Ethnographic Sampling List (Essential Ethnographic Methods – SSL)

Convenience sampling: “any group readily accessible to the researcher that reasonably might be assumed to possess characteristics relevant to the study” (233)

Criterion-based sampling: “choosing study participants or units because they possess characteristics related to the study’s central questions” (235)

Extreme or dichotomous case selection – “representing the ends of a defined population continuum” (237)

Typical case selection – “involves selection based on a known average for the population” (238)

Intensive case selection – selection based upon the assumption that those studied in a geographic proximity will exhibit similar behavioral, cultural, and cognitive patterns. (We have to be careful not to generalize these observations to wider demographics of the population, as just because it is observed here does not mean that this trait is endemic to the whole population.)

Unique case selection: “selecting for study a nonreplicable person, event, or situation” (239)
– looks for least common cases, outliers to the normal flow of events or not-truly-repeatable patterns

Reputational case selection – “asking experts from, or participants in, the community who are familiar with the criteria of interest to the researcher to recommend individuals for participation in the study” (240)

Chain-referral selection – “asks participants in an activity or people who possess specific characteristics to identify others known to them who share those activities or characteristics” (241) – see also, snowballing. Index respondents are the points of reference from which the researcher starts their chain of reference. Chain-referrals can also be combined with randomized sampling techniques

Bellwether/ideal case selection – the most similar to a perfectly conducted experimental success state, case study

Comparable case selection: cases that are “selected because each exemplifies as closely as possible characteristics of interest to the researcher” (244) – looks for similarities in separate sites across time to examine similar processes; also called meta-analysis

Quota sampling – not probablistic, but does attempt to select an appropriate amount of representativeness from each known group within a population to be shown in a sample. Is useful at early stages of research to show the breadth and variety of behaviors and characteristics to be studied.

Targeted selection – like quota sampling, but only possible when researcher has plenty of secondary data in which they can frame their sampling frame, and designate high-priority areas of review

    Probabilistic Sampling Methods

Systemic – must determine what is desired to be sampled and how many units of that sample will be necessary to complete the study. Selection of a sampling interval – estimated population divided by desired units = sampling interval (every 4th respondent, etc.)
– perks: doesn’t require every single unit to be sampled to be identified in advance and placed on a list, used in most studies of social interactions where ideas cannot be ennumerated

Random – must know characteristics of entire population in advance; every member or unit in the population must be placed on a numbered list from the first to last; each member of the population must be available to the researcher — can select sampling interval and sampling frame from random number list

Stratified sampling – identifying significant groups within the population and sampling from each of these groups separately. Aids in comparison of samples, generalization of population and representation

Cluster sampling – Identifying natural groupings of the population (ie classrooms, neighboorhoods) and creating sample from these units

Cross-sectional – “snapshot”- examines population at one period of time

Trend studies – replication of a study several times over designated period of time

Cohort studies – follows a population of interest over designated intervals of time; helps in studying historical/developmental influences

Panel studies – selection of a specific sample or group and follow same sample over time

— when do we end? “Informational saturation” (aka sufficient redundancy (262)- when our study produces no new information.
— how big should my sample be? The greater the heterogeneity of the target population, the greater the sample size needs to be to be representative of the population.

Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: