PSYC30013 Lecture Notes - Lecture 3: Effect Size, Confidence Interval, Consistent Estimator
10 May 2018
School
Department
Course
Professor
Lecture 3
- Use the term relationship as a generic word to describe the focus of virtually all
research investigations
- The particular type of relationship we wish to investigate in a research study is
distinguished by using terms such as association, prediction, or difference
- An association between two constructs is the most simple kind of relationship
considered - e.g. is depression associated with stressful life events?
- The notion of association implies a systematic co-occurrence of scores on two
construct measures
- How we investigate an association, and undertake an analysis, depends on how we
measure the two constructs
- Continuous measurements → correlations
- Categorical measurements → contingency tables
- The only practical distinction in data we make is between either continuous or
categorical observed scores
- Continuous (e.g. age, summed scores on questionnaires, IQ scores, etc) → no
distinction between integers and real numbers, values always imply some meaningful
ordering, may not be strictly interval in scaling
- Categorical (e.g. religions, football teams, gender, subject grades, etc) → no
distinction between nominal or ordinal scaling, numbers assigned to categories are
arbitrary in meaning
- For associations, we only consider variables that are either both categorical or both
continuous
-
find more resources at oneclass.com
find more resources at oneclass.com
-
- Population, relationship, constructs
- Whenever we undertake a research investigation, we should inspect some kind of
distribution plot of the scores: boxplot, histogram, Q-Q plot (or P-P plot)
- Assists in understand and identifying the general shape of the distribution (symmetric
or asymmetric), if there are obvious outliers, if there might be coding errors (e.g.
entered as 89 rather than 9, etc), if missing responses or scores have been
appropriately specified in the data file
find more resources at oneclass.com
find more resources at oneclass.com
-
- Scatterplot of scores on two variables
-
- Covariance - concurrent or co-occurring variation, high scores on a variable
correspond with high scores on other variable or low scores correspond with low
scores
find more resources at oneclass.com
find more resources at oneclass.com
Document Summary
Use the term relationship as a generic word to describe the focus of virtually all research investigations. The particular type of relationship we wish to investigate in a research study is distinguished by using terms such as association, prediction, or difference. The notion of association implies a systematic co-occurrence of scores on two construct measures. How we investigate an association, and undertake an analysis, depends on how we measure the two constructs. The only practical distinction in data we make is between either continuous or categorical observed scores. Continuous (e. g. age, summed scores on questionnaires, iq scores, etc) no distinction between integers and real numbers, values always imply some meaningful ordering, may not be strictly interval in scaling. Categorical (e. g. religions, football teams, gender, subject grades, etc) no distinction between nominal or ordinal scaling, numbers assigned to categories are arbitrary in meaning. For associations, we only consider variables that are either both categorical or both continuous.
