Help
Options
Didn't know it?
click below
click below
Knew it?
click below
click below
Don't Know
Remaining cards (0)
Know
retry
shuffle
restart
0:00
Embed Code - If you would like this activity on your web page, copy the script below and paste it into your web page.
Normal Size Small Size show me how
Normal Size Small Size show me how
stats week 1
| Question | Answer |
|---|---|
| population (partameter) | entire group of entities info is wanted about |
| sample (statistic) | part of the population that is examined to gather info. used to infer info of parameter without spending too much time or money etc |
| satistic | gotten from data, calculated |
| data | what is measured |
| bias | something, like a statistic, systemically favoring an outcome |
| convenience sample | a sample of chosen individuals who are convenient to reach but may not represent a diverse/accurate population |
| voluntary response sample | consists of people who choose themselves by responding to a general appeal. Show bias as ppl w/ strong opinions r more likely to respond. |
| random sample | consists of a random assortment of individual participants |
| random sample problems | under coverage, non response, response bias (untruths) |
| undercoverage | when not enough groups are picked for the sample size to represent the parameter |
| non response | when someone is picked but does not participate (not picking up phone or not doing a survey) |
| response bias (untruths) | a member of the sample responds untruthfully, maybe to get it over with or answer is embarrassing etc |
| sample types | simple random sample of size N, stratified random sample, multistage sampling design. |
| simple random sample of size N (SRS) | picking at random intervals of a population to get sample. example: have 100 ppl, pick every 10th person. |
| stratified random sample | divide population into groups of similar entities (strata), choose separate SRS in each strata, combine all the SRSs to form the final sample |
| multistage sampling design | random selection is done at several stages.(i.e, selecting over several days or different states and cities) |
| anecdotal evidence | not reliable, not necessarily based on a sample. Just asking ppl around you that you know as an example |
| observational study | gather data without interfering with the sample ( i.e, survey ) |
| experiment | deliberately impose a treatment on our entities (or individuals) to see what happens |
| 1 experimental unit | the unit we do the experiment on (i.e: human, man, dog, girl) |
| 2 treatment | the experimental condition applied to the units |
| 3 factors | explanatory variables (the outcomes of the experiment may depend on these variables). they are assigned/controlled by researchers intentionally at different levels |
| 4 levels | specific values of the factors |
| 5 response variables | what is measured for each unit (collected as data) |
| things to watch out for in experiments | lurking or confounding variables, placebo effect, bias, lack of realism |
| lurking or confounding variables | variables not taken into consideration |
| placebo effect | the response might be affected by the subject's condition |
| bias | results skewed in a certain direction |
| lack of realism | the experiment reproducing the "real" situation correctly |
| 3 principles of experimental design (to show causation) | 1. control group, 2 randomization, 3 replication |
| control group | not getting treatment |
| randomization | random assignment |
| replication | enough results |
| 3 types of experiments | completely randomized designs, randomized block design, matched pair design |
| completely randomized design | random assignment at beginning |
| randomized block design | groups separarted then random assignment |
| matched pair design | 1. two groups, same treatment, diff order, one score 2. one group, same treatment, two scores |
| causation | can only be shown through a carefully controlled experiment; but, when not possible needs strong variable association, consistent association, higher values of 1 variable imply higher value of other, alleged cause preceeds effect and is probable |
| ethics (humans) | need approval before starting experiment by review board before human subject. Informed consent is necessary and confidentiality too |
| Review board | approve an experiment before it happens if human subjects are involved |
| informed consent | tells a person what is going to happen to them in the experiment. subject signs, minors need adult consent. |
| confidentiality | researcher guarantees the subjects wont be identified if results are published or spoken about |
| ethics (animals) | replacement, reduction, refinement |
| replacement | were non animal subjects or lower species like cells considered before |
| reduction | show using the minimum number of animals needed for experiment |
| refinement | explain how animals will be treated and how they will be euthanized after experiment. |
| role of statistics | steps: 1. questions about population, 2 sample design, 3. experimental design, 4. get data, 5. use statistics to infer |
| categorial variables | a word, not number. (ie gender) |
| quantitative variable | number (ie, age) |
| data plotting | for each unit we collect a value of the response variable. Look at each value and count how many times it appears (frequency) examining the distribution of the response variable to look for patterns |
| categorial variable graphing | bar graph, pie chart |
| bar graph | horizontal axis-value of variable, vertical axis- frequency (number or %), height of bars- frequency, variables dont need to appear in any particular order. Bars dont touch |
| pie chart | full circle-100%, frequency of each variable: % |
| quantitative variable graphing | stemplots, histograms, boxplots |
| stemplots | need to take data and sort from smallest to largest. A hand drawn plot of data. used for smaller data sets |
| histogram | Used for larger data sets, touching bars. Horizontal axis: continuous range of values for variable, Vertical axis: frequency (# or %) corresponding to different times, Vertical bars for each bin |
| p% percentile | p% of the observations fall at or below. quartiles like 25th percentile and 75th percentile with 50% as the median percentile. |
Created by:
user-1742075