Quant Final Exam

What is an expiremental research design?

• Examines cause and effect relationships with a control group, sample randomization, and manipulation of the IV

• Use when you want to explain something

What is a quasi-experimental design?

• Nonrandomized control trial and examines cause and effect relationships

What is a correlational design?

• Determines whether a relationship exists between two constructs and assesses the strength of that relationship

• Used when you want to predict something

What is a case study design?

• Two groups are compared, those with the condition and those without

• Tries to identify what factors from the past might have contributed to the condition

What is a single subject design?

• Studies the behavior of a small number of participant where each participant is their own control, then data is collected repeatedly across phases

What is a cross-sectional design?

• Gathers data at one point in time

What is internal validity?

• Validity = accuracy → measuring what it is supposed to measure

• Ensures results are due to tested factors and not other influences

What is external validity?

• The ability to generalize findings to a broader population

• Quasi experimental has more external validity than experimental

• Descriptive stats and correlational designs have stronger external validity → we are more confident that they will generalize to the population (less confident in the internal validity that our IV is causing the change)

What is the difference between the IV and DV?

• IV: causal variable, x axis

• DV: the outcome variable, y axis

When should you use a one sample t test?

• Continuous outcome (DV), 1 exposure group

• Comparing sample group to the normative population

When should you use a independent t test?

• Continuous outcome (DV), 2 groups

• When you want to compare means

When should you use paired/dependent t tests?

• Continuous outcome

• Pre/post test → same group is being tested twice

When should you use a one way ANOVA?

• Continuous outcome

• 3 or more groups

When should you use a pearson correlation?

• Continuous continuous correlation

• One predictor (variable)

• Tells us the strength and direction of the relationship

When should you use linear regression?

• Continuous continuous correlation

• Multiple predictors

When should you use a chi square test?

• Categorical outcome variables

• 1,2, or >2 groups

What are the measures of central tendency?

• Mean: used for non-skewed continuous data, always report SD with mean → most commonly used; interval/ratio

• Median: used when there are outliers → ordinal/ratio/interval

• Mode: used for identifying the most common value for categorical (nominal) data

What do you want to report for pearsons correlation?

• Report r and p value

What is the F statistic in a one way ANOVA?

• The ratio of variance between/within and tells us if a difference exists, but doesn’t tell us where

• F(dfbetween, dfwithin)=____, p=___

What is E2 in a one way ANOVA?

• Enta squared is the effect size used for ANOVA

• >0.06 is medium

• >.14 is large

What is a chi square test?

• Compares the observed frequency with the expected frequency

• Determines how observed values compare to the critical value → if obtained (observed) is > critical value (expected) we reject the null

What is parametric statistics?

• Statistics used for the inference from a sample to a population that assumes certain characteristics about the population and the sample selection

• Typically requires at least 10 cases

What is non-parametric statistics?

• Distribution free statistics that do not require the same strict assumptions as parametric stats

What is nominal and ordinal levels of measurement?

• Nominal: data that can be categorized by qualitative characteristics, but there is no meaningful rank between categories

• Ex: gender, hair color, nationality

• Ordinal: data that can be categorized and ranked in meaningful order

• Ex: sports teams, likert scale, SES, education level, surveys

What is interval and ratio levels of measurement?

• Interval: numerical data that has known equal intervals, without a meaningful zero point

• Ex: temperatures, IQ, credit score, calendar years, time on 12h clock

• Ratio: numerical data that is ordered/ranked with equal distance between points and has a true zero

• Ex: weight, height, age, income, distance, time

What are strategies to maximize power of a research study?

• Increase sample size

• Use higher significance

• Reduce error, choose reliable and valid measures

• Increase effect size

What is internal consistency?

• Reliability that is measured with Cronbach’s alpha

• Do they ask the same thing?

• 0.70 or higher = acceptable

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

• Type I error: false positive (saying something is happening when it’s not) → fail to reject null

• Type II error: false negative (saying something isn’t happening when it actually is) → reject null