cas 301 chapter 12

Restriction of Range?

-A problem when scores on a variable are limited to a small subset of their possible values -if you have a limited range, you will not get a very large correlation or it might look like they are not related when they are

Curvilinear Relationship?

-Pearson r is designed to detect only linear relationship -If curvilinear, can appear that there is no association even when there is -Scatterplot will help to visualize -other type of correlation can be used to determine the strength

Measures of variability?

-The spread/dispersion of scores about the mean -Interval or ratio scores only -Standard Deviation -Range

Correlations (Pearson r)- What they are?

-describe the strength and direction of the relationship between 2 variables (you can use 1 to predict the other) -association between 2 variables -usually interested in a predictor variable and a criterion (outcome)

sensitivity?

-high ability to detect difference -Is there variability? To show individual difference.... When variability is low there is a restriction of range... Bad for correlations (won't find association) -ceiling and floor effect

Two ways we need/use statistics?

1. Descriptive statistics= To describe the sample data 2. Inferential statistics= To make inferences and draw conclusions about a population based on sample data

correlations-Important considerations?

1. Restriction of Range 2. Curvilinear Relationship 3. Sensitivity

the normal curve?

50% fall above average and 50% fall below average- perfectly symmetrically

Mediating variables (and how they differ from confounding variables)?

A variable that helps explain the relationship between 2 other variables ( a mechanism) What comes between/explains A leads to B, leads to C

Statistical control?

A way of taking into account or "controlling" for, third variables (potential confounds) -researchers need to measure the variables -use statistics to remove its contribution/effects

Frequency distributions (know different types and how to read them)?

Bar graph, pie chart, histogram, frequency polygon

correlations-Direction and strength?

Direction= positive or negative Strength= how large is the absolute value of the number

Basics of structural equation modeling, SEM (what it is)?

Expected pattern of relationships among a set of variables

correlations- What types of variables/scales can be used?

Interval or ratio; ranges from -1 to 0 to 1

multiple regression?

MR= unique association between multiple predictors and a single criterion variable (shared variance is removed)

Measures of central tendency?

Mean (average) Median (middlemost score) Mode (most frequent score)

Basic measurement scales?

Nominal, ordinal, interval ,ratio

Regression equation?

RE= allows prediction of one behavior/score (criterion) when the score on the other (predictor) variable is known. Ex: Predicting the score on Exam 3 (criterion-Y) from the number house spent studying (predictor-X)

Standard Deviation?

The average deviation from the mean Most common used variability SD score large= when most ppl have scores far away from mean SD score small= when most ppl have scores close to the mean

range?

The difference between the highest and lowest score in a distribution

Partial correlation?

The part of the whole correlation that doesn't include the 3rd variable Provides a way of statistically controlling 3rd variables. Removes any overlap (possible contribution of 3rd variables)