types of variables 1 - methods of research

basic research

new knowledge w/ no intention of solving a social or clinical problem. theoretical in nature - theory: general explanation that attempts to explain relationship between cause and effect

applied research

(think about vaccines) solves problems that require immediate attention (clinical research)

field research

everyday setting (home, schools, clinics)

laboratory research

less contrived settings outside the mainstream of daily lives

experimental research (2 requirements)

1. random assignment of participants to one or more conditions (best way to ensure equivalency and eliminate bias)

2. selection and manipulation of conditions

non experimental research

lack measures of comparison ( group or prior measurement) - typically observe behavior without inferring cause and effect

quasi-experimental research

almost experimental but not quite (usually lacks random assignment)

if an experiment does not meet requirements of experimental research it is probably quasi-experimental

if subjects are randomly assigned to groups it is a what experiment

true

if subjects are not assigned randomly to groups but there is a control group or multiple measures it is a what experiment

quasi

if subjects are not assigned randomly to groups and there is not a control group or multiple measures it is a what type of experiment

non-experiment

what is a variable

focus of interest for behavior scientists; concepts that take on different quantitative or qualitative values

independent variables

the controlled and manipulated variable (the presumed cause; conditions that cause changes in behavior)

quasi-experimental research design (independent variable)

cannot control or manipulate the independent variable, because it is fixed

dependent variables

“if x, then z” - presumed effect (consequence, focus of observation, data, behavior that is changed)

bivariate research

1 independent, 1 dependent

multivariate research

more than 1 independent, more than 1 dependent

active variables

can be manipulated by the researcher (ex: types of language tests, treatment, procedures, noise, and other conditions that are readily changed)

assigned variables

measured but not manipulated (human characteristics - age, gender, intelligence, occupation, hearing sensitivity)

continuous variables

take on a range of variables and possess the property of order ( ex: age )

categorical variables

people or objects are assigned to categories based on whether they possess some characteristic or not; no order (ex: young, middle, old)

intervening (extraneous) variable

potential nuisance variables (any variable that affects the dependent variable but is NOT the independent variable) - may be observable but can compromise the study

operational definitions

how you operationally define (explain) your variables - describe activities necessary to measure and manipulate variables - instructions for selecting subjects, measuring behaviors, and carrying out procedures

must be valid and reliable for the population of interest

operational definitions is based on

previous studies or research literature

limits of operational definitions

some concepts are difficult to define - operational definitions may differ across research studies

research data

consequences of observing or otherwise gathering information for study - observations are coded (usually coded in numerical data - according to specific rules is known as measurements)

research data must be

reliable

reliability procedures (blinding)

interobserver reliability

(across people) - ex: took the same scoring sheet and gave it to two different people and looked at their agreement to see if they come up with the same score

intraobserver reliability

(within the same person) ex: gave scoring sheet to one person and had them score it on two separate occasions and comes up with the same score

in interobserver and intraobserver reliability you are looking at

percentage of agreement and point by point agreement

an agreement index of below blank is usually unacceptable

80%

internal validity

degree to which their is a relationship between the independent and dependent variable without the influence of extraneous variables

degree to which the independent variable leads to a change in the dependent variable - when this is weak, we cannot imply that the independent variable produced the effect observed on the dependent variable (may be confounding variables that weaken the internal validity)

problems with what are the most common weaknesses in research studies

internal validity

possible threats to internal validity

ATP, differential selection effects, history effects, maturational effects, statistical regression effects, attrition effects, multiple testing effects, instrumentation effects, additive and interaction effects

ambiguous temporal precedence (ATP) effects

the treatment must occur before a change in the dependent variable is observed - atp effects are present when the direction of the relationship is not clear

differential selection effects

subjects are assigned to experimental and control groups in a way that results in unequal distribution of the subjects’ unique characteristics - can alleviate this by random assignment and matching (how many people have had this and divide evenly among groups)

history effects

research that require observations over long periods of time (longitudinal studies) are vulnerable to this - includes outside effects (extraneous variables) that may influence the dependent variable during the course of the study

can alleviate this by reducing time between pre and post testing and statistical procedures

maturational effects

internal events - changes in physical abilities and mental processes (ex: children’s development of speech and language or motor skills) - issue when maturation is not the independent variable/focus

long or complicated tasks may produce maturational effects like boredom and fatigue

how to control: reduce time for completing observations, add a control group, provide planned rest times or comfortable surroundings

statistical regression effects

occurs when subjects who score very high or very low on a test regress toward the mean on the next administration of the test

how to control: don’t select subjects who have extreme test scores

attrition effects

occurs when there is a loss (dropout or mortality) of participants (often seen in long studies)

can control by adding additional participants to offset the possible loss

multiple testing effects

subjects are tested more than once and become sensitized to the test or improve due to practice

can control by counterbalancing tests and plan an interval of time between tests

instrumentation effects

occurs when there are unwanted variations in instruments used to measure human behaviors - mechanical instruments may change over time and human observers may change over time

can control by calibrating equipment, training, and perceptual anchors (having ways to train people to accurately identify good and bad training measures)

additive and interaction effects

one or more of the many potential threats to internal validity may be present and interact with one another