A small sociological investigation
Turn an interest into a project
You now have enough vocabulary to ask a sociological question and enough methodological knowledge to recognize an impossible proposal. The final task is to design a small investigation whose evidence could bear on its explanation. You do not need to discover a universal law or collect information from strangers. You need to make a question, a comparison, and a limit fit together.
We will work with a new packet from the fictional San Francisco learning center. It concerns registration across several courses during four Saturdays in an invented term. These figures are separate from the twelve-seat allocation and old-versus-new examples used earlier. The packet supplies everything required for this exercise. Its records describe an imagined process, not actual residents or an evaluation of a real center.
Start with an ordinary concern: people have difficulty returning with missing documents before the deadline. That concern matters to the people involved, but it is not yet a research question. We must decide whether to study their experience, the recorded outcome, the rule's history, or the effect of changing it. Each would require different evidence. Our packet can support a bounded comparison of rules and recorded outcomes.
The supplied document packet
Document A is the registration notice used on the first two Saturdays. The counter opens at nine in the morning and closes at four in the afternoon. Applicants whose required documents are incomplete at their first contact may return by noon to complete registration that day. Returning later means beginning a new attempt on another day. The notice does not explain why noon was chosen.
Document B replaces that notice for the final two Saturdays. Opening hours remain nine to four. Applicants with missing documents may now return until four for same-day completion. The document requirements remain the same. The notice says the change is intended to reduce unsuccessful repeat visits. That sentence records a purpose; the outcome still needs examination.
Document C is a supplied aggregate of eighty distinct first registration attempts across the four Saturdays. Each appears once. Later contacts on the same day are attached to that attempt rather than counted as new applicants. Document position is recorded at first contact, before assistance. Completion means that registration is accepted by four that day. The table covers the recorded attempts in this packet, not everyone interested in taking a course.
| Position at first contact | First two Saturdays: completed / attempts | Final two Saturdays: completed / attempts |
|---|---|---|
| All required documents ready | 16 / 20 | 17 / 20 |
| Required documents missing | 4 / 20 | 10 / 20 |
| Total | 20 / 40 | 27 / 40 |
Document D adds timing information. All four completions among the initially missing-document group in the first period occurred by noon. Of the ten in the final period, three occurred by noon and seven after noon but by four. The packet does not say when people who failed to complete returned, whether they returned at all, or what prevented them from doing so.
Document E is a staffing note stating that two employees were scheduled at the counter on each of the four Saturdays. It gives no information about experience, language skills, interruptions, or time spent assisting individual visitors. No applicant work schedules, travel times, course preferences, or records of people deterred before arrival are supplied.
Choose a question the packet can address
A broad first question might be: Does the center exclude people with inflexible jobs? The concern is sociologically meaningful, but this packet cannot answer it. It contains neither work schedules nor people who never arrive. Describing missing documents as a proxy for inflexible employment would silently substitute one concept for another without evidence that the substitution is justified.
A more tractable question is: How did recorded same-day completion differ before and after the return deadline was extended, particularly among applicants missing documents at first contact? This identifies the outcome, the comparison, the relevant group, and the timing of classification. It remains descriptive. We can then ask whether the timing pattern is consistent with a deadline constraint, while retaining alternative explanations.
The mechanism we want to examine is straightforward: a later deadline permits an applicant to return during an interval previously unavailable for same-day completion. The packet contains evidence directly relevant to that interval. It does not contain the counterfactual behavior of those same applicants under the earlier rule. Some might have returned earlier if required; others might not have been able to complete at all.
This distinction improves the project. It permits a useful conclusion about what happened in the supplied records and a qualified interpretation of why, without promising an effect estimate the design cannot deliver. A small study is strongest when its ambition matches its evidence.
Define the concepts before calculating
Write the unit as a first registration attempt. It is not a visit, because an attempt can involve a return visit on the same day. It is not a seat in a single course, because the packet concerns several courses. It is not necessarily a unique person across every possible future term. The supplied definition establishes what a row would mean in the underlying fictional record.
Write the outcome as acceptance of registration by four on that Saturday. That definition excludes completion on a later day. Someone classified as unsuccessful here might eventually enroll, so the result should not be described as permanent exclusion. A later-outcome study could matter, but it would need a follow-up period and evidence the packet does not contain.
Write the grouping variable as required documents ready or missing at first contact. Avoid replacing those categories with prepared and unprepared people. The latter labels imply a broader judgment about character or competence. A missing document might result from an institutional delay, an unclear requirement, a lost item, or a person's oversight. The records do not distinguish those possibilities.
These definitions are not decoration at the beginning of a report. They determine which arithmetic is meaningful and which interpretation is permitted. If the underlying records were ambiguous, the project would first need to resolve that ambiguity or narrow its question accordingly.
Make the comparison transparent
Among applicants with documents ready, completion changes from sixteen of twenty, or 80 percent, to seventeen of twenty, or 85 percent. Among applicants initially missing documents, it changes from four of twenty, or 20 percent, to ten of twenty, or 50 percent. Overall, it changes from twenty of forty, or 50 percent, to twenty-seven of forty, or 67.5 percent.
The recorded increase is larger in the group directly affected by the return deadline: thirty percentage points compared with five. The composition of the two document groups is the same across periods in this packet, so that particular compositional change does not explain the overall increase. Unlike the previous chapter's example, we are not comparing periods with sharply different shares of missing-document attempts.
That does not make the groups interchangeable. The missing documents could be easier to obtain in the final period. Applicants could differ in travel distance, available time, or familiarity with the process. Identical group totals do not establish identical people or circumstances. We have addressed one possible difference, not all differences.
The timing record adds a useful observation: seven completions in the final period occurred during the newly permitted interval. It connects the recorded pattern to the proposed mechanism more directly than the overall percentage does. Yet seven is not an estimate of seven additional completions caused by the rule. That claim would require knowing what those applicants would have done under the previous deadline.
Put the objection inside the design
A serious alternative is that the later period happened to include easier missing-document cases. If retrieving a document took ten minutes rather than several hours, more applicants might complete regardless of the deadline. A future study could record the kind of document needed and relevant circumstances at first contact, using consistent categories. That would help examine whether case differences explain the pattern.
Another alternative concerns staff assistance. The staffing note establishes the scheduled headcount, not the amount or quality of help. Two experienced employees might perform differently from two newly trained employees. An absence or a long interruption might also matter. Treating a fixed staff count as proof of identical service would give a limited record more authority than it deserves.
The group with documents already ready provides a useful comparison because the deadline change should affect it less directly. Its small increase suggests that something beyond the return deadline might also have changed. However, this group is not a guaranteed control group: the policy could alter congestion or staff availability for everyone. A comparison group earns credibility through an account of the process, not through the word control.
Your proposal should state the strongest feasible objection and identify evidence that could bear on it. A generic sentence saying the sample is small is less useful than explaining which difference might produce the observed pattern and how the proposed method could investigate it.
Draft a one-page proposal
Here is a model proposal. It is an original example of the final task, not a report of research that has occurred.
Question and scope. I will compare recorded same-day registration completion across four fictional Saturdays before and after the return deadline changed from noon to four. I will examine applicants missing documents at first contact separately from those whose documents were ready. The study concerns recorded first attempts across the supplied courses, not all people who wanted to enroll.
Concepts and explanation. Completion means acceptance by four on the same day. The proposed mechanism is that the later deadline creates additional time to retrieve required documents. The document category is measured at first contact. I will avoid treating missing documents as evidence of low motivation or an inflexible work schedule.
Evidence and method. I will compare the two notices, calculate completion rates from the supplied aggregate, and examine the timing record. I will report numerator and denominator for each group and period. The staffing note will establish scheduled headcount only. This is a document and record analysis using fictional supplied material; it involves no participant recruitment.
Comparison and alternatives. The ready-document group provides a limited comparison, but may also be affected through congestion or assistance. Differences in document difficulty, staff experience, and applicant circumstances could explain part of the pattern. Additional consistently recorded information on those factors would strengthen a later study.
Expected contribution and limit. The analysis can describe the pattern and assess whether it is consistent with the proposed time-constraint mechanism. It cannot isolate the causal effect of the deadline or describe people absent from the records. I will report contrary evidence and preserve those limits even if the observed increase is large.
Review the proposal as a reader
A reader should be able to trace each proposed claim to a source. The notices support statements about formal requirements. The aggregate supports completion rates under its definitions. The timing record supports the existence of completions during the extended interval. The staffing note supports scheduled headcount. No source in the packet supports a claim about participants' employment conditions.
The proposal also makes a result capable of disappointing its author. If the timing record showed no completions after noon and no other evidence of a deadline-related process, the proposed explanation would receive less support. It might still operate through some other route, but that route would need to be specified and investigated. A revised explanation should be acknowledged as a revision.
Finally, the proposal distinguishes an empirical contribution from a policy judgment. A person may favor a later deadline because it offers flexibility, even before a causal effect is estimated. Another may worry about staff workload. The records can inform those judgments, but they do not measure every cost or decide which value should take priority.
Carry the practice beyond this case
For a different topic, preserve the same discipline while changing the evidence. A study of a public park's posted rules might compare dated notices and the activities they permit. A study of a published survey might examine question wording and population definitions. A historical institution might require an archive. None needs to begin with an intrusive encounter or a collection of identifiable private information.
The intellectual move is to make a broad interest answerable. Specify the setting, identify the unit, explain the proposed relationship, and choose evidence that could bear on it. Keep a distinction between what people say, what records contain, what you infer, and what you think ought to happen. These distinctions make social life more interesting because they reveal questions that a quick character judgment or a familiar slogan would close too soon.
Application
Allow thirty to forty minutes. Write your own one-page research proposal using the supplied packet, or a clearly identified public document set you can actually inspect. Include a question, scope, concepts, main method, comparison, proposed explanation, alternative explanation, and limitation. You may adapt the model, but choose the wording and emphasis yourself.
Before finishing, check five things:
- Could the named evidence actually answer the question?
- Are the population, unit, period, numerator, and denominator clear where relevant?
- Does the explanation identify a process rather than merely attach a theory label?
- Is there a credible alternative and an observation that could weaken the preferred account?
- Are descriptive findings, causal claims, and value judgments kept distinct?
For the supplied packet, a satisfactory proposal correctly calculates 80 to 85 percent for ready-document attempts, 20 to 50 percent for missing-document attempts, and 50 to 67.5 percent overall. It treats the seven late completions as relevant timing evidence, not a proven causal effect. It does not infer work schedules or experiences of people outside the records. Success is a feasible, honest investigation whose evidence can change the explanation.