Non-probability sampling techniques are methods where the selection of participants is not based on random chance, meaning not every member of the population has an equal probability of being selected. Instead, selection relies on the researcher's subjective judgment, convenience, or specific criteria. These methods are often used in qualitative research, exploratory studies, or when a complete sampling frame is unavailable. While they limit generalizability, they are valuable for in-depth understanding and hypothesis generation.

Here are the different types and their conditions of usage:

1. Convenience Sampling (Accidental or Haphazard Sampling):

   • Description: Participants are selected based on their easy accessibility and proximity to the researcher. It's the simplest and least costly method.
   • Conditions of Usage: Often used in exploratory research, pilot studies, or when time and resources are limited. It's suitable when the specific characteristics of the population are not critical to the research question, or when the goal is to gather preliminary insights rather than generalize.
   • Example: A researcher surveying students in a university cafeteria about their opinions on campus food. The sample consists of whoever is available and willing to participate at that moment.

2. Quota Sampling:

   • Description: The researcher identifies specific characteristics (e.g., age, gender, ethnicity, socioeconomic status) relevant to the study and sets quotas for each subgroup. Participants are then selected non-randomly (often conveniently) until the quota for each subgroup is met.
   • Conditions of Usage: Used when the researcher wants to ensure that specific subgroups are represented in the sample in certain proportions, but random selection within those subgroups is not feasible or necessary. It's an attempt to make the sample somewhat representative of the population's demographic structure.
   • Example: A market researcher wants to interview 100 people about a new product, ensuring that 50 are men and 50 are women. They will stop people on the street until they have reached their quota for each gender.

3. Purposive Sampling (Judgmental or Expert Sampling):

   • Description: The researcher deliberately selects participants based on their expert knowledge, specific experiences, or particular characteristics that are highly relevant to the research question. The selection is intentional and guided by the study's objectives.
   • Conditions of Usage: Ideal for qualitative research where in-depth insights from specific individuals are crucial. It's used when the researcher needs to target a particular group or individual who can provide rich, detailed information due to their unique position, experience, or expertise.
   • Example: A researcher studying the challenges of implementing a new educational policy might purposively interview school principals, district administrators, and experienced teachers who have direct involvement with the policy.

4. Snowball Sampling (Chain-Referral Sampling):

   • Description: Initial participants are identified and recruited. These participants are then asked to identify and refer other potential participants who meet the study criteria. This process continues, creating a 'snowball' effect as the sample grows.
   • Conditions of Usage: Primarily used when studying hard-to-reach, hidden, or marginalized populations where a complete list or sampling frame is unavailable (e.g., undocumented immigrants, drug users, specific professional groups, or individuals with rare diseases). It relies on the social networks of initial contacts.
   • Example: A researcher studying the social networks of homeless youth might start by interviewing a few initial contacts and then ask them to refer other homeless youth they know.

While non-probability sampling methods are efficient and practical, researchers must acknowledge their limitations, particularly regarding the generalizability of findings. They are best suited for generating hypotheses, exploring complex phenomena, and gaining deep contextual understanding rather than making statistical inferences about a larger population.