stats (im gonna kms)

s

Sample SD

σ

Population SD

M

Sample mean

μ

population Mean

df

represents the amount of independent information available to estimate a statistic. Let you know which t- distribution to consult when finding t-values

Variability

How convincing is the difference how spread out are the scores in relation to their mean?

Cohens d

Measure standardized effect size how large is the difference Between two means relative to SD.

Ho evaluated by independent groups t test

Ho: μ=μ

Ho: μ-μ= 0

Paired design

Each participant is tested in each of two contusions (sensitive eliminates some SE)

N

Number of data points

X

Individual score

Experimental design

Has one or more variables that are manipulated. Random sampling

DV

Observed (anxiety levels)

DV

Observed (anxiety levels)

Operational definition

A precise procedure for measuring values that represent a construct. (Scores on a specified happiness questionnaire)

Reliability

Consistency (obtaining similar results)

Validity

Accurate (is it measuring what it says it’s measuring)

Nominal

Numbers represent characteristics according to a simple code.

No numerical meaning (qualitative)

Ordinal

Numbers are assigned based on rank amount of characteristic.

Can be compared: some results are better then others

Differences not interpretable (dif distances between them)

Interval

Numbers are assigned based on relative quantity of characteristic.

Has an arbitrary 0 (ex. Years)

Ratio

Numbers are assigned based on the absolute magnitude of a characteristic.

The scale has a True 0 (salary)

SD

a statistical measure that tells you how far apart data points in a set are from the average (mean).

Smaller=scores are more similar

Sensitive to outliers

Z-score

Expresses an individual score in SD units. How many SD away from the mean did a person score?

Provides standardization

Comparisons across different scales

Standard Normal Distribution

Mean= 0

SD= 1

68% of scores fall within 1 SD

95 within 2 SD

99.7 within 3 SD

SEM

Inferential stat that captures variability in our estimate of the population mean.

Small= increase precision

Large= decrease precision

How precise is our estimate of μ

CI and MoE

Captures a range of values from data that most likely contain the true value of what was estimated about the population mean. 95% will likely capture the true value.

MoE- measure of precision. Is half the CI

Cat eye Picture

Tells us about plausibility→ middle= more plausible, ends= less plausible

Z distribution

Uses population SD

SND

T distribution

uses sample SD

Graph wider to account for uncertainty

Shape depends on df (N-1)

Increase N approach’s z score

P value

Tell us the likelihood of obtaining a result, or a result more extreme if the null hypothesis were true.

The further the notice is from the point estimate to lower the P is

How surprising would the observation be if the null were true??

obtain t > critical t

Reject Ho

Obtain t <_ critical t

Fail to reject Ho

Increasing N

• Increases power, More likely to detect real effects

• Decreases sample error, increasing sample size gives more precise estimate of error

• Potentially decreases P value, makes it significant results easier to catch

Increasing N

• Increases power, More likely to detect real effects

• Decreases sample error, increasing sample size gives more precise estimate of error

• Potentially decreases P value, makes it significant results easier to catch

Ho

There is no effect/difference.

• Rejecting = We have enough evidence to suggest there is an effect

• Failing to reject = we don’t have enough evidence to suggest effect

Type 1 error

we reject the null hypothesis Despite it being true. We notice an effect when there really is none.

Type 2 error

Failed to reject all hypothesis and Noel is actually false

We miss an effect that is actually there.

Alpha

• protect against type one errors, decreasing alpha decreases the chance of mistaking chance for real effect

• Increases risk of type two errors, decreasing alpha increases the chance of this taking real effect for chance

  
  

Beta

• Decreasing sample size increases chance of making a type two error

• Increasing alpha decreases beta, which increases the risk of type one error

• Increase effects being studied decreases beta