Use LEFT and RIGHT arrow keys to navigate between flashcards;
Use UP and DOWN arrow keys to flip the card;
H to show hint;
A reads text to speech;
100 Cards in this Set
- Front
- Back
nominal categories
|
yes/no
apples oranges not numeral |
|
ordinal
|
1st 2nd 3rd
oredered levels |
|
interval
|
intervals between levels are equal
no abs 0 |
|
ratio
|
has a 0
|
|
three ways of describing results
|
1. comparing group percentage
2, correlating individual scores 3.comparing group means interval and ration data |
|
comparing group percentage
|
used when data is nominal
(comparing males and females .. do you like to travel?) |
|
correlating (co-relating)
|
individual scores- r of individual on 2 distinct score
(where one sits in class and their grade... ranges from 1 to -1 |
|
graphing frequency distribution
|
pie. bar, freq polygons, histograms
|
|
pie
|
for nominal data shown in the percent form
|
|
bar graph
|
use separate and distinct
|
|
histograms
|
values are continuos ie age blood pressure
|
|
frequency polygons
|
ratio interval.. freq distribution is shown with a line
|
|
mean
|
score that divides the group in half
|
|
mode
|
most frequent score
|
|
variability
|
the amount of spread in the distribution of scores
|
|
standard deviation
|
(s) indicates the avg deviation from the mean, small where scores are clustered around the mean
|
|
variance
|
(s^2) square root of standard deviation
|
|
correlation coefficients
|
describe the strength of relationships
|
|
strength of relationship
|
abs value of r
|
|
pearson r sorrelation coefficient
|
need two data points from each P, interval/ ratio data
|
|
direction of relationship
|
+ if one variable increases and the other does too
- if one increases and the other doesnt |
|
which is has the stronger relationship??
-.66 or +.56 |
-.66
|
|
restriction of range
|
need to get as wide a range as possible for each variable to determine the magnitude/ strength of r. If Ps are too homogenous then you cannot see the true strength
|
|
curvilinear relationship
|
cannot be seen with r it is designed to look only at linear relationships btw two variables
|
|
effect size
|
strength of association btw variables
|
|
indicator of effect size
|
pearson r coefficient
|
|
advantage of effect size
|
consistent across all types of studies
|
|
effect size ranges..
|
0 to 1 and is generally reported as small med and large
|
|
transforming effect size to a percentage (percent of shared variance btw the two variables)
|
square value
ex. r=.5 becomes .25 which means between the two variables 25% is shared leaving 75% still unaccounted for |
|
statical significance
|
one wants to know if the results will hold up if the experiment is repeated several times (inferential statistics)
|
|
inferential statistics
|
infers weather the results will "hold up"... makes researchers focus on internal validity
|
|
regression equations
|
calculations used to predict a persons score on one variablr when that persons score on another variable is already known
ex college uses act |
|
regression equation form
|
y= a +bX
|
|
y=
|
score you wanna predict
|
|
x=
|
score that is known
|
|
a=
|
constant
|
|
multiple correlation
|
combines a number of predictor variables to increase the accuracy of prediction
|
|
multiple correlation equation
|
y=a+b(sub one) x (sub one) + b (sub 2) X (sub two) ........
|
|
each variable doesnt equally predict in ...
|
multiple correlation equation
|
|
partial correlation and 3rd varable problem (often seen in no experimental research)
|
provides a way of statically controlling extraneous third variables
|
|
structural models
|
allow researcher to test how obtained data fit in a theoretical structural model
|
|
structural models
|
expected pattern of relationships among a set of variables
path analysis |
|
Inferential statistics are necessary because..
|
the results of a given study are based on a single participant pool (sample) and the researcher really wants to know if the results would be true for the whole population
|
|
inferential statistics
|
Allows researchers to make inferences about the true difference in the population on the basis of the sample data.
|
|
inferential statistics
|
Gives the probability that the difference between means reflects random error rather than a real difference
|
|
null hypothesis
|
pop means are equal, the observed diff is due to random Error NOT THE IV
|
|
research hypothesis
|
THERE WAS AN EFFECT OF THE IV
|
|
statical significance
|
how strongly we can infer that that the research hypothesis is correct..( by doing this we reject the null hypotheses)
|
|
you can never PROVE the hypothesis but you can...
|
REJECT the null hypothesis
|
|
statical significance
|
high sta sig means a very low chance that the difference or results are due to chance or error
|
|
probability
|
the likelihood of the there being a diff in the sample means when there is no diff in the pop means.. so if random error is operating at 1% then that means a 99% chance
|
|
sample size
|
as you increase sample size random error influence drops
|
|
increase in sample size
|
so as your sample size increase you are more confident that the outcome you find in your experiment is actually diff than the null hypothesis (the means are equal in the pop) and you can reject the null hypothesis and accept the research hypothesis. you cannot say you have proven the research hypothesis
|
|
t test is used to..
|
reject the null hypothesis
|
|
the t test is used with experiments with
|
2 mean (simple experiments)
|
|
with the t test you look at the ..
|
mean difference between groups.. the null hyp expects the to be 0
|
|
the value of t increases as..
|
the difference between the sample means increase
|
|
t=
|
group difference/
within group variability t is a ratio of two aspects of the data, the difference between the group means and the variability within groups |
|
within group variability is measured by
|
variance (square root of Standard deviation) which tell us how far scores deviate from the mean
|
|
within group variance becomes lower when..
|
you increase sample size
|
|
you want within group error to be small so..
|
t can be bigger
|
|
you can make t bigger by...
|
having a larger sample size
|
|
you need to know two things to know if the t test is significant
|
deggree of freedom and
one-tailed vs. two tailed |
|
degree of freedom
|
(df) total number of participants minus the number of groups (ex with 2 groups each with 10 par, its 10+10-2=18)
|
|
degree of freedom is useed to..
|
get the critical value by selecting it from a table by using the degree of freedom
|
|
also in the table you must choose either one tailed or two tailed to get the critical value
|
one tailed two tailed
|
|
the f test is used when
|
used when > 2 levels of one of two types of variance: systementic and error.
|
|
systematic variance
|
deviations of the group means from the grand mean ( which is the mean score of all individuals in ALL groups)
|
|
error variance
|
the deviations of individual individuals scores in each group from their respective group means (“within group variance”)
|
|
statical significance( alpha)
|
chosen significance level (alpha level) indicates how confident you wish to be when making the decision if your sample results are reliable , and do reflect the population. in psychology we tend to use .05 .01 or .001 alpha keveks (95%, 99% 99.9 %)
|
|
knowing t or f are useful for
|
knowing the effect size of a correlation coefficent
|
|
type I error
|
Made when the null hypothesis is rejected but the null hypothesis is actually true... means saying there is a diff due to the IV when there isnt
|
|
type II error
|
the null hypothesis is accepted when the research hypothesis is true.. means saying there is not a diff due to the IV when there really is
|
|
Researchers traditionally have used...
|
either a .05 or a .01 significance level in the decision to reject the null hypothesis(alpha = .05 or .01)
|
|
Researchers generally believe that the consequences of making a Type _____ error are more serious
|
I,
for jurors, incarcerating an innocent person is worse (I) than letting a guilty person walk (II) |
|
power analysis
|
statical test that determines the optimal sample size based on probability of correctly rejecting the null hypo. in other words what is the optimal number if subjects need in the experiment considering the effect size and alpha level
|
|
Smaller effect sizes require....
|
larger samples to be significant
|
|
Researchers usually use a power between...
|
.70 and .90
|
|
Scientists attach ____ importance to results of a single study
|
little
|
|
scientists look at results of
|
replicated studies
|
|
reporting the significance of the pearson r correlation coefficient
|
is the ‘co relationship” correlation between the two variables stastically significant
|
|
Software Programs
|
SPSS
SAS Minitab Microsoft Excel |
|
steps in analysis
|
Input data
Conduct analysis Interpret output |
|
One Independent Variable– Two Groups Only (control and experiment)
Nominal Scale Data ordinal interval or ratio |
Nominal Scale Data- chi square test (
Ordinal Scale Data- mann-whitney U test Interval or Ratio Scale Data- t-test |
|
One Independent Variable– Three or More Groups Ordinal Scale Data
nominal ordinal interval |
Nominal Scale Data- chi square test
Ordinal Scale Data- kruskal- wallace H test Interval or Ratio Scale Data- ANOVA |
|
Two or More Independent Variables
nominal ordinal interval or ratio |
Nominal Scale Data - chi-square
Ordinal Scale Data – no appropriate test is available Interval or Ratio Scale Data – two-way analysis of variance (ANOVA) |
|
generalization to other participants of research participants
|
college students can be too restrictive to many studies
volunteers can have prior knowledge and seek approval more gender cultural ...ok to use theses groups if the same result would be found in a different population |
|
cultural consideration
|
today college students are more culturally diverse
|
|
important to be aware that the ways in which the operational definition of the construct are grounded in..
|
cultural considerations
|
|
experimenter's influence need to be kept..
|
constant throughout the experiment (equal across groups)
ex. personality, gender, experience participants do better when tested by the opposite gender |
|
pretests and generalization
|
not like the real world
but allows researcher to assess mortality effect use soloman 4 group design |
|
mundane realism
|
weather the experiment bears similarity to events that occur in the real world refers to the superficial physical characteristics of the situation and usually has little external validity (bo bo doll)
|
|
experimental realism
|
wether the experiment has impact on the participants, involves them psychologically and makes them take the experiment seriously and makes them behave normally while under similar constructs (standford study... prison)
|
|
mutual benefits of lab and field findings
|
are complimentary and allow greater mundane and experimental realism and greater control of the confounding variables
|
|
importance of replications:
exact replication |
an attempt to replicate precisely the procedures of a study to see wether the same results are obtained
|
|
conceptual replications
|
the use of differential procedures to replicate a research finding
(just has the same concept) |
|
literature reviews :
|
summarize , tell what findings are strongly supported and which ones are weakly supported, points where research is inconsistent or lacking, discuss future directions for research
|
|
meta analysis
|
method for determining the reliability of a finding by examining the results from many different studies
researcher combines results from diff studies using diff measures of a construct important in allowing effect size to be measured when having the effect size of many studies |
|
goal of psychology
|
to promote human welfare
|
|
impact of psychology
|
health
law and criminal justice education work environments |