Search
Create
Log in
Sign up
Log in
Sign up
IRM week 2
STUDY
Flashcards
Learn
Write
Spell
Test
PLAY
Match
Gravity
Characteristics of experimental research
Terms in this set (43)
Science
Assumes everything has a cause - Study of the relationship between variables - seeks to identify casual relationships, and thus to control phenomena
Independent variable
condition or event manipulated by the experimenter
Dependent variable
aspect of behavior thought to be affected by the independent variable
Extraneous variable/s
factors besides independent variables that might affect the dependent variables, and need to be controlled - can be controlled so that they are equally likely to influence performance in both conditions/groups e.g., age, gender, previous experience with stats
Confounding variables
linked in a way that make it difficult to sort out their effects - usually when extraneous variables have not been controlled e.g., comparing performance in statistics in Stage I psychology students and maths students - it might be reasonable to assume that more maths students did statistics in high school than did psychology students
Random assignment
when subjects have an equal chance of being assigned to any group or condition in the study
Third variable problem
an unobserved variable that affects both of the observed variables
Controlling extraneous variables so they don't become confounding variables (x6)
Randomization - Holding variables constant - Matching variables - Building variables into experimental design - Within-subjects design - Analyzing covariance
Observing variables
involves recording information about variables, but not manipulating anything e.g., recording how many students had coffee before the lecture; recording how many students enjoyed the lecture
Manipulating variables
allows us to establish more certainty about a cause and effect relationship e.g., giving half the students in the class coffee before the lecture; recording how many students enjoyed the lecture
Correlation
examines the relationship between two variables
Correlation used in what studies?
Naturalistic, case, archival
Correlation useful when
examining whether a relationship exists e.g. pilot study. Examining variables that cannot be manipulated. Examining variables that it would be unethical to manipulate
Correlation coefficients
A statistic that expresses the strength and direction (positive or negative) of the relationship between two variables.
Experimental manipulation
can establish a cause and effect relationship between two variables - experimenters manipulate the independent variable to see whether the dependent variable changes systematically e.g. within subjects or between subjects
Within subjects
Each subject undergoes both experimental condition and control condition (subject as own control)
Between subjects
The experimental group consists of participants who receive special treatment. The control group consists of similar subjects who do not receive treatment given to experimental group
Constructs
Clearly defining. an objective measure of both IV's and DV's e.g. amount of sleep and mood. Internal attributes or characteristics that cannot be directly observed but are useful for describing and explaining behavior. If we want to measure a concept (e.g. personality), we need to define it clearly beforehand
Measurement
the assignment of numbers or categories to objects or events
Operationalisation
when we measure a variable we define beforehand exactly how the variable is going to be measured
Validity
is the extent to which a test measures what it is designed to measure
Face validity
does it look like we are measuring the right thing
Construct validity
Does your test actually measure the concept being studied?
Predictive validity
how well the test predicts future validity
Convergent validity
how well the test results fit with other tests measuring the same construct
Internal validity
the change in the DV occurs because we change the IV
External validity
extent to which the result applies in the real world
Reliability
the extent to which the test provides a reliable (repeatable) measure
Test-retest reliability
if the same person does the test twice, are the results similar both times?
Split-half reliability
Does the score from one half of the test give a similar result to the score from the other half?
Measurement error
any instrument will overestimate as much as it underestimates
True score
the score after infinite tests
Obtained score
the true score ± measurement error
4 levels of measurement (lowest specificity to highest)
Nominal, Ordinal, Interval, Ratio
Nominal measurement
measurement that deals with relations, classes, frequencies:
e.g., male or female. must be analyzed using nonparametric techniques, which assume no particular underlying distribution shape. cannot be added, subtracted, multiplied or divided
Ordinal measurement
when numbers reflect rank but not the distance between scores: e.g., 1: strongly agree; 2: agree; 3: neutral; 4: disagree; 5: strongly disagree. cannot be added, subtracted, multiplied or divided
Interval measurement
measurement is when numbers reflect rank and distance between scores. continuous. Zero is arbitrary (does not mean the absence of a trait). e.g., temperature. can be added or subtracted, but it is not meaningful to multiply or divide them
Ratio measurement
when numbers reflect rank and distance between scores.
continuous. Zero is absolute. e.g., income, height, weight. rare in psychology.
Non-parametric
do not assume the data or population have any characteristic structure or parameters e.g. suitable for examining the order in which runners complete a race
Parametric
Statistical methods used when the distribution of the data is known. Requires prior knowledge of the nature of the data under examination; data must fall under a normal (Gaussian) distribution and be measured on an interval or a ratio scale
Discrete measurement
restricted to whole numbers e.g. Counting
Continuous measurement
infinite possible values e.g. measuring
Increase of specificity
from nominal to ratio does not imply that continuous measures are better than less discrete or categorical measures. No level of measurement is better or worse than another: they are simply different. The research design, through operational definitions, determines the scale of measurement, the kinds of data analysis that can be used, and therefore the scope of possible conclusions.
;