Need to know

What are the two types of statistical samplings?

1. Random sampling

2. Non-random sampling

What is considered random sampling?

• SRS: simple random sampling

• Stratified sampling

• Cluster sampling

• Systematic sampling

What is considered non-random sampling?

Convenience sampling

What is SRS (simple random sampling)

Every member within a population has an equal chance of being chosen for the study.

• Ex: A high school has 100 students in the senior class and 5 students are randomly chosen to take a survey about the lunch

What is stratified sampling?

The target population is divided into groups (strata) and from these groups a random number of people will be chosen to be in the survey

• Ex: You want to survey a university so you break the population into grade types and then you randomly select students from each group.

What is cluster sampling?

You have a target population that you divide into clusters and then from these clusters you randomly choose which will be used within the study and those in these clusters will be surveyed.

• Ex: The target population is all high schools in the state so the clusters will be the high schools and then schools will be randomly selected to be surveyed.

What is systematic sampling?

For the target population, you choose every i^th number from a random starting point.

• Ex: You want to survey 10 students from a school population is 1000 so you survey every 10th student.

What is convenience sampling?

Data is gathered from an easily accessible/available group of people rather than random selection

• Ex: A professor hands out a mental health survey in her own psychology class because they are in front of her.

What are the different types of experimental design?

• Blocked randomization design

• Matched pair design

• Completely randomized design

• Double blind experiment

What is a blocked randomization design?

When participants are divided into subgroups based on shared traits and then are randomly assigned treatments within each block

What is a matched pair design?

Similar subjects are grouped into pairs based on shared characteristics before randomly assigning one person to a treatment and the other to a control

What is a completely randomized design

Treatments are assigned to experimental units entirely at random

What is a double blind experiment?

When neither the research study nor the participants know who is receiving the active treatment and who is receiving an inactive placebo.

What are the different types of bias?

• Response bias

• Non-response bias

• Volunteer bias

What is response bias?

When participants provide inaccurate or false answers on surveys, interviews, or questionnaires.

What is non-response bias?

The collected data fails to represent the target population before certain groups opt out or cannot be reached.

What is volunteer bias?

When people choose to join a study have different traits from people who do not, which skews the results and makes them less accurate for the general population

For Q x C what graph will be used?

• side by side

• groupwise summaries

For C x C what graph will be used?

• Stacked bar charts

• Side by side barplot

• Two way absolute frequency

• Conditional relative frequency table

For Q x Q what graph will be used?

• Correlation

• Scatter plot

What univariant summaries is used for just Q?

• Mean/median

• SD

• IQR

• Variance

• Histograms

• Five number summary

• box plot

What univariant summaries is used for C?

• Absolute frequencies

• Pie chart

• Relative frequencies

• Barplot

What is categorical nominal?

No natural ordering

• Major, residence hall, eye color, ethnicity, etc.

What is categorical ordinal?

Some natural order but group distance is unknown

• Pain scale, app rating in stars, etc.

What is quantitative discrete?

Integers

• Age in years, number of pets, number of falls when skating, number of siblings, etc.

What is quantitative continuous?

Measurement on interval

• exact age, tumor size in mm, time spent on hw in minutes, height in ft, etc.

What are the 4 assumptions for binomials?

B Bernoulli outcome: only one of two possible outcomes in each trial

I Independence: outcomes of n trials are independent

N Number of trials, n, are fixed

S Success probability, p, is constant for each trial

What are the 4 assumptions for poisson?

P probability of a single event occurring within an interval is proportional to the length of interval

O occurrences within an interval can be theoretically infinite

I Independence among events

S Situations involving count data (discrete)

How to calculate the whiskers on a boxplot?

1. Calculate the min, Q1, Q2, Q3, and max

2. Find the bounds

   1. If the values are within the bounds then there are no outliers and the whiskers will go to the min and max value

   2. If the values fall outside of the bounds then there are outliers and the whiskers will go the value within the data set that is before the outlier

How to calculate the outliers of a boxplot?

First need to calculate the boundaries which is:

• Lower: Q1 - IQR(1.5)

• Upper: Q3 - IQR(1.5)

The outliers will be the variables that are after these bounds

For scatter plots explain:

1. Direction

2. Strength

3. Form

1. positive, negative, neither (horizontal line)

2. strong, weak, none

3. linear, nonlinear, or no relationship (line or blob?)