Rule 1: There should be the possibility of surprise in social research

Guidance to Relevant Questions - Objections

• The no-surpises objection - if the research premise will clearly do what has already been done or will clearly lead to a known result.
• The so-what objection - if the research premise will clearly bring nothing useful.

Investigating Prior Research

• Challenge prior research
• Clarifying prior research
• Extend prior research

Causal Chains

Introduction of mediating variables

Identify pathways by which independent variable X affects dependant variable Y through mediating variable Q, where Q is are various affecting pathways.

In other words, various potential causal pathways that can be identified between two conditions.

Introduction of nonmediating variables

Identify an association between variable X and variable Y through variable Z, where the link through Z between the two variables X and Y is not causal.

In other words, two conditions are linked by a third condition through association that does no necessarily mean X leads to Y. Sometimes, through the addition of control variables, the association between X and Y can become partly causal, partly not

Introduction of moderating variables

Moderating variables condition effect between variables.

In other words a characteristic that can be identified which has an overall effect between two conditions.

Sampling

• First Principle of Sampling - effectively states that one grouping can allow you to identify characteristics of the whole
• Second Principle of Sample - effectively states that representative sampling is more important than size.
• Second Principle of Sample - effectively states that representation should produce contrasts.
• First Principle of Sample Size - effectively states the sample size need not scale linearly with population size
• Sample Size Corollary - effectively states that generalisations can be made from small but representative groups.

References

Seven Rules for Social Research, Glenn Firebaugh, 2008