This site is 100% ad supported. Please add an exception to adblock for this site.

iss 305 exam 2 deck


undefined, object
copy deck
- People who chose their lottery ticket wanted four times as much for it as people who had someone else chose it. this is an example of what?
illusions of control.
what's the gambler's fallacy?
tendency to see links between events in the past and events in the future when the two are really independent
what is the Representativeness heuristic?
probability judgments are often based on how representative certain features of events are to what we know
what does a gee wiz graph create?
an impression of larger vs. smaller differences
Means are sensitive to what scores?
Every score
what does the median consist of and is it sensitive to extreme scores?
medians are made up of the middlemost scores and are fairly insensitive to extreme scores.
when will the mean equal the median?
anytime distribution is symmetric
can the mean equal the median when distribution isn't symmetric?
What is the mode?
the most popular score
why would one choose mean vs. median?
to create an impression that scores are especially high or low.
what is the range?
the difference between the highest and lowest scores.
what is the relation between range and extreme scores?
ranges can be too sensitve to extreme scores
what is the interquartile range?
the difference between the 75th and 25th percentiles.
what is the variance?
mean of squared deviated from the mean
if we let the levels or possible values of Variable A to be the rows and the levels of Variable B be the columns what’s the smallest simplest table possible to establish a relationship?
a 2 x 2 table.
what is required of variables to cause a relationship?
when one variable changes, the other must change (dependence).
what must you show in a 2x2 table in order for there to be a relationship?
either 1. That the row relative frequencies differ within either column Or 2. that the column relative frequencies differ within either row
what information do you need in order to draw a valid conclusion about whether there is or isn’t a relationship between two variables?
The information that would permit a comparison of the appropriate relative frequencies
what would a perfect relationship in a 2x2 table be?
that & D>O and C & B=0 Or the other way around
how rare are perfect relationships?
Almost no relationship in nature (and hence, of scientific interest) are perfect relationships because most things in nature (like lung cancer) have multiple causes
When we say that variables A and B are related what do people assume to be true?
that a causes b (NOT true)
if people get high scores on pretests do they usually repeat their scores on the posttest?
No, hard to believe eh katie?
if people are selected to be in the study because of the extremity of their scores at pretest, we should expect such changes even without any manipulation. why?
Because the chance factors which make a person especially high at pretest are unlikely to make them again especially high at posttest
what is the problem with the "base" number of a percentage being small?
a small change in numbers can cause a big change in the %
what does skewness summarize?
the degree of ASYMMETRY
what must skewness equal for distribution to be symetric
zero, katie, it's zero!!
when does the distribution have a tail in the positive direction?
when skewness is greater than 0
when does distribution have a tail in the negative direction?
when skewness is less than 0
what does the use of a mode as a central tendency statistic assume?
the most common response is the most descriptive of the whole distribution
what does a percentage make a comparison between?
a "base" number and a second number to be compared with that base
what are some problems with using proporions or percents?
having a small base, combining percents when the bases are different, comparisons of increases or decreases in percents, when the base is unclear
Most common response is discrpitive of whole distribution -bimodal-2 humps -trimodal-3 humps
how does the gee wiz graph create its impression
by o Stretching/compressing the axes of a graph (usually Y axis, sometimes X) And/or breaking the Y axis (leave out some or most of the full range of Y axis)
how can one avoid the gee wiz graph mistake?
1.Make sure all axes are clearly labeled 2. Make sure that either there’re no breaks in an axis or that one is fully aware of the break 3. Make sure the range through which the graph varies is a meaningful one
what is a one-dimensional picture?
- Using two (or even three) dimensional objects comparing two groups or conditions which are actually only being compared on a single dimension
what does one mean by comparing apples and oranges?
- Focusing on one feature of a graph when it is another feature which is actually being depicted
what are the errors in judging simple relationships?
single cell error, man on the street interviews, reasoning from one's own experience, case studies
what's wrong with man on the street interviews?
3 Reporters often solicit opinions or testimonials from isolated people
what is wrong with 3. Reasoning from one’s own experience?
we assume they are right about their experiences
when do we let exceptions disprove the rule?
with negative testamonials, man who..., and historical counter examples
why do most commit single cell errors?
- Because we assume that all relationships are perfects ones (rules without exceptions)
o If every real rule has no exceptions, then we have only 3 possibilities. what are they?
1. Experience and effectiveness always go together 2. Inexperience and effectiveness always go together 3. There’s no relationship at all
what's the vividness effect?
o We’re more likely to pay attention to vivid, testimonial evidence that pallid base rate information
what causes most stereotypes?
the vividness effect
- If vivid info is more likely to lead to the single cell error, what makes evidence especially vivid to us?
o We pay special attention to “man/woman who” evidence when the man/woman involved is our self
when do we mostly make the single cell error?
1. We tend to make the single cell error more when we receive the data piecemeal than all at once. 2. We tend to make the error more when we are simply rying to figure out if there is a rule vs. trying to profit from it.
what kind of people are most likely to fall victim to the single cell error?
people who chronically rely on heuristics
what are single row/column errors?
- These are errors that result from having data in one row or column of the contingency table, but not the other - takes a number of forms including superstitious behavior
how did the single cell error form
- We suggested that single cell error stemmed from assuming that all relationships were perfect
why do we make marginal errors?
We assume that all relationships were perfect
- What would marginals look like if there was a perfect relationship?
they'd be identical
what do we usually do if we already believe that there’s a relationship?
o perceived r>> actual r when we expect that r is large o we tend to overestimate the degree of relationship o and even to see relationships when none exists o for example, the use of invalid projective tests, such as the draw a person test or the Rorschach test
what do we usually do if we have no reason to belive there's a relationship?
o Expected r = 0. Perceived r
what are some things a critical thinker should ask?
is a relationship/lack of one between two variables suggested? what are the variables? is there enough info to conclude if there's a relationship?
what type are most psychological relationships?
probabilisitic (if one variable changes the other seems to change)
what do we mean by chance?
- When something is due to chance, we don’t mean it couldn’t be predicted or explained, we mean it’s caused by factors that we’re unaware of and unable to measure. (something caused registrar to fuck up Steiner’s records.
why do we assume choice always gives us control?
choice often does mean control, even when the outcomes are determined completely by chance.
what's the conjunction fallacy?
judging that probability of (A and B) > min (Prob (A), Prob B)). -Suppose A= Jim passes course. B= Jim is a German major -What are reasonable estimates of all?
what is the belief in a just world?
people get what they deserve/ bad things only happen to bad people.
what's a null hypothesis?
a single source of data for all subjects (no relationship)
what's the alternative hypothesis?
-Different sources of data for different subjects -Depends on how strong relationship is and how large the samples are (there is a relationship)
what is the goal of hypothesis testing?
to choose between the null and alternative hypothesises
what are the procedures for hypothesis testing?
-Draw a sample. -Compute an effect size from your samples -Assume that the null hypothesis is true - find out what would be unusually big effect under this assumption -Even the smallest observed effect will be judged to be statistically significant if the sample size is large enough
alpha is usually 5%. why not set it at 0?
Because the lower alpha, the bigger the chance of making the other kind of mistake (type 2 error).
what is beta and what does it depend on?
probability of type 2 error. -Depends on sample size -Depends on how strong the actual relationship is.
what is a spurious relationship?
when an outside variable causes both A and B.
possiblities when variables are correlated:
-A causes B -B causes A -spurious
How and when can one get from correlation to causation?
One can measure and check the influence of other, potential causal variables. -An inability to think of a plausible “third variable” does not mean that it doesn’t exist. -It is the use of a particular type of method that gives rise to this ambiguity, not whether or not alternative causes are apparent. -Need to establish the relationship using methods which guarantee all three necessary conditions the experimental method.
what is a true experiment?
a method in which the investigator creates differences in one and only on variable, A, and then measures a second variable, B
according to stanovich, what are the 3 conditions of scientific thinking?
1. comparison 2. manipulation 3. control
what is the most difficult condition of scientific thinking? why?
control. -Failure to control or eliminate other possible causes besides A creates rival or alternative hypotheses -There are lots of ways to screw up.
what are the two general approaches to experimentation?
-Approach 1: measure the same people on the dependent variable B both before and after the introduction of the purported causal, independent variable A. (pretest/posttest design) -Approach 2: take two groups of people, vary or manipulate their exposure to levels of the independent variable a, and then measure the dependent variable B. (independent groups design)
what are the rival hypothesises?
history, maturation, pre-testing, and mortality
explain the history hypothesis
any other event besides the independent variable occurring between times 1 and 3 could also cause change
explain the maturation hypothesis
sometimes, just the passage of time results in people changing, even without any external cause
explain the pre-testing hypothesis
sometimes the act of measurement itself at pretest can change behavior at posttest
explain the mortality hypothesis
Sometimes, you lose or gain people between pre and posttests and this change in the composition of the sample itself can produce changes on the dependent variable
- It’s possible the two groups being compared already differed on the dependent variable even before manipulation was introduced. what is this problem called?
selection error
how can one solve the selection error?
1. Better than nothing- match the two groups on all variables known or suspected to affect the dependent variable -Why not perfect? You need to know all relevant variables, and be able to measure them all reliably and validly 2.Good- get a pre-test on the dependent variable to make sure that the two groups don’t differ -Why not perfect? Can sometimes get groups that are both extreme but equal at pretest, and regression leads to posttest differences 3.Best- use random assignment – every person has equal chance to end up in each group or condition being compared
what are treatment confounds?
-Where the experimenter has created differences not just on the independent variable but other variables too -Ideal experiment has no confounding variables, good experiments have no plausible confounding variables -You need to ask not just what has the experimenter tried to manipulate but what are all the variable they actually manipulated
what are experiementer effects?
-The experimenter can sometimes create what type expect to see (knowingly or unknowingly) oIt’s up to the experimenter to decide what is and isn’t an aggressive act oCan be avoided by “double blind” experiments
what are demand characteristics?
-Clues which tell subjects (right or wrong) what the experimenter is studying and what they expect to find. Sometimes this alters subject’s behavior -Can be avoided by making the purpose of the experiment unclear oThis can create new ethical problems of deception (Milgram’s study)
weak manipulation?
if only tiny differences have been created on the ind. Var., then we’re not surprised at no differences on the dep. Var.
- Invalid manipulation of ind. Variable?
unless one has really created differences on ind. Var. don’t expect effect on dep. Var. (even if AB)
Floor/ceiling effects?
if scores on dep. Var. in one condition are already as high as they can go then no manipulation can make them any higher.
Differences between producers and consumers in goals
oProducer- often the primary goal is knowledge oConsumer- primary goal is decision making
Differences between producers and consumers in need to decide
oProducer- can and usually does suspend judgment oConsumer- often doesn’t have the time or can’t afford to suspend judgment
Differences between producers and consumers in time perspective
oProducer- sees research as a long-term, accumulative process oConsumer- usually wants and often has to rely on immediately available evidence
Differences between producers and consumers in ways of evaluating and deciding
oProducer- accepts and exploits convergence principle oConsumer- often exhibits short and long term negativity bias- if there’s something wrong with the study , then it’s of little value.
Diagonal error
If you only have the diagonal boxes you can not figure the rest of them you can have many hits but still no relationship
Floor effect
scores in one condition are as low as they can go:no chance to find lover score in other condition -When someone claims there is no relationship check to make sure there is no floor effect for eaither variable
Gambler's fallacy
- the tendency to see links between events in the past and events in the future when the two are really independent  Ex) winning lottery ticket for January 2nd was 372 so are unlikely to pick the numbers 372 for the next day, even though the lottery numbers are completely independent
sampling error
making inferences about large populations from small samples -the smaller our sample, the bigger the sampling error
proceduers used in hypothesis testing
draw a sample, compute effect size from your samples, assume that the null hypothesis is true
significance/alpha level
the most unusual 5% of possible outcomes identified
illusion of control
Because often choice means control we assume that choice always gives us control, even when the outcomes are determined entirely by chance
motivated misperception of probability and chance
Belief that people get what they deserve, bad things happen to bad people. If raper is a respectable person, people are more likely to think the victim did something wrong to deserve it
possibilities when a & B are correlated
1- A causes B 2- B causes A 3- Some other variable causes both A and B
the experimental method
a method in which the investigation creates differences in one and only variable A and then measures a 2nd variable B. Variable A=independent Variable B= dependent
what is cofunded with the independent variable?
age, race, student/nonstudent status, etc
what part of the study increases experimenter effects?
the use of a subjective rating by a confederate to rate helpfulness
Ceiling Effect
-scores in one condition of an experiment are as high as they can go: no chance to find a higher in the other colums
"no control group error"
-single cell error -example:In 2001 Argentinas currency and economy collapsed-->report said that 40 of 100 deaths of elderly person in Argentina have been suicides--->not saying there is or isnt a relationship, we can not tell if there is from the information given
Superstious Behavior
-Single row/column -Laura and emily drink the night before an exam and they 3.0 so they think the drinking helped them on the exam and now do it for every exam-->But they dont know what they would of got if they did not drink before
Depression and Single cell error
-In a study:depressed were accurate whether or not they had control over pressing a button and a green light going off (response and enviromental relationship) -non-depressed saw more control than there was when they experienced/wanted control and less control than there was when they didnt like the outsomes
what aspect decreases the chances of selection bias the most?
randomly assigning
what aspect increases the chance of demand characteristics
using a very unusual event (someone bleeding in a lab setting)
sometimes the act of measurement itself at pretest can change behavior at posttest (even without any manipulation of the independent variable) *Red flags- when the act of measurement os disruptive or unexpected, when the interval is short, or when the dependent variable is an attitude or opinion
decay or instrumentation
sometimes the measurement instrument changes in a systematic way from pretest to posttest, producing apparent differences when there really are none *Red flag- when humans judge
rival hypothesis (subject mortality)
sometimes you lose (or gain) between pretest and posttest *red flag- when the interval is long, when only group data is available- not individual data, or when the manipulation is especially attractive- causing people to migrate in and vice versa
2 things that make a real & good experiment
manipulation- an experimenter creating differences rather than simply observing naturally occuring ones 2- random assignment- since selection is always a threat
3 things to look out for in an experiment
treatment confounds (when the experimenter has created differences in not only the independent variable, but others) 2- experimenter effects 3- demand characteristics (clues telling subjects what the experiment/expected actions should be)
difference in goals between consumers and producers
producers- often primary knwoedge (basic research) consumers- often decision making
difference in need to decide between consumers and producers
producers- can and usually do suspend judgement consumers- often can't/don't suspend judgement
difference in ways of evaluating and deciding between consumers and producers
producers- accepts and exploits convergence principle consumers- often exhibits short and long term negativity bias- if there is something wrong with the study its of little value
Age and Single cell errors
-Example: students in different grades asked the same question: and in the case that 6 cars all started with an additive, then asked if it helped the yolder the people got the less said that just cause they all started means that the additive helped
When do we make single cell error
-When we recieve the data seperatly than all at once -just trying to figure out the rule vs. trying to profit from it
Moral 2 of cell error
-Knowing all 4 entries is not nessary
Moral 1 of cell error
knowing all 4 entries is suffcient but not necessary, sometimes you can figure out all 4 cells from what is available
Expectations about relationships bias
if we are trying to find if A and B are related, we tend to pay mroe attention to the positive.-->we remember or notice confirmations more than diconfirmations
Underutilization of base rate information
-Ex: Group of 100 men-70 engineers and 30 lawyers -base rate=.7 for engineers and .30 for lawyers, read self-description saying they like puzzles and dont like politics and are to guess if they are a lawyer or an engineer-->people think the probability of the person ebing an engineer is much higher
Ivan Steiner
Story illustraring the power of chance and probability, told he had the 2nd highest GPA and did not get a scholarship...he really had the highest but the registar made a mistake, then he ended up being a psych teacher instead of a lawyer(which is what he always wanted to be)
Representative Heuristic
-Judgements based on feature of events that we are used to happen -Like in a coin flip most people woul dnot think that there would be any long streaks and there should be about a 50/50 distribution
Letting an exception disprove the rule
Negative Testimonials, Man Who Errors, Historical Counterexamples
implications of the vividness effect (making single cell error)
we are more likely to pay attention to vivid, testimonial evidence than pallid base rate information -events that are rare or unusual catch our attention and are more vivid -can help explain how stereotypes are formed
beliefs we hold about members of a social group sometimes can be true (baseball players are tall) but often they are not and almost always overgeneralized *stereotypes form through socialfication
personal evidence and the single cell error
we pay special attention to "man/woman who" evidence when the man/woman involved is our self ex) if you got struck by lightning you attach different meaning to coincidences than when the exact same thing happens to someone else
Summarizes the degree of Asymmetry -Skew=0-->Symmetric distribution -Skew<0-->Tail in negative direction -Skew>0-->Tail in positive direction
Establishing Relationships
from uni-variable to bi-variable -More interesting questions are not about single varriables but if pairs of variables go together
Reasoning from ones own experiences
Ex:My dad says dont be lazy cause if you work hard you will succeed, because he worked hard all his life and he is now the president of his company" -implied=Finanical success is compared to how hard you work
Case studies
-Single cell error -In depth examination of a single instance or event->case -Ex:Lorenzos oil:A disease that degentrated the body, this oil was found to slow down or stop progress -implied-Taking this oil will slow or stop the disease
"Man on the Street" interviews
You ask people who you think will answer the way you want them to, and most of them are just opinions or testimonals.
-Single cell error -Observation is only a single cell and cannot establish a relationship.....what you say might not work for everyone
when are you more liekly to use heiristics to make single cell error?
-when we lack motivation to think carefully about what evidence we have or need -when we lack the ability to think carefully about what evidence we have or need
"one dimensional picture"
using 2 or 3D objects to compare two groups, which are only being compared in 1 dimention
How many Variables does it take to establish a relationship?

Deck Info