PSYB01H3 Study Guide - Final Guide: Partial Correlation, Null Character, Social Class
51 views13 pages
For unlimited access to Study Guides, a Grade+ subscription is required.
Dec. 6th, 2010 Lecture
Scales of measurement: A review
No numerical, quantitative properties
Levels represent different categories or groups
2. Ordinal – minimal quantitative distinctions
Order the levels from lowest to highest
3. Interval – quantitative properties
Intervals between levels are equal in size
Can be summarized using means
No absolute zero
4. Ratio – detailed quantitative properties
Can be summarized using mean
Analyzing the Results of Research Investigations—What to do with data?
Three basic ways of describing the results:
1. Comparing Group Percentages
-used for nominal scale variables
Example: Ask boys and girls whether they like school.
-Like or dislike is nominal (categorical variable)
-Ask 100 boys and 100 girls
-Find that 60 boys and 75 girls like school
-What would you report?
-Perform statistical analysis to determine if difference between groups is significant.
*Here, we are thinking about ways to become familiar with data, and to describe it. Comparing the
number of people in each nominal group (in the form of a percent score) is one way of presenting the
information, and of relating our findings to our research question.
2. Correlating Individual Scores
-Obtain pairs of observations from each subject (each individual has two scores; one from each of the
-Ask whether variables go together in a systematic fashion by calculating Pearson r correlation
(Determines strength and direction of relationship)
Another way of describing data is to state how the scores are related to each other (i.e. do the variables,
as operationalized by the scores obtained) vary systematically together.
***What types of data may be used? only ratio and interval data may be used
Notice the variability in the data. The linear relationships in these examples are quite realistic.
Individuals vary in their responses, yet we can still see a pattern of how the variables are related to
3. Comparing Group Means
For Experimental Designs
-Compare the mean (average) response of experimental group with the mean response of control group
-in experiments, remember that we are sampling from a population, and then randomly assigning to our
control and experimental groups (2 groups in most simple design; may have more). When we collect data
(in the form of scores) within each of the groups, the data is pooled within each group and then compared
across groups to see if there are differences in the overall scores. This is done by calculating the mean
score within each group.
*Think about how variability within each group might affect these pooled scores.
♦Want to study the effect that stroking a dog has on resting heartrate
♦How might we design this experiment??
♦20 subjects; have 1 very cuddly dog named Kaija….
assign to experimental and control group. Have them do everything the same except patting the dog
for 20 minutes in the experimental group….(your independent variable is the contact condition: no
contact or dog patting)
Sample Data: (Resting Heart Rate)
Dog Patting Group No-Dog Group
Here are our individual scores…notice how we calculate the average/mean score for each group (we will
look at the formula for calculating the mean score later in the lecture). Basically, we add each score
together (within each condition or group) and then divide by the number of participants within that
Graphing frequency distributions
1. Pie Chart: useful for nominal data
2. Bar Graph: might be visually helpful for nominal data as well
3. Frequency Polygons: Used for interval/ ratio scales.
Each point on the graph represents the number of participants who ended up with a particular score on a
measure. This example shows the number of people who scored each score, as per the group they
belonged to. Visually, we get a sense of how the groups, and the individual scores within each group,
differed from each other
4. Histogram: Depicts scores for comparison
5. Stem and leaf plots : Useful because they give us a sense of the shape of our data, as well as the
actual scores. Plots are constructed based on the range of scores seen.
E.g: Range = 48 - 95. Stem40 and the leaf = different variants of all the scores in the 40’s
It’s always good to graph individual scores for each group to get a visual sense of what data looks like.
Found by adding all the scores and dividing by the number of scores (X=X/N)
(Where X=sum of values in a set of scores; N=number of scores)
* x (bar) stands for mean.
Indicates central tendency with interval or ratio scales
**Graphing our means in comparison can be useful but we have to note the scale used on the y axis!!
Because it may be deceiving to view…
***Outliers have been tossed out. Therefore mostly used for data analysis
2. Median (Mdn)