PSYB01 - Chapter 12
Understanding Research Results: Description and Correlation
Two reasons for statistics:
1) Statistics are used to describe data
2) Used to make inferences on data
Scales of Measurement: A Review
Variable must have operational definition + Have 2/more levels of variable .
Levels of variables can be described using four scales:
1) Nominal
2) Ordinal
3) Interval
4) Ratio
Scale used determines types of statistics that appropriate when results of study are analyzed .
Level of nominal scale variable: Have no numerical, quantitative properties .
Levels are simply different categories / groups .
Eye colour, hand dominance etc.
Variable with ordinal scale level: Minimal quantitative distinctions .
Rank order level of variable from lowest to highest .
Internal and ration scale variable: More detailed quantitative properties
Internal scale variable: Intervals between level are equal in size .
- Generally have 5/more quantitative levels .
- No absolute zero point, no absence of mood .
Ratio scale variable: Equal interval, have absolute zero points .
- e.g. time, weight, length .
Three ways in describing results:
1) Comparing group percentage:
e.g. 60% of female like to travel, 40% of male do not .
2) Correlate scores of individual on 2 variables:
Do not need distinct groups of subjects, measured on 2 variables with numerical value.
3) Comparing group means:
Compare the mean responses of participants .
.
Frequency Distribution: Indicate # of individual that receive possible score on variable .
e.g. how many students receive score on variable .
Pie chart: Useful when represent nominal scale .
Bar graph: Useful when represent nominal scale .
Frequency polygons: Useful when represent interval/ratio scales .
- Use line to represent frequencies .
Histogram: Uses bars to display frequency distribution for quantitative variable .
Able to indicate any outliers .
Descriptive statistics: Allow to make precise statements about datas .
Two statistics are needed to describe the data .
Used to describe Central tendency: How participants scored overall .
Used to describe Variability: How widely the distribution of score is spread .
Central Tendency: Mean, median, mode . Mean: Appropriate when score are measured on interval or ratio scale .
Median: Appropriate when score are measured on ordinal scale, take accounts of rank order .
Useful with interval and ratio variables .
Mode: Most frequent score, Appropriate when score are measured on nominal scale .
Median/Mode can be better indicator than mean if few unusual score bias the mean .
Variability: Using Range/ Standard deviation, average deviation of scores from mean .
Appropriate when score are measured on interval and ratio scale .
Effect size: Refers to strength of association between variables .
Value range from 0.00 to 1.00 (Small=0.10-0.20 , Medium=0.30, Larger= above 0.40) .
For experiment more than 2 variables/conditions .
r2 : Percentage value represents % of variance in one variable, that account for second variable .
Criterion variable: In predicting future behavior .
Predictor variable: On basis of person’s score on other variable .
Used in Regression equation = Y= a+bX .
Multiple correlation: Used to combine #

More
Less