Chapter 6 – Methodological Control in Experimental Research
-Two general types of experimental design…
(A) Between-Subjects Design: Groups of subjects contribute data to different levels of the
independent variable I.V.
-A subject receives either level A or level B but not both; the comparison of conditions A and B will be a
contrast between two groups of level.
Example: Effects of Alcohol (I.V.)
Level A – 0 oz
Level B – 2 oz
Subjects are assigned randomly to one of the group.
-An independent variable I.V. (memory test) with subject variable (gender, age) must be between-
-The only time a subject variable won’t be a between-subject variable is when behaviors occurring at
two different ages are being compared, and the same persons are studied at two different times in their
lives; also marital status (subject variable) and same person is studied before and after a marriage or a
- Subjects enters the study fresh and naïve with respect to the hypotheses to be tested
- They don’t bring along practice effects.
- Large numbers of people may need to be recruited, tested, and debriefed during the course of the
experiment; more groups require more subjects thus researcher invests more energy.
Example: 6 levels will require 20 subjects in each level
-Between group differences could be due to the I.V but may also be due to individual differences; each
person differs from one another (confounding var. family background, age, traits, religion, etc.)
Solution – to prevent such case, you can random assign subjects to groups so that we have equivalent
groups – all groups are equal to each other in every important way except for the levels of I.V.
-There are two techniques to create equivalent groups in between-subject design:
1. Random Assignment (By chance)
- A method for placing participants, once selected for the study, into the different groups.
-The purpose of random assignment is to individual difference variable among subjects and spread them
out evenly among groups so that you do not have a disproportionate number of individual difference
variable in one group compared to the others.
-The more people in a study involved (large sample size) the better the results of random assignment.
Example: A B C D (levels of I.V.) 15 subjects per condition = total of 60 subjects
-Different experimenter (confounding var.) or all of the level A subjects were picked first then level B’s
-To complete the assignment of participants to conditions to ensure an equal number of subjects per
group, researcher can use blocked randomization.
-Ensures that each condition has a subject randomly assigned to it before any condition is repeated a
second time. Each block contains all conditions in a randomized order.
Example: Exhaust all conditions in one step, Block 1 contains only A,B,C,D once that is done, you do
another Block 2 containing conditions A,B,C,D;
-Can be done by hand (in a hat), using a table of random numbers or computer program.
-The process of random assignment is normally associated with laboratory research, but can also be
possible in field research.
2. Matching – Subjects are paired together on some subject variable such as their characteristics (level
of anxiety) and then distributed randomly to the different groups in the experiment.
-In the memory study, “anxiety level” would be called a matching variable.
-It’s used to create equivalent group when the sample size is small.
-when the number of subjects is small
-variables should be reasonable to the experiment; matching variable is correlated with the dependent
Example: In memory study, anxiety is correlated with recall or math performances match on students
GPA, or IQ.
-Drawback of Matching is its inconvenience (have to bring in subjects twice, once to test the subjects on
their IQ to match, and then on the actual experiment test.
-Also, matching concerns the number of matching variables to use; in a memory should groups be
matched for anxiety level or intelligence level? With many matching variables it becomes difficult to
accomplish matching and results in eliminating participants because close matches cannot be made.
-Thus, experimenter prefers to make effort to recruit enough volunteers to use random assignment
even if there might be some extraneous var. which correlates with the dependent variable. They simply
believe that in a large sample size, these extraneous factors will be distributed evenly.
(B) Within-Subjects Design: (repeated measures) same group of subjects contribute data to all the
levels of the independent variable I.V.
-Same subjects are tested repeatedly in the different level; receives both A and B levels of the I.V.
Example: Alcohol consumption (I.V.), Alcohol level: Level A - 0oz and Level B -1oz
Subject 1 is tested on both A and B conditions
Subject 2 is tested on both A and B conditions
-If each participant receives both levels A and B, both levels exist within each individual. Advantages:
-Fewer subjects are needed in a study; when comparing two conditions and want to test 20 people for 1
condition/level, you’ll need 40 subj. for a between-subjects study but only 20 subj. for a within-subjects
-Eliminates equivalent group problem; gets rid of the individual difference variables
-This design is more sensitive to finding significant differences between groups.
-Pick on clues and have the practice effect
-The repeated exposure could influence the D.V. This is called an order effect: can either get better at it
(practicing over and over again) or can get tired of doing it repeatedly (long process leads to fatigue)
-The performance therefore can improve with practice or decline from trial to trial, these two effects
can be referred to as progressive effects because performance changes steadily (progressively) from
trial to trial.
-Also some sequences might produce effects different from those of other sequences, called carryover
Example: p.194 – the effects of noise on a problem solving task – anagram: level 1 unpredictable noise
UPN and level 2 predictable noise PN:
Subjects running level 1 UPN first then level 2 PN, may do poorly in level 1 and then this poor
performance might discourage them and carry over to condition level 2 PN “here we go again” and not
try as hard to solve a problem, however, if subjects are running level 2 PN first, the subjects may do
better at their performance, and some of the confidence might carry over to the level 2 of the study –
UPN “here we go again but I’ve handled the noise before, can handle it again”. Thus performance in
condition UPN might be much worse in the sequence UPN-PN than in the sequence PN-UPN. Because of
this effect experimenters usually decide to use a between-subject