BBH 310 Exam 4 (FINAL)

What is scientific research and how does it rely on the “ways of knowing” we discussed in class?

• scientific research helps us identifiy facts and understand world around us

• works best when ways of knowing are balanced since we use diff approaches and tools to understand

What are the 3 requirements for demonstrating cause and effect?

• covariation - cause and effect must be associated; cause occurs, effect should also occur

• temporal precedence - cause must come before the behavior

• alternative explanation - rule out confounds or other potential causes

What is the difference between correlation, prediction, and cause-and-effect?

• correlation - association between two variables

• prediction - relationship between variables forcasts an outcome

• cause-and-effect (causation) - relationship where one variable directly produces change in another

When do we use the terms X/Y?

• used to describe relationships

• emphasizes correlation

When do we use the terms predictor/outcome?

• used in forecasting or estimating outcomes

• emphasizes prediction

When do we use the terms IV/DV?

• used in experimental or causal research

• emphasizes cause-and-effect relationships, testing causal effect

How do random sampling and random assignment differ?

• rando sampling - pick ppl from large population so sample reps population; generalizability

• rando assignment - participants go into diff groups in a study (experimental vs. control), internal validity

How do random assignment and random sampling influence validity?

• random sampling = external validity, more rando sample is, more can generalize

• random assignment = internal validity, reduce bias and confounding variables (cause anf effect)

What is an operational definition?

set of procedures used when you measure or manipulate variables; includes what is measured and observed

What is the difference between reliability vs validity?

• reliability - consistency or stability of a measure, yield same results every time

• validity - accuracy of a measure, measures what it’s supposed to

What are the pros of using self report? (3)

• access personal day-to-day info that is otherwise hard to assess

• affortable

• easy to deploy

What are the cons of using self report? (2)

• lacks clear units

• can be biased or inaccurate

What are the pros of using observational/behavioral data? (2)

• see complex behaviors in natural setting

• detailed data

What are the cons of using observational/behavioral data? (3)

• confounds

• no control of participants or active intervention

• unable to find definitive cause and effect

What is central limit theorem?

sample distribution is normal (bell curve), closer to population

How does central limit theorem (CLT) relate to reliability and null hypothesis testing?

• reliability - standard error comes from CLT, CLT justifies why larger samples give more consistent, reliable estimates of population

• null hypothesis testing - test statistics based on sample means, CLT tells us if sampling distribution of mean is normal (bell curve) → allows us to compute z, t-scores and p values

What are confounds?

variables outsides IV that may influence the DV

How do we control for confounds in a true experimental (randomized controlled trial)? (3)

• random assignment

• control group

• blinding to reduce bias

What are basic elements of a basic experiment? (6)

• independent variable

• dependent variable

• experimental group

• control group

• pretest and posttest

• randomization

Why do we use independent variables in a basic experiment?

• introduce controlled changes

• represents the cause we are testing

What type of statistics does central tendency and variability fall under?

descriptive statistics

What is the difference between central tendency and variability?

• central tendency - statistic that tell you what sample as whole or average is like (Ex. mean, median, mode)

• variability - spread of data, how data varies from central tendency measures (Ex. standard dev, range)

What are the measures of central tendency?

mean, median, and mode

What are the measures of variability?

standard deviation and range

What is the difference between a type I vs type II error?

• type 1 error - reject null when actually true, false alarm → see effect isn’t real (false alarm)

• type 2 error - fail to reject null, missed detection → not seeing effect that is real (MISS)

What does p-value mean exactly?

• tells you how likely your results are due to random chance

• lower p value - reject null, effect probably real

What is a t-test ?

tests whether the means of 2 groups are statistically different from one another

What types of variables does t test use?

• independent variable - categorical (labeled) w 2 levels

• dependent variable - continous

  • group variable categorical, outcome variable continous

What does t test reflect exactly?

difference between 2 groups - variability and sample size

What is ANOVA?

• analysis of variance

• tests whether means of more than 2 groups are statistically significant from each other

How does ANOVA differ from t-test?

• ANOVA compares mean of 3 or more groups, t- test only 2 groups

• ANOVA uses F statistic, t -test uses t statistic

• ANOVA control for type 1 error well

higher the numerator in t - test equation means…

higher t value = tells us more accurately whether groups are different

What is a Pearson correlation ?

measures the strength and direction of the linear relationship between two continous variables

What types of variables does Pearson correlation use?

both variables must be continous (measured)

What is regression?

statistical method used to predict the value of one continuous variable (the outcome, Y) from one or more other variables (the predictors, X)

How does regression differ from a Pearson correlation coefficient? (2)

• correlation shows relationship, regression shows prediction and quantifys change

• regression give slope (change)

Why would a researcher use regression over Pearson correlation? (4)

• regressions allows for …

  • prediction

  • amount of change (slope)

  • multiple predictors

  • detailed outputs

Why might a researcher use ANOVA over regression?

when independent variable is categorical and trying to compare means across groups

What is the difference between paired vs. unpaired t-tests?

• unpaired - compares mean between 2 separate groups (exp. vs. control)

• paired - compared mean between the same group or individual (pre test vs post test)

What does Pearson correlation exactly reflect? (3)

• direction - positive or negative

• strength - r range from -1 to 1, value near 0 no relationship

• linearity - how well data fits into a straight line

How are the Results and Discussion section different?

• results = statistic and basic intepretations

• discussion = “what does it all mean?”

What 3 statistics should we report when describing a finding in the Results section?

• p-value

• test statistic - t, F, r

• indicator of confidence - confidence interval or standard error

What is a confidence interval?

a range of values that likely contains the true population value (mean, correlation, regression slope, etc.).

How is a CI similar to a p-value? (3)

• inferential stat

• tells you whether smth statistically significant

• based on sampling distributions

How is CI different than a p-value?

• p value - tells you whether effect is statistically significant

• CI - significance AND effect size AND precision

What is effect size?

a statistic that tells you how large or meaningful a finding is in the real world.

What does effect size tell you?

• How big the effect is (magnitude)

  • large effect size = the difference or relationship is strong and meaningful.

  • small effect size = effect is weak, even if it’s statistically significant.

How are confounds controlled for in quasi-experimental designs? (3)

• use regresion or ANOVA to account for confounds

• match key variables in participants in intervention and comparison groups

• look at trends before/after intervention

  • NO random assignment

How to control for confounds in pre-experimental designs? (3)

• most difficult to control for

  • repeated measures

  • pre and post test

  • homogenous sample

What are the 4 measurment scale types?

• interval scale → interval but no true 0

• ratio scale → real 0

• nominal scale → name/label

• ordinal scale → order/rank

What 2 measurement scales are continous variables?

interval scale and ratio scale

What 2 measurement scales are categorical variables?

nominal scale and ordinal scale

What is an interval scale? What is an example?

• numeric scale where distance between two points matters

• no true 0

• Ex. Celsius

What is a ratio scale? What is an example?

• distance between two points matter

• true 0

• ratios possible

• Ex. heart rate

What is a nominal scale? What is an example?

• groups with no rank order

• Ex. gender, ethinicity

What is an ordinal scale?

• groups with rank order

• Ex. SES

Why do we use dependent variables in a basic experiment?

• observe the effect of the IV

• provides evidence of whether the intervention worked

Why do we use an experimental group in a basic experiment?

shows what happens when the IV is applied

Why do we use a control group for a basic experiment?

• serves as a baseline for comparison

• helps rule out confounds

Why do we use a pretest and posttest for a basic experiment?

• allows us to measure change over time

• establishes differences occurred after IV was introduced

Why do we use randomization (random assignment) for a basic experiment?

• balance unknown and known confounds

• increase internal validity

• strengthen causal inference

Research question of group 1

Do participants that engage in outdoor exercise compared to indoor exercise exhbit lower levels of stress?

Research question of group 2

Do individuals ages 18-25 have fewer depressive symptoms if taking SSRIs and practice yoga?

Research question of group 3

E cigarette smoke exposure and sinus health in college student ages 18-22

Research question of group 4

In college students ages 18-22, how does the presence of Binge Eating Disorder (BED), specifically with foods of low nutritional value affect bone mineral density?

Research question for group 5

Does engaging in 30 minutes of yoga lessen anxiety and stress prior to taking an exam in college students?

Research question for group 6

In PSU student athletes, is burnout more likely to occur in team sport athletes or individual sports?

Research question for group 7

How does daily caffiene compared to energy drinks affect average BP levels in college students ages 18-25 over a 3 month period?

Research question for group 8

How does increasing adherence time to the Mediterranean diet influence cardiovascular risk biomarkers over a 2 year period in older adults, ages 65-80 at elevated risk?

To reject the null hypothesis for Cl…

• 0 is not include

• Both bounds should be positive or negative

To fail to reject the null hypothesis for Cl…

• 0 is included

• one bound positive, one bound negative

What are descriptive statistics that assess variability? (3)

• range

• standard deviation

• variance

What are descriptive statistics that assess central tendency?

• mean

• median

• mode

conformity to powerful person/group

authority

cultural beleifs and information from other not in a position of authority

folk wisdom/peers

empiricism or experience, finding patterns from observation (specific to general)

induction

make specific conclusion based on one premise (general to specific)

deduction

How to differentiate from induction vs deduction?

induction → INcrease specific to general

deduction → DEscend from general to specific

p value is less than 0.05 →

reject the null, results are significant

p values is greater than 0.05 →

fail to reject the null, results are insignificant