some other parish subgroup. 

One must be very precise here because one cannot generalize from a sample of one population to another population. One cannot take a sample of the people at Mass at a particular Sunday and generalize to the whole parish because all parishioners do not have an equal chance of being selected. Parishioners not present, for example, have a 0% chance of being selected.

Once the target population has been defined, options of reaching the population can be brainstormed. The options depend upon the depth of involvement. Getting a sample of registered parishioners to express their opinions via a simple questionnaire will be considerably easier than getting them to attend an evening meeting so that the pastoral staff can hear their opinions firsthand. 

Sometimes costs or other constraints make it impossible to reach the target population. In one diocese a questionnaire was distributed to people present at all Masses on a particular weekend instead of being mailed to all registered parishioners. The constraint was cost. By distributing the surveys at Mass the diocese saved over $200,000. The downside, of course, was that the results could not be generalized to all parishioners. The diocese was willing to trade the greater generalizability for the considerably lower cost. One could argue that it is important to reach out to all parishioners, not just those at Mass, but that would be a philosophical issue not a research issue.