Paul Wallin and Leslie Waldo, two sociologists from the 1960s, embarked on a study to understand the impact of social class on children's academic performance. They attempted to gauge social class by asking 2,002 eighth-grade students to detail their fathers' jobs. However, they encountered a problem when trying to sort these jobs into social classes. They found that 17% of the descriptions were too ambiguous to categorize, and 5.5% of the students didn't provide any information at all.

This situation underscored a significant obstacle in social science research, which is the process of converting raw data into substantial evidence that can back up theories and ideas. The process involves three key elements: data, evidence, and ideas. Data transforms into evidence when it's used to bolster an argument that promotes an idea. However, for data to evolve into compelling evidence, it needs to be accurate, reliable, and valid.

Wallin and Waldo came to understand that their data was too flawed to make any definitive statements about social class and its effects. This is a typical problem when relying on self-reported data without any independent verification. These known sources of error should be seen as constant threats that need to be mitigated through rigorous methods, rather than anomalies that occur once in a while.

The individuals who gather the data also play a crucial role. For instance, interviewers might skew responses to wrap up quicker. Similarly, record keepers might manipulate data to serve their organization's interests. Therefore, researchers need to take into account how each person involved in the data collection process can affect its validity as evidence.

Moreover, difficulties encountered in conducting research properly can shed light on new areas to explore. Instead of overlooking persistent errors in data collection, researchers should view them as valuable hints about the underlying social phenomena they're studying.