The correct answer is b. self-selection.
Online surveys commonly suffer from self-selection bias, which is a significant threat to the external validity and generalizability of the findings.
Self-selection occurs when individuals choose to participate in a survey or study on their own accord, rather than being randomly selected from a population. This can lead to a sample that is not representative of the target population, as certain types of people may be more likely to participate than others.
Some reasons why self-selection bias is common in online surveys include:

1. Voluntary participation: Online surveys are often advertised or shared, and people voluntarily decide whether or not to participate.
2. Access and interest: Individuals with internet access and an interest in the survey topic are more likely to self-select into the study.
3. Non-response bias: Certain groups or individuals may be less likely to respond to online surveys, leading to an unrepresentative sample.

Regarding the other options:
a. Poor reliability: While reliability is an important consideration in survey design, it is not a factor that is inherently associated with online surveys. Reliability refers to the consistency and reproducibility of measurements.
c. Probability sampling: Probability sampling is a method where each member of the population has a known, non-zero probability of being selected. Online surveys often use non-probability sampling methods, such as convenience or snowball sampling, which are more prone to self-selection bias.
d. Over-sampling: Over-sampling refers to the practice of deliberately sampling certain subgroups at a higher rate than others, typically to increase the precision of estimates for those subgroups. It is not a common issue associated with online surveys specifically.
Therefore, self-selection bias is the most common issue that plagues online surveys, as the voluntary nature of participation can lead to a sample that is not representative of the target population.