WGU C955 Probability and Statistics (2022/2023) Already Passed

WGU C955 Probability and Statistics (2022/2023) Already Passed Boxplot an image that has min, Q1, median, Q3, max Histogram A graphical representation -- bars, measuring the frequency within each interval Skewed right Not a symmetric distribution, the tail is on the right, i.e. extra stuff on the right Measures of center Median, the mean (and mode) Measures of spread Range, IQR & standard deviation Standard Deviation Rule 68% of the data are within 1 standard deviation, 95% are within 2, 99.7 are within 3 standard deviations from the mean. For skewed data, use these for center and spread In this situation, we use median (for center) & IQR (for spread) Explanatory variable In a study, what we think is the "cause" Response variable In a study, what we think is the "effect" Scatter plot A graphical representation of Q -> Q Two way table A graphical representation of C -> C Side-by side box A graphical representation of C -> Q Linear relationship "shaped like a line" Correlation coefficient, r Between -1 and 1; measures how close the points are to the line and if the trend is uphill (positive) or downhill (negative). r = -0.2, for example This is an example of a correlation coefficient that represents a weak negative correlation. r = 0.9, for example This is an example of a correlation coefficient that represents a strong positive correlation. Linear regression line A line that fits the data as close as possible, used to make predictions Interpolation Making predictions *within* the range of your data. This is usually accurate. Extrapolation Making predictions *outside* of the range of your data. This is generally a bad idea. Simpson's Paradox When split up, each data set can have a pattern which goes away when all the data is combined. Only way to prove causation Experiments, because they account for lurking variables Observational study A type of study where we measure or survey members of a sample without trying to affect them. Cannot prove causation. Experimental study A study where you split subjects up randomly and impose a change on one group to study the effect; can prove causation Prospective study Is a study that's done over time to find results / A study watching for outcomes during the study period Retrospective study A study that looks backwards to assess outcomes and possible causes after the fact Control group randomly assign people or things to groups. One group receives a treatment and the other group does not. This is the group that does not receive treatment Placebo A substance or procedure that has no effect used for comparison to the real substance or procedure Placebo effect A beneficial effect produced by the belief of the patient/subject, not by the intervention itself. Experimenter effect When the persons running an experiment affects its results by influencing the subjects inadvertently Open question A question that gives the responder freedom to answer in many different ways -- harder to analyze with statistics. Closed question A question with limited choices, e.g. multiple choice or yes/no. Easier to analyze statistically. Unbalanced response Giving more options that are negative than positive options (or vice versa) which biases the responses towards the more common option Matched pairs Grouping two similar subjects and giving different treatments/procedures to each in order to compare the differences. For example, having one twin take a medication while the other twin does not. Blind experimental study When information of a study isn't revealed to the participants Double Blind Study When information of study is hidden from the researcher and the participant. Population The entire group you are trying to describe or understand. Sampling frame List of group from which you choose your sample. Sample The group that is actually picked to be included in a study Simple random sample making a selection by following a random pattern and selecting without replacement. Unbiased. Systematic sample Sample is selected by listing the sampling frame, then making a selection by following a simple pattern (eg. Every 20th name). Unbiased. Voluntary sample Members of the sample may choose not to respond. Similar to Non Response. Convenience sample Participants are easy for researcher to access. Tends to increase bias. Cluster sample Sample frame is divided into groups, we select a few groups, then selecting ALL of the members of those groups. Stratified sample Sample frame is divided into groups. Then we choose a random sample (usually the same size) from within EVERY group. Multi-stage sample Multiple rounds of randomness and grouping. For example: randomly selecting a few groups, then choosing a small random sample just those selected groups. Often a combination of Cluster and Stratified Sampling. The formula for all simple probabilities (number of possible outcomes for the specific event) / (total number of possible outcomes) For independent events, P(A and B) = ? P(A)*P(B) For independent events, P(A|B) = ? P(A) [this is the right side of which probability formula?] P(A|B) = ? P(A and B) ÷ P(B) P(A or B) = ? P(A) + P(B) - P(A and B) In general, P(A and B) = ? P(A)*P(B|A)