Week 1 lecture CNTD, Standard deviation, z scores, measurement, COIS, correlation and psychometrics Flashcards
What is a standard deviation and standard error of the mean?
SD: Variance gives us a measure in units squared, taking square root of variation gives SD
SEM: measure of spread, what SD would look like if measured in real world
What are the 5 measures of spread? (CRISS)
Confidence intervals
Range
interquartile range
Standard deviation
Standard error of mean
What are confidence intervals and what represents them? (what couldn’t you figure out how to make elliott’s first assignment ?) 95/100…
Defines range taht will theoretically capture true mean 95% of the time (95 times out of 100 )
Represented by error bars
Why does the mean, mean nothing and what’s possible when a chart doesn’t have this?
Means nothing as there’s no measure of spread, when a chart doesn’t have a measure of spread they either are stupid or think im stupid
How do we know when error bars are significant or insignificant?
When error bars overlap quite a bit means probably insignificant, when error bars do not overlap significantly means it’s usually statistically significant
What three things all represent the same thing? (squares,very,S..)
sum of squares variance and standard deviation all represent same thing just different amounts
What are z scores and it’s formula?
Expresses score in how many Sd’s it is away from the mean
z= x-x(bar) ~(divided) SD
What are the 5 property’s of z scores ? cuts off top/bottom, lie between…
(cuts off 2…, lies: 1..,2..,3..) (95% of,99% of,99.9 %of)
1.96 cuts off top 2.5 percent of distribution
-1.96 cuts off bottom 2.5 percent of distribution
95 percent of z scores lie between 1.96 and -1.96
99 percent of z scores lie between 2.58 and -2.58
99.9 percent of z scores lie between 3.29 and -3.29
what does p=0.49mean as an alpha level and what does it mean if p=0.5 is alpha level of expirement?
P=0.49- less than 5 times out of 100 an error can happen and we are okay with that, alpha level set to show error we are okay with getting
P=0.5- we are okay with error being 5/100 (error rate we are okay with having)
What is a correlation and it’s 3 factors? (change, weight scale, two things don’t cause eachother)
Co variation/relation between 2 variables
3 factors:
variables change together
usually scale variables
correlation doesn’t mean causation
what is a correlation coefficient and it’s 3 factors (yay or nay, 1, STRONG based upon..)
statistic that quantifys a relation between two variables
3 factors
can be positive or negative
falls between -1.00 and 1.00
value of number indicates strength
What is a positive and negative correlation and what kind of relation do they have? (smiles for positive)
Positive : One variable has high score and other variable tends to as well
-direct relation between two variables
Negative: One variable has high score while other has low score
-inverse relation between variables
What are two scale variables (NOIR) and what’s a pearson correlation coefficient and what two letters can it be represented by? (what do you get when making personal best)?
two scale variables: interval and ratio
Pearson coefficient is a statistic that quantifys linear relation between two variables
Can be represented by Italic R for sample data or Italic P for population parameter
How many strengths of correlation are there and what numbers do they represent? ( like shirt sizes)
Small- 0.10
Medium-0.30
Large- 0.5
What are the two limitations of correlation and what does it mean? (no full popu…, switch from negative to positive)
Restricted range- smaller range than full population, correlation can become smaller
Effect of outlier- outlier can make correlation much stronger than supposed to and even change from negative to a positive
What are the 6 steps to correlation hypothesis testing? ( Identify, state,find and determine, determine, calculate, make..)
- identify population, distribution and assumptions
2) state null and research hypothesis
3) Find and determine characteristics and compare
4) determine cortical values
5) calculate statistics and tests
6) make a decision
What do psychometric and psychometricians do? (2 factors) development..
Psychometric are used in development of tests and measures
psychometricians use correlation to examine 2 aspects of development of measures
2 Factors
1) reliability 2) validity
Why are correlations used in psychometrics?
to help assess reliability of pro sports team or establish validity of personality test
What are the 2 meanings of stats and the 2 ways to represent data? (maths) (g.., history..)
2 meanings of Stats: 1) field that applys to mathematical techniques to interpret and summarize data 2) Mathematical techniques themselves
2 ways to represent data
1) frequency distributions (histograms)
2) Graphs
What does a frequency histogram represent and when do the bars touch and not touch ? (gender,age and weight)
Represents frequency data in bar form
Non continuous variables (gender) bars don’t touch
Continuous variables (age, weight) do touch
What type of statistics is used to measure thousands of high school grade point averages and what else does it measure? (2) measures: CT, V
Descriptive statistics to obtain precise scores, measures central tendency and variability
Where is the frequency of scores the strongest and weakest?
Strongest/greatest near the mean and progressively decreases towards extremes (ends)
What are the three frequency scores? 60.. something falls between what and what SD’s away from the mean?) (-1,2,-3)
68 % of scores fall between -1 and 1 standard deviations away from the mean
95% fall between -2 and 2 SD’s away from mean
99% fall between -3 and 3 SD’s away from mean
What’s the difference between Parametric and Non parametric stats(and what’s another name for it)? (one needs somthing, other doesn’t need it)
Parametric needs population parameters to do inferential procedures
Non parametric can be called distribution free stats as it doesn’t need population parameters to do inferential procedures
What kind of data is non parametric and parametric stats, and what types of tests fall under both (2 each) ?(sneaky and always going..)
Non parametric- discrete data : Chi square, Kruskal wallis,spearmen,kendall’s tau
Parametric- continuous data : T tests, Anova
Which tail design has more power (1 or two tail) and why?
One tail designs have more power as they can detect significant difference because they’re less conservative (more likely to find something)
What does a perfect correlation (constant) mean and inferential statistics? (inferences)
Perfect correlation means variables always stay the same
Inferential stats is information about a sample that can be used to make inferences about a population
