STATS exam 3

0.0(0)
Studied by 0 people
call kaiCall Kai
learnLearn
examPractice Test
spaced repetitionSpaced Repetition
heart puzzleMatch
flashcardsFlashcards
GameKnowt Play
Card Sorting

1/44

encourage image

There's no tags or description

Looks like no tags are added yet.

45 Terms

1
New cards
Sampling
the act of selecting observations from which to collect data
2
New cards
Sample
observations selected
3
New cards
Population of interest
all observations that we are interested in studying or making a statement about
4
New cards
Inclusion criteria
the way statisticians and data scientists define who or what the population of interest is by specifying inclusion criteria based on certain attributes
5
New cards
Estimates
we call the value we use as a stand-in for the parameter
6
New cards
External validity
the degree to which a sampling strategy supports making a generalization inference
7
New cards
Parameters
the summary statistics for attributes based on an entire population of interest
8
New cards
Sample statistic
when we compute summary statistics for a sample
9
New cards
Generalization inference
the act of using a sample statistic to determine the value of a parameter
10
New cards

Representativeness

the extent to which the observations in the sample are like the observations in the sample are like the observations in the population of interest, in terms of important underlying baseline characteristics.

11
New cards
Sample bias
when a sample is not representative of a population of interest, it is not appropriate to use sample data to make generalizations about the population of interest
12
New cards
Who’s not here
one easy way to help you remember to think about representativeness
13
New cards
Volunteer sampling
a sampling strategy in which you make a survey or study available to some or all respondents who meet the inclusion criteria
14
New cards
Another term for volunteer sampling
is convenience sampling
15
New cards
Probability sampling
a sampling strategy in which all eligible observations are assigned a non-zero probability to be included in the sample, and then some observations are selected using a random process
16
New cards
Strata
statisticians and data scientists often incorporate some sort of grouping into their sampling strategy; similar groups in the population of interest are called strata.
17
New cards
Quota sampling
a sampling strategy in which you have specific quotas for specific strata of respondents within the population of interest
18
New cards
Social exchange theory
posits that in exchange for giving their data, respondents should get something in return
19
New cards
Survey incentive
the act of offering something in return for a respondent providing their data
20
New cards
Non-response
when a respondent selected for the sample chooses not to respond to the survey, either to specific questions or to the entirety of the survey.
21
New cards
Bias
any systematic manner in which the data that has been collected has fundamental problems that will prevent the statistical methods from producing a correct estimate
22
New cards
Measurement bias
any bias in the data due to the problems in the measurement process
23
New cards
Sample bias
any bias in the data due to problems in the representativeness of a sample
24
New cards
Causal bias
any bias in the data due to differences between the controls and the observations in the treatment group
25
New cards
Estimate’s precision
how much uncertainty we have, and therefore how much room for error we give, when we make an estimate of a parameter
26
New cards
Posterior probability distribution
is a special type of distribution that specifies how likely different potential values of a parameter are given the sampling strategy, data, and estimate that was collected.
27
New cards
Point estimate
our estimate of the population parameter. Without incorporating ish-ness
28
New cards
Interval estimate
instead of using a single number as an estimate, we can use the entire middle 95% of the posterior distribution to provide an entire range of numbers as our estimate
29
New cards
Evidence-based testing
a form of statistical testing that begins with evidence in the form of data, creates a posterior distribution for a parameter, and subsequently evaluates hypotheses by comparing them to the posterior distribution
30
New cards
Hypothesis-based testing
a form of statistical testing that begins with a hypothesis, creates a probability distribution based on that hypothesis and the estimation error, and subsequently evaluates estimates from data by comparing them to the probability distribution
31
New cards
Null model
represents your expectations for what the data will be
32
New cards
P-value
the probability of seeing something even further from what was expected than the observed data
33
New cards
Regression
a method that is based on keeping the distance from all the points to the line as small as possible
34
New cards
Residual standard error
describes the typical prediction error
35
New cards
Prediction interval
what we think the value of the response variable will be for a single observation, if we know that observation’s values for the explanatory variables
36
New cards
Prediction interval margin of error
= 2 * residual error
37
New cards
Prediction interval lower limit
= predicted value +/- 2*residual error
38
New cards
Actual value
= predicted value + predicted error
39
New cards
Actual value - predicted value
= prediction error
40
New cards
Prediction z score
= prediction error/residual standard error = (actual value-predicted value)/residual standard error
41
New cards
Estimate z score
= (estimate-hypothesized value)/estimation error
42
New cards
How to create an interval estimate using equations:

1) Compute the point estimate from the sample. 2) Get the estimation error (standard error) by simulation. 3) Margin of error = 2 * estimation error. 4) Lower limit = point estimate - margin of error. 5) Upper limit = point estimate + margin of error. 6) Report both limits: “The middle 95% of the posterior distribution lies between [lower limit] and [upper limit]. Therefore, we believe that the true value of the [parameter] of [attribute of interest] for [population of interest] is probably somewhere between [lower limit] and [upper limit].”

43
New cards
P-value scale (image)
Left end: p = 0.001, data is probably not consistent with the null model. Then p = 0.02, then p = 0.10. Right end: p = 0.25, data is probably consistent with the null model.
44
New cards
R-squared scale (image)
R² = 0 and R² = 0.1: No Utility. R² = 0.3: Slight Utility. R² = 0.5: Moderate Utility. R² = 0.7: Strongly Utility. R² = 1.0: Perfect Utility.
45
New cards
Prediction z score scale (image)
Left end: z = -4, the actual value is surprisingly lower than predicted. Then z = -2, z = 0, z = +2. Right end: z = +4, the actual value is surprisingly higher than predicted.