Search
Create
Log in
Sign up
Log in
Sign up
research methods exam 2
STUDY
Flashcards
Learn
Write
Spell
Test
PLAY
Match
Gravity
Terms in this set (95)
Discrete data
simplest way of naming/ identifying what is being measured, describe the presence or absence of some characteristic or attribute
Categorizing
presence/absence of characteristic, no inherent value to categories, ex.demographics, sex, political affiliation
Continuous level data
variables take on quantity, intensity, or magnitude. Values can differ in degree, amount, or frequency (not yes or no), and these differences can be ordered on a continuum
requirements for discrete data
-Mutually exclusive- meaning no one can be in two categories (completely separate),
-Exhaustive (all possibilities listed)
-Equivalent (equal)
-At least two categories
Ordinal data
data measured by ranking elements in logical order from high to low or low to high. Ex. 1st, 2nd, 3rd there is no gap to tell difference and are only relative together like Houston is better than Austin example
Interval
data measured based on specific numerical scores or values (distance between each interval is even, balanced each side) ex. Strongly agree/ strongly disagree
Ratio data
zero is now an option
Likert-type scale
participants are given a statement and asked to respond indicating the degree to which they agree or disagree with the statement, for each number or quantity there is a meaning (strongly agree, disagree)
Semantic-differential scale
using a stimulus statement, participants are asked to locate the meaning they ascribe to a stimulus, the response scale is anchored by two opposites, sometimes used a 7 point scale (miserable 1 to enjoyable 7)
Validity
(truth/accuracy) extent to where it measures what you want it to measure
Internal validity
accuracy of conclusion drawn from data depends on how the research was designed and the data collected, and there is a true difference in scores of individuals in a scale (Mean girls feeling her boob to predict rain while it is already rating would have a lack of internal validity)
Face validity
the measurement looks and feels as if it will capture the construct we want to measure, weakest type of validity (Guy who buys his gf the most flowers is the better boyfriend)
Content validity
degree to which the items in the measuring device represent the full range of characteristics (Not just how many flowers but everything he does)
Construct validity
ensuring that the core concept that was intended was measured and not something else (obedient experiment with shocking)
Concurrent validity
if participants' scores from the two measurements are highly related, or correlated, both old and new measuring instruments both measuring the same things are related
Predictive
when measurement predicts performance or behavior
Reliability (consistency)
degree of stability trustworthiness, and dependability of a measurement, a reliable measure is one that is consistent and gives very similar results each time it is used.
Internal reliability
degree to which all of the items on a scale invoke the same response from the person responding to questionnaire ex. Same questions in a survey but with different wording
Interrater reliability
more in qualitative, everyone on research team is measuring research in the same way
Test-retest reliability
measures the relationship between scores at two administration of the same test to the same participants at two different times, effects on time intervals, measured between 0-1 ex. How much you learned after class -> 2 weeks later how much retained
Split-half reliability
the test would be split into two parts ex. Different half of the questions in a survey the first time, then the other half the second time, same content different questions
∝ Chronbach's Alpha
how we measure reliability, from a scale of 0-1, .70 or higher is considered reliable, the closer to one the more reliable it is.
Cohen's kappa
measurement of interrater reliability between two raters, closer to zero is chance, closer to 1 the greater the agreement
Maturation
threat to internal validity of research as participants change or mature over the course of observations (their beliefs changed or they became overly familiar with same questions and answer untruthfully)
Attrition
threat to internal validity because participants can no longer be a part of the study because they have dropped out
External validity threat
the ways researchers find and select their samples because the threat weakens the generalizability of the findings (convenience sampling)
Ecological validity threat
degree to which participants are like the ones the researcher is really interested in and the extent to which the setting is natural (telling people to flirt in a room/lab versus observing at a club)
Manipulation
intentionally changing the independent variable to see the affect on the dependent variable
Random assignment of participants
does not matter which sampling is used to select participants for either the control group or experiment group (50 swearing, 50 none)
Main effect
the affect of one independent variable on its own in a factorial experiment
Interaction effect
how independent variables can combine to affect the dependent variable
Control group
participants who get no treatment or a baseline treatment
Treatment group
the group of participants who receive stimulus, or anything that the researcher is interested in studying
Manipulation check
that the participant did in fact regard the independent variable in the various ways that the researcher intended (Does it work?)
Experimental design
manipulation of independent variable and random assignment of participants, purpose is to determine causation
Quasi-experimental design
manipulation of independent variable but NO random assignment, goal is to minimize limitations from classic, natural occurring independent variable (sex), self-select treatment (unequal)
Descriptive design
no manipulation and no random assignment, no temporal order, the independent or dependent isn't first (chicken or the egg)
Posttest research
measurement during or after the participants' exposure to the independent variable only once (Experimental, Quasi
Pretest-posttest research
measurement before and after exposure to the independent variable and measurement of dependent variable (Experimental, Quasi
Longitudinal research
multiple long-term measurements of the dependent variable
Factorial research
introduces 2nd independent variable to demonstrate a complex-causal affect
Field experiments research
experiments that are done in a natural environment, lack of degree of control
Experimental design- Strength
best opportunity to argue causation
experimental design- weakness
Not always natural/realistic, some variables cannot be manipulated
Quasi-experimental- Strength
more natural/realistic can sometimes argue causation
quasi experimental-Weakness
no random assignment may mean lack of equivalence
Descriptive-strength
can't manipulate independent variable, ethically/realistically, allows us to collect info about those contexts
descriptive-weakness
cannot argue causation
Descriptive statistics
numbers that summarize essential and basic information about the dataset as a whole ex. Mean, median, range, # of cases/participants
Inferential statistics
statistical tests that provide information about the relationships between or among variables, used to draw conclusions about a population based on a sample ex. Differences between groups, relationship between variables
Normal Curve
bell curve, mean, median, and mode would be same value, distributed at peak
Positively Skewed
a distribution in which there are very few scores on the right side of the distribution, tail of bell pointing to right, mean is higher than median and mode, most scores below mean
Negatively Skewed
distribution in which few scores are on the left side, long tail pointing to left, most scores higher than mean
Number of Cases
data points, capitalized N, population (the higher the # the more reliable)
Frequency
number of times a particular value or variable occurs
Percentage
a comparison between the base, which can be any number, and a second number that is compared to the base
Central tendency
the primary summary form for data, acts to represent summary of scores of one variable (mean, median, mode are all measures of central tendency)
Mean
most common central tendency, the average, adding up all variables dividing by number of cases
Median
the middle of all the scores on one variable
Mode
the score that appears the most often in a data set
Range
the difference between the highest and lowest score
Standard deviation
standard calculation and representation of the variability of a dataset, how far apart the scores are from one another
Significance level
a criterion for accepting or rejecting hypothesis and is based on probability (how much error researcher find acceptable)
Probability level
significance level, established for each statistical test prior to computing the statistical test (less than .05 is acceptable)
Type 1 error
when the null hypothesis is rejected even when it is true
Type 2 error
occurs in opposite, researchers fail to reject, or they accept, the null hypothesis and reject the alternative hypothesis, reject alternate even when it is true.
Degrees of freedom
the way in which the scientific tradition accounts for variation due to error (df), specifies how many values vary within a statistical test (df=number of categories - 1)
Observed frequency
number of times the category actually appears, literal
Expected frequency
set by researcher, number of times the category was expected to appear/occur, typically equal among all categories
Contingency table
2 nominal variables (categorical in nature ex. Gender), two variables in relationship to each other, (number of rows - 1) X (number of columns -1)
Between group variance
the differences or change between categories or groups of participants (Blacks vs Whites)
Within group variance
difference or change within ONE category or group of participants (differences among Whites or differences among Blacks)
Chi-square (one way)
the statistical test for determining if differences in how the cases are distributed across the categories of one categorical, or nominal, variable are significant, tests difference between observed and expected frequency
Chi-square (two way)
2 nominal variables (categorical in nature ex. Gender), two variables in relationship to each other
t test
used to test hypotheses that expect to find differences between two groupings of the independent variable on a continuous-level dependent variable, looks at one discrete level independent variable with two categories and one continuous level dependent variable.
Two tailed (t test)
non-directional
One tailed( t test)
directional hypothesis
Independent sample ( t test)
shouldn't know each other, not dependent on one another, set to compare two sets of scores (post test only design)
Paired samples ( t test)
compares two paired/matched scores (pretest/protest, time ½)
ANOVA
analysis of variance (F), more than two groups of independent variable on the dependent variable
One way (ANOVA)
tests for significant differences in the dependent variable based on categorical differences on one independent variable (only one independent variable tested, at least two categories of IV
Two way (ANOVA)
two categorical independent variables and one continuous level dependent variable
Factorial (ANOVA)
not limited to a two-way design, used with multiple factorial designs with three or even 4 dependent variables, at least two levels of two categorical dependent variables
Curvilinear relationship
representing a u shaped curve
Linear relationship
one unit change in one variable is associated with a constant change in the other variable
Beta coefficient (beta weight)
unit of standardized scores in regression, indicates difference in a dependent variable associated with an independent variable, estimates the magnitude of change
Regression line
line drawn through data points on a scattergram that best summarizes the relationship between dependent and independent variables, best to minimize distance from points to line
Coefficient of determination
r^2, the percentage of variance two variables have in common
Spurious correlation
relationship between two variables in which a third variable, either known or unknown, influences the other variables
Pearson product-moment correlation coefficient
(r) statistical test that examines the linear relationship between two continuous level variables, allows us to determine if both scores increase or decrease, allows us to determine if scores have no relationship, allows at 2 levels continuous variables
Correlation
r, statistical test that examines the linear relationship between two continuous level variables
Coefficient
shows direction of relationship, the closer to one the more perfect the relationship, the variables are almost identical in the direction they move, the scale goes from negative 1 to positive 1, 0 means there is no relationship
Regression
a set of statistical techniques that predict some variable by knowing others, allows researchers to address many types of research questions and hypotheses and different types of data, looking for a relationship between two variables
Linear regression
R, a form of regression in which the values of the dependent, or criterion, variable are attributed to one independent variable, or predictor variable. The line of best fit is used to determine relationship
Multiple regression
allows a researcher to test for a significant relationship between the dependent variable and multiple independent variables separately and as a group, multiple correlation coefficient is used as an index of magnitude of the relationship among the variables, the significance level has to be .05 or less or use beta weights
;