Office hours : Tuesday 3:00 om - 4:00 pm
labs = every week (not this week!) hump 219
Ta list = in course outline
Main library has SPSS supposedly at queen’s learning common
Follow what we are supposed to read on syllabus (has schedule)
hand in assignments at beginning of class AND email it to [email protected]
for assignments: ex: Friday 12th, assignment is posted, due on tuesday 23rd
First exam = Thursday evening
read ch. 1 and 2 by thursday
validity of research depend on (go see definitions on ppt)
1- statistical conclusion validity**this course**= associtation, is there a relationship between two
variables? not necessarily causal. Depends on minimizing Type I and Type 2 errors. much error
= little research for large sample/subject, little error = lots of research on small sample/subject.
2- internal validity**psych 203 (design of the study). Is there a CAUSAL relationship between
variables. They are related, is it causal or is there other variables that influence it? Depends on
control of confounding factors and invoke principle of ceteris paribus (with other things the same
aka changes only with independant variable)
3- construct validity (interpretation of variables, their meanings). Not only causal relationship
what are the case and effect constructs (higher-order constructs). Is your scale legitimately
measuring what it says. nb on a scale comes to mean higher-order construct, which could be
depression. Depends on the meaning associated with choosing operationalizations of the
dependant and independant variables. Must be able to JUSTIFY CHOICES
4 - external validity. (issues of generalizability, can animal research be generalized to
humans…) generalizable across populations, settings and times. Depends on how variables are
selected (participants, settings, times…). Promoted by RANDOM SAMPLING.
must be a particular population of our interest
size can vary greatly can refer to almost anyway (trees, people, businesses…)
Problems : can be very large and thus very difficult to measure or observe. Therefore, we draw
a smaller population/ representative population (rather than all adolescents in the world, look at
selected set of individuals usually inteded to represent the population of the research study.
For generalizability, bigger sample = more accurate representation, but if sample is smaller and
Look at picture in lecture notes online.
parameter: value that describes a population
statistics : value that describes a sample
Research steps: 1- question parameter, 2- get info about stats, 3- generalize back to parameter
datum/ raw score = 1 observation or measurment
data or data set = collection of observations or measurments.
descriptive stats : summarize, organize and simplyfy data
inferential stats : techniques that allow to use samples and make generalizations about the pop
(ex: t-tests, chi-square, ANOVA)
samples are generally representative but is not expected to be perfectly acurate. **
important **sampling error is the amount of error existing between a sample statistics and the
corresponding population parameter.
How do we know when we have a reasonable sample? We can estimate how good the sample
we have is.
Step 1 : experiment/collect data
step 2 : descriptive stas (summarize)
step 3 : inferential stats (interpret)
ex : there is a 3 point difference between samples. explanations : could be random
sampling error OR there could be one teaching method that is better.