Exam 4 Lecture 1 Flashcards
Why people use statistics
We have questions, so we collect data. The goal is to answer our questions by comparing variables or groups using math and probability theory.
Types of questions that statistical analysis can answer
Are things different?
Are things changing?
Are things related?
Is the likelihood of something different?
Who is more likely to eat red meat: resistance exercisers and aerobic exercisers? (Q: I eat red meat- Yes/No)
Who is more likely to exercise: freshmen or seniors? (Q: Do you exercise- Yes/No)
Is marijuana use (Q: Do you use- Yes/No) related to vaping (Q: Do you use- Yes/No)
BOTH variables are categorical. Data are Cross-sectional.
We would analyze these data with chi-square.
Are things different?
Is resistance exercise different than aerobic exercise in terms of calories burned per hour?
Are exercise science majors more likely to exercise than physics majors?
Do marijuana users get lower grades than non-users?
1 variable is categorical. Data are cross-sectional.
We would analyze these data with…
t-test- only when there are 2 groups
Analysis of Variance (ANOVA)- 2 or more groups
Are things changing?
Does muscle mass change after 30 minutes of resistance exercise?
Do exercise science majors get smarter after taking basic stats?
Do people get lazier after they stat using marijuana?
1 variable is categorical. Data are longitudinal.
We would analyze these data with…
Paired t-test- 2 groups
Repeated measures ANOVA- 2 or more groups
Are things related?
Is more resistance training related to more muscle mass?
Is a higher GPA related to higher school spirit?
Is more alcohol use related to less exercise?
Both variables are continuous. Data are cross-sectional.
We would analyze these data with correlation.
Is one thing changing the other?
Does more resistance training lead to more muscle mass?
Does a higher GPA align with a higher first-job starting salary?
Does more alcohol use lead to less exercise?
Both variables are continuous. Data are longitudinal.
We would analyze these data with regression.
To get valid answers…
Statistics allows us to observe characteristics or behaviors of a smaller group (or a snapshot of time) and infer about the larger group.
Sampling is the core of statistics (EX)
Group A: 1200 calories/day x 2 weeks, no exercise
- Loses, on average, 3.2 lbs (Std Dev: 0.5)
Group B: 1500 calories/day x 2 weeks, 20 min swimming/day
- Loses, on average, 3.6 lbs (Std Dev: 0.8)
Group B’s value is certainly larger BUT does this mean that… everyone would lose weight faster by eating more but working out?
To decide this, we need statistics.
Sampling is the core of statistics. What must we know when we sample?
When we sample, we have to understand that the sample may or may not actually represent the population.
Statistics asks: how good is your sample?
- Does it really reflect the whole population? If it’s not that good, your results might just be ‘chance’ (sampling error!)
A statistic is a…
Parameter Estimate
Everyone
Population
Parameter
What you want to know about a population
Some of a population
Sample
A statistic
What you can compute from a sample to estimate a parameter
How good a statistic is depends on…
How good your sample is
Accuracy
How likely your statistic (estimate) actually reflects your parameter
Variances measures
How similar people in a group are (or variable values are). Statistical tests are calculating whether the variance can be EXPLAINED (by another variable that you measured)
Variance- is key to stats
How far is each value from the central tendency (mean)?
In the lower versus upper graph, the individual data points are closer to the central tendency line.
In statistical terms, we’d say that the differences in completion time (y-axis) is better explained by lift time (x-axis, lower) than run time (x-axis, upper)
By the time you were 2 years old, your doctor was predicting how tall you were going to be.
Method 1: predict adult height by multiplying height at 2 yeas old x 2
Method 2: predict adult height by looking at parent height
Are they accurate? Sometimes
What you have vs. what you want
In prior example:
What you have: height as toddler or height of father
What you want: height of kid as an adult
Independent variable:
- What you have.
- It is independent of other variables or the study you are doing.
- It isn’t affected by ‘what comes next’.
Dependent variable:
- What you want.
- It is dependent on the independent variable or the study you are doing.
- You manipulate something and see if this changes.
Independent Variable
- Height as toddler or height of father
Is it qualitative/categorical or is it continuous?
Dependent Variable
- Height of kid as an adult
Is is qualitative/categorical or is it continuous?
This is essential for which statistical test you use.
What test to use when dependent and independent variables are categorical
Chi square test
Fischer’s exact test
Likelihood that endurance exercisers also engage in resistance training
What test to use when independent variable is categorical and dependent variable is continuous
t-test
ANOVA
Comparing endurance exercisers to resistance exercisers on daily protein intake
What test to use when independent variable is continuous and dependent variable is categorical
Regression
Logistic
Likelihood of developing heart disease based on endurance exercise history
What test to use when independent and dependent variable is continuous
Correlation
Regression
Relating daily protein intake to skeletal muscle mass