Please enable JavaScript.
Coggle requires JavaScript to display documents.
U4 Concept Map (Sampling Errors (Convenience Sampling (When one is…
U4 Concept Map
Sampling Errors
Convenience Sampling
When one is attempting to gather a convenience sample, they are trying to choose individuals who are easiest to reach the results that they are looking for.
-
Undercoverage
is when some of the population is misrepresented and it effects the data as a whole. It occurs when some groups in the population are left out of the process of choosing the sample.
Good Sampling Techniques
SRS
A simple random sample- these are difficult for large groups of items or people, and not practical. It can be difficult to get an accurate list of the population, and can be time consuming.
Stratified SRS
Classify the population into groups of similar individuals (strata), then choose a separate SRS in each strata and combine them to have a full sample. Works best when the population of a strata are similar in a way that affects the variable being measured, with large differences between groups.
Cluster Sampling
A cluster sample is when you divide the population into smaller groups, and chose an SRS of the clusters. Easier to use when a population is large and spread over a large area. Selects individuals that are "near" each other. Clusters may still have variations in the results. An example could be a school administrator wanting to give a survey to 200 students, and taking an SRS of 8 homerooms instead of 200 students, and give the survey to to the students in the selected homeroom.
-
Bias
is when one makes a mistake while sampling and their results of their data turn out to be poor because of it. A bias can effect the data as a whole and mess up certain percentages and probabilities.
Sources of Bias
could occur depending on the different characteristics of the samples/ subjects and the different kinds of treatment that they get.
-
Observational Study
An observational study is NOT the same as an experiment, no treatment is given, the investigators only observe, compare, and analyze.
Experimental Units
The unit of statistical analysis. Must be able to receive different treatments. When the experiment or study involves people, see "Subjects".
Subjects
Refers to the population. When the data includes people, they are refered to as subjects, not experimental units.
Lurking Variable
A variable that is neither a explanatory or response variable, but can still influence the response variable.
-
-
Population
is the amount of observations that one can make with a set of data. When referring to people, they are called subjects. When referring to anything else, call it/them experimental units.
Confounding
Confounding occurs when two variables impact a response variable, but the two variables effects cannot be distinguished from each other.
Double Blind
A double blind experiment, neither the subjects or the people who interact with them know which treatment they are giving/receiving, this is the most effective type of experiment since it eliminates bias.
-
Treatment
Treatment is the kinds of conditions that are applied to the experiment. For example if someone is experimenting with how sunlight effects certain plants, the treatments could be changing the amount of sunlight that the plants have access to.
-
-
