ch. 13
Content Validity:It focuses more upon the ____ of the material
It focuses more upon the appropriateness of the material
1.In chi-square testing, you compare how well an ________ breakdown of people over various categories fits some ________ breakdown.
1. Observed; expected
Power for chi-square tests for independence: What would you like to do to get the maximum power?
1. You want as many participants as possible. 2. You want as simple a contingency table (that is, as few categories in each direction) as possible.
• Degrees of freedom
= (Ncategories [number of groups] - 1) = (4 - 1) = 3 *Always skewed to the right*
"Chi-square test for independence" e.g.)
E.g.) Contingency table
E =
Expected frequency
• All of these types of hypothesis testing are ...
General Linear Model.
Any problems regarding this type of reliability?
It is not easy to decide which way to split the halves.
Chi-square test =
It shows how well an observed frequency distribution fits an expected (or predicted) frequency distribution
df=
NCategories-1
O =
Observed frequency
1. The difference between the phi coefficient and Cramer's phi is that: a. The phi coefficient divides by the number of participants b. The Cramer's coefficient divides by the number of participants multiplied by the dfsmaller c. The phi coefficient is used for designs greater than 2x2 d. The Cramer's coefficient is used for 2x2 designs
The Cramer's coefficient divides by the number of participants multiplied by the dfsmaller
Needed sample size for chi-square tests for independence: The more df there are (the more categories in each variable being crossed), the ____ participants you need for the same amount of power.
The more df there are (the more categories in each variable being crossed), the [__m_o_r_e__] participants you need for the same amount of power.
C:
The number of people observed in this cell's column
R:
The number of people observed in this cell's row
N:
The number of people total
The chi-square table: Use a chi-square table to find the cutoff and make a ___
conclusion!
t-test for [_____________] means (Comparison between two groups)
independent
•The Chi-Square Statistic: If the categories are all divided evenly then that would be our ...
null hypothesis
The phi coefficient is the ...
square root of the result of dividing the sample's chi-square by the total number of people in the sample
what makes a question "Good" or "Bad" in terms of alpha?
SPSS will report "alpha if item deleted", which shows how alpha would change if that one question was not on the test. Low "alpha if item deleted" means a question is good because deleting that question would lower the overall alpha. Questions with high "alpha if deleted" tend to have low inter- item correlations (Pearson's r).
interrater Reliability: Different raters who score the essay, interview, observation, or test should come to similar conclusions, if the coding manual is ...
"reliable."
• Power for Chi-Square Test for Independence : For maximal power:
(1) include as many participants as possible (2) simplify your contingency table as much as possible (minimize your categories)
Example for Chi Square Test of Independence: E=
(Column Total x Row Total Percentage) OR (Column Total x Row Total/ Total Sample Size)
• Chi-square test for independence
(MORE THAN ONE categorical/nominal variables) o Ex: OU students who like outdoor activities are more likely to drive Chrysler cars than Ford or GM cars and OU students who like indoor activities are more likely to drive GM cars than Chrysler or Ford cars.
t-test for dependent means: SM (the standard deviation of the distribution of means of difference scores) =
(sign)S2M
Cronbach's alpha ranges from
0 to 1.0.
The Chi-square statistic reflects the overall ________________ between the expected and observed frequencies.
1. lack of fit.
Conducting Chi-Square for Independence; involves ___ categorical variables (nominal variables, each with several categories)
2
1. How many degrees of freedom does a chi-square test have? How do you find the df?
A chi-square test for independence has one df. To find it, use the following formula: df=(N_Columns-1)(N_Rows-1
• With a chi-square test, we can do hypothesis testing for variables whose values are categories (nominal variable) for_____independent and dependent variables.
BOTH o Example: IV: basketball players, baseball players, or football players; o DV: Democrat or Republican
• When you are interested in the effect of extraversion on the level of satisfaction, what test would you use?
Bivariate Regression
chi-square test for independence. Independent and dependent variables:
Categorical variable with two or more values (for both)
1. What kind of variable is used to conduct a chi-square test?
Categorical/nominal variables
χ2=
Chi-square statistic
Hypothesis: Students at OU are more likely to drive a Chrysler car than a Ford car or GM car---- Chi-square test for goodness of fit or Chi-square test for independence?
Chi-square test for goodness of fit
The potential problems using test- retest reliability
E.g.) Some scales are trying to measure the "[__c_h_a_n_g_e__]" so high test-retest reliability would not make sense (we may want to measure something that does vary over time like moods, test scores, etc.) E.g.) You can't use this approach if taking a test once would influence the second taking. E.g.) It is not practical or appropriate, especially for [__c_h_ild_r_e_n__] (assuming their rapid developmental changes) E.g.) You need to do survey at least twice. So, it is costly and we will [___lo_s_e___] some participants.
How to assess a measure's reliability
E.g.) odd-numbered and even-numbered questions E.g.) first half and second half
• When you are interested in the effect of Big Five in the level of satisfaction, what test would you use?
Multiple Regression Analysis
the difference between phi coefficient and Cramer's phi is that instead of dividing by N, you divide by ____
N times the degree of freedom of the smaller side of the contingency table.
First Assumption for chi-square tests: No individual can be counted in ____
No individual can be counted in more than one category or cell
Expected frequency:
Number of people in a category or cell expected if the null hypothesis were true.
Interraterreliability=
NumberOf AgreedCodings /Total NumberOf Codings
5. What is the difference between reliability and validity?
Reliability: Confidence that a result can be replicated. Validity: Confidence that a result shows what it is supposed to show.
t-test for dependent means: S2M (the variance of the distribution of means of difference scores) =
S2/N
5.True/False: You want to observe a poor fit between observed frequency and expected frequency.
True: We want to observe a poor fit between observed frequency and expected frequency.
. What can we do to demonstrate the content validity of my optimism scale?
Use pilot test to assess appropriateness. Employ subject matter experts in the related field.
. Criterion Validity: Concurrent Validity =
When data are recorded simultaneously
7. Which of these statements is correct?
a. A study can be reliable but not valid
Chi-square test for goodness of fit involves a ____
a. Involves a single nominal variable
Chi-square test for independence involves a ____
a. Involves two nominal variables, each with several categories.
To get a cutoff, we use the ...
chi-square distribution
• The chi-square statistic reflects the ___ between the expected and observed frequencies
lack of fit *We want to see a huge difference between observed and expected values in order to reject the null hypothesis
Chi-square statistic = Statistic that reflects the overall [_____________] between the expected and observed frequencies.
lack of fit *should make us excited
Power for chi-square tests for independence: Like all other hypothesis-testing situations, second Assumption for chi-square tests The [ more ] degrees of freedom there are (the [ more ] different categories that are crossed with each other), the____ power there is.
less
0 to +.30 =
little or no association.
• ß shows unique contribution to the ...
outcome variable when the impact of all other predictors is controlled statistically.
_____ correlations (Relationships between 2 or more variables)
pearson's
This type of reliability is the similarity of the ratings between times, instead of ...
raters
• Cohen's conventions for effect size for Cramer's phi depend on the degrees of freedom for the____
smaller side of the table
+.70 to +1.00=
strong positive association
chi-square statistic:
χ2 = Σ (O - E)2/ E
Chi-square tests =...
χ2 test
• Chi-square test for goodness of fit
(ONE categorical/nominal variable) o NO independent variable or predictor variable o Ex: OU students are more likely to drive Chrysler cars than Ford cars.
Cronbach's Alpha if Item Deleted" shows the alpha if that item isn't included in the calculation. -Items with high values (higher than the overall α ) have to be dropped.
-
t-test for independent means: step 1
-Population 1: People that work out -Population 2: People that sat indoors watching Netflix -Research Hypothesis: µ1 ≠ µ2 -Null Hypothesis: µ1 = µ2
Generally, a good measure should have a Cronbach's alpha of at least ____and preferably closer to .90.
.60
The widely-accepted social science cut-off is that alpha should be ___ or higher for a set of items to be considered a scale.
.70
in the chi-square tests, you compare how well an observed breakdown of people over various categories fits some ...
.expected breakdown.
1. After conducting a chi-square test using a sample of 60 participants, a researcher found a chi-statistic of 5.67 with degrees of freedom 3 and using a significance level of 5%. How would this researcher report their results in APA format?
.χ2 (3, N = 60) = 5.67, p <.05
he phi coefficient has a minimum of _ and a maximum of _
0; 1
When the smallest side of the table is 2, the degree of freedom is ____. Thus, the effect sizes given in the table for this situation are the same as for the ordinary phi coefficient.
1
1. To increase power with a chi-square test for independence: a. Increase participants b. Increase the number of categories in the nominal variables c. Decrease the number of categories in the nominal variables d. Both a and c
1. -a, Both a and c
Chi-square test = (steps 1,2, and 3 out of 6)
1. Determine the observed frequencies in each category. 2.. Determine the expected frequencies in each category. 3. in each category, take the observed minds expected frequencies
SPSS
1. Enter the frequencies. 2. Select Analyze, Non-parametric tests, and Chi-square 3. Move your variable to the box of test variable list 4. Make sure "all categories equal" is selected for "expected values" 5. Hit OK! 6. The second table should show observed and expected frequencies, as well as differences between them 7. The third table should show the chi-square value, df, and its significance level.
SPSS: Chi-square test for Independence:
1. Enter the scores into SPSS, and click "Analyze." 2. Click "Descriptive Statistics," and "Crosstabs." 3. Click on one variable. Then click the arrow next to the box labeled "Row(s)." Click on the other variable. Then click the arrow next to the box labeled "Column(s)." 4. Click "Statistics," the box labeled "Chi-square," the box labeled Phi and Cramer's V, and "Continue." 5. Click "Cells," the box labeled "Expected," "Percentages," and "Continue." 6. Click "OK
4.True/False: When you find the cutoff for a chi-square test you need to report (+/-) in front of the cutoff value.
1. False: Chi square distributions are skewed to the right meaning that the distribution is always positive. No need to report (=/-)
Contingency table and differences: steps
1. Find each row's percentage of the total. 2. For each cell, multiply its row's percentage by its column's total.
SPSS provides four tables in the output.
1. It shows the number of individuals for each variable. 2. "Crosstabulation" table gives the contingency table of observed and expected values for the two nominal variables. 3. "Chi-Square Tests" table shows the actual result of the chi- square test for independence. *The first row shows the chi-square value, the degrees of freedom, and the exact significance level. 4. "Symmetric Measures" table shows the effect size measure
Assumptions for Chi-Square Tests:
1. No individual can be counted in more than one category or cell • You cannot use chi-square testing if the scores are based on the same people being tested more than once (repeated measures) because this would create scores that are non-independent • Ex: People who watch both baseball and football = excluded, because they are not independent in the chi-square test) 2. None of the expected frequencies should be less than 5 for the results to be valid • If we have an expected frequency of less than 5, we can use the non-parametric test, Fisher's Exact Test (non-parametric test for chi-square—advanced stats, DO NOT need to know for PSY 2510!)
1. The chi-square statistic is concerned with ____ frequencies and ______ frequencies.
1. The chi-square statistic is concerned with expected frequencies and observed frequencies.
chi- square tests in research articles; In research articles, reports of chi-square tests usually include:
1. The frequencies in each category or cell as well as the df, 2. Total number of participants, and 3. The sample's chi-square and significance level
1. What is the null hypothesis for a chi-square test?
1. The null hypothesis for a chi-square test will state that participants are equally likely to pick any of the available choices. In other words, the expected frequency will be the same for observed frequency for all cells.
The shape of the chi-square distribution depends upon the _________________.
1. degrees of freedom
2. List the ways you can check the reliability of the scale shown below.
2. List the ways you can check the reliability of the scale shown below. • Inter-rater reliability • Test-retest reliability • Split-half reliability • Cronbach's alpha (α)
Chi-square test = (steps 4,5,6 out of 6)
4. square each of thee differences *to get rid of the direction of the difference, since the interest is not only in how much there is. 5. divide each squared difference by the expected frequency for its category 6. add the results of step 5 for all categories.
Power for chi-square tests for independence (.05 significance level) E.g.) 2 x 4 study (df = 3) of 50 people with an expected medium effect size. Power will be...
40. If the research hypothesis is true, and there is a true medium effect size, there is about a 40% chance that the study will come out significant.
Cramer's phi =
= An extension of the ordinary phi coefficient that we can use for contingency tables larger than 2 x 2
The [__________] of the ratings between raters
= Interrater Reliability ex- essay - 2, 2, and essay e 1,1
cronbach's alpha= The average of the split-half correlations from all possible splits into ...
= The average of the split-half correlations from all possible splits into halves of the items on the test.
contingency table
= Two dimensional chart showing frequencies in each combination of categories of two nominal variables.
Chi-square test for goodness of fit:
A single nominal variable (with 3, 4, 5...categories).
1. What are the two assumptions of Chi-Square Test?
Assumptions for chi-square tests are that (1) scores must not be based on the same people being tested more than once and (2) expected frequencies should be greater than 5 for results to be valid
6. Multiple choice: The degrees of freedom for Chi-Square Test is dependent upon ____________ A. Sample size only B. Number of categories only C. Both samples size and number of categories
B. Number of categories only
◆ Hypothesis: OU students who like outdoor activities are more likely to drive a Chrysler car, whereas OU students who like indoor activities are more likely to drive a GM car. Chi-square test for goodness of fit or Chi-square test for independence?
Chi-square test for independence
6. What are the different types of validity, and what are the methods or statistics associated with them?
Content validity = Peer review, expert opinion, preexisting definition in the literature Construct validity = Pearson's correlation coefficient Predictive validity = Regression analysis External validity = Replication
8. What is the difference between content validity and criterion validity?
Content validity assesses the appropriateness of the measure often through peer review, whereas criterion validity measures appropriateness through comparison to objective measures.
Convergent Validity: Between (1) my optimism scale AND (2) somebody's pessimism scale
Convergent Validity: Negative, strong, ideally p<.05
• For a larger design than 2x2, use ___
Cramer's phi: an extension of the ordinary phi coefficient
To resolve these issues, we use
Cronbach's alpha (a)
3. What is the most widely used measure of reliability?
Cronbach's alpha (α)
4. Cronbach's alpha should be at least ... for a set of items to be considered a reliable scale.
Cronbach's alpha should be at least .70 for a set of items to be considered a reliable scale.
1. Content Validity
Degree to which instruments cover aspects or dimensions of underlying construct.
. Construct Validity
Degree to which instruments related with other constructs For example, Does my optimism scale measure the participant's personality or temporary positive mood? Does my optimism scale measure the participant's level of optimism or level of non-pessimism? To answer these questions, first we need to collect data!
construct validity: Between (1) my optimism scale AND (2) positive mood scale:
Discriminant Validity: Positive, but weak, maybe n.s
E:
Expected frequency for a particular cell
If the inventory has a high reliability, it must be an example of important psychology research. true or false
False! Reliability is different from importance. It is possible to produce findings that are highly reliable but not important!
Interrater Reliability
If there are two or more raters of each participant's behavior or material, then we can get the interrater reliability!
SPSS Input
In Variable view: • 2 categorical variables: o 1st variable: sports ♣ click values (ex: 1 = football, 2 = baseball) ♣ change measure to nominal o 2nd variable: drink ♣ click values (1 = beer, 2 = pop, 3 = water) ♣ change measure to nominal In Data view: • Enter scores into SPSS (2 columns, 40 rows) 1. Click "Analyze" 2. "Descriptive Statistics" "Crosstabs" 3. Select one variable then click the arrow next to the box labeled "Row(s)" (ex: sports) 4. Select the other variable and click the arrow next to the box labeled "Column(s)" (ex: drinks) 5. Click "Statistics" check boxes "Chi-Square" and "Phi and Cramer's V" a. Phi and Cramer's V is used for effect size 6. Select "Cells" check boxes labeled "Expected" and "percentages" (row and coulmn) "Continue" and "OK"
How to assess a measure's reliability
Instead of measuring twice, researchers may split a test into 2 or more parts to determine if answers to one part are similar to answers on another part of the test.
. Construct Validity: It has to do with the logic of items which ____
It has to do with the logic of items which comprise measures of concepts.
Chi-square test for goodness of fit": ...
It involves a single nominal variable (with 2, 3, 4, 5...categories).
"Chi-square test for independence": ....
It involves two nominal variables, each with several categories. E.g.) Contingency table
Chi-square test:
It shows how well an observed frequency distribution fits an expected (or predicted) frequency distribution
2nd Assumption for chi-square tests (could be controversial :)
None of the expected frequencies should be less than 5, for the results. of a chi-square test to be valid.
o Observed frequency:
Number of individuals actually found in study in a particular category (the numbers that are actually observed)
Observed frequency:
Number of individuals actually found in the study to be in a category or cell.
o Expected frequency:
Number of people in a category expected if the null hypothesis were true~everyone evenly distributed (the null hypothesis would be that everyone is equally distributed) (100/N; N being number of "cars")
Expected frequency:
Number of people in a category or cell expected if the null hypothesis were true.
_____ ANOVA (Comparison among three or more groups)
One-way
If you are interested in the differences in the level of satisfaction among people who took PSY2510, those in PSY2500, and those in PSY1000, what statistics would you use?
One-way ANOVA
Reliability
Our confidence that a given finding can be reproduced again and again
Validity
Our confidence that a given finding shows what it is supposed to show.
validity: Now which statistics would you like to use?
Pearson's correlation coefficient
t-test for dependent means: populations
Population 1: People who go to Applebees Population 2: People who show no difference
what stats would you like to use?
Regression Analysis
1. What is the difference between reliability and validity?
Reliability is the confidence that a given finding can be consistently reproduced and validity is the confidence that a given finding shows what it is supposed to show.
t-test for dependent means: research vs null
Research hypothesis: Population 1's mean difference score is different from the Population 2's mean difference score (which is zero!). (μ1 ≠ μ2) ◆ Null hypothesis: Population 1's mean difference score is the same as the Population 2's. (μ1 = μ2), which is zero!
Example for Chi Square Test of Independence: Row Total=
Row Total= (Row Total / Total Sample Size)
t-test for dependent means: S2 (the estimated population variance of difference scores) =
SS/df
If you are interested in the effect of score of the final exam in the level of satisfaction regardless of class, what statistics would you use?
Simple bivariate regression analysis
Figuring chi-square
These steps can also be stated as a formula: E: Expected frequency for a particular cell R: The number of people observed in this cell's row N: The number of people total C: The number of people observed in this cell's column 𝐸=𝑅/N ×𝐶=𝑅×𝐶/N
validity: we can collect data from..?
We can collect data from (1) my optimism scale AND (2) positive mood scale. We can collect data from (1) my optimism scale AND (2) somebody else's pessimism scale.
First Assumption for chi-square tests: We can't use the chi-square tests if the number in each cell are based on the ___ ___ being tested more than once.
We can't use the chi-square tests if the number in each cell are based on the same people being tested more than once. *one key assumption is that each frequency must not ave any special relation to any other frequencies.
. Criterion Validity: Predictive Validity =
When data from the new instrument are used to predict observations at a later point in time
Effect size for chi-square tests for independence:
With a 2 X 2 contingency table, the measure of association of the two nominal variables is called the phi coefficient (φ ) Here is the formula: ..
Cronbach's alpha (a) The idea is:
You can divide the measure into halves in all possible ways and figure the correlation using each division, then average all these split-half correlations.
1. In a contingency table, the number in each cell is a. a mean. b. A number of people
b. A number of people
_____Regression (Relationships from 1 predictor to outcome)
bivariate
Contingency table and differences: We have to figure differences between observed and expected for each combination of _____ (that is, for each cell of the contingency table).
categories
With chi-square test, we can do hypothesis testing for variables whose values are ....
categories (nominal variables).
With a chi-square test, we can do hypothesis testing for the dependent variables whose values are ____
categories; nominal in SPSS.
E is the expected frequency, which is based on what you would expect if there were equal numbers in each ...
category.
To decide whether to reject the null hypothesis, you can compare the ____
chi-square cutoff and the chi-square of the actual sample.
If a test has a low variance, the scores for the group are close together. Unless the participants truly are ..
close in ability, the test is not useful
Most cases, IV and DV can be determined by a research design, but there are times that can go either way. That is nothing to do with which variable goes to ____
column or row.
. Criterion Validity Degree to which an instrument measures what it claims to measure through ____
comparison to objective criteria.
the denominator divides by "N" multiplied by the degrees of freedom on the smaller side of the ___
contingency table
Chi-Square Tests: Each person can contribute to only one cell of the ___ ___. This is because we use % in each row to get the expected frequencies and the total of percentage in each row needs to be 100% to meet the conditions in which people are equally divided into each cell. Due to this reason, we cannot use a chi-square test on a repeated- measures design.
contingency table.
Next, we need to know whether the chi-square statistic we have just figured is a [__b_i_g_g_e_r__] mismatch than we would expect by chance, so, basically we need the ..
cutoff!
1. To increase power with a chi-square test for independence: a. Increase participants b. Increase the number of categories in the nominal variables c. Decrease the number of categories in the nominal variables d. Both a and c
d. Both a and c
Fisher's Exact Test or calculating a Bayes factor is an option when the ...
data violate the assumption.
t-test for ____means (Comparison between before and after)
dependent
• T test, ANOVA, correlation and regression all require that the ___ is equal interval (scale), which allows us to get the mean.
dependent variable
• Chi-Square (Goodness of Fit); Table is relatively short because Chi-Square does not involve the number of participants to ____
determine the cut-off score like the T-table or F-table do.
The shape of the chi-square distribution depends on the ...
df
The shape of the chi-square distribution depends on the ___
df. df = NCategories - 1
Chi-Square Tests: Same people cannot be in ___ ___
different cells.
The phi coefficient is the _____ for a chi square test for independence for square test for a 2 X 2 contingency table
effect size
t test, ANOVA, correlation, and regression, all required that the dependent variable has scores that are _______ in SPSS).
equal-interval scale
External Validity Degree to which the experiment's results can be ____
extrapolated to a real-life population.
_____ANOVA (Comparison between 2 or more dimensions)
factorial
1. True or False: Chi-square test = X test
false
We are going to focus on how many people or observations fall into different categories, the____ E.g.) Four car types among OU students (n = 100) Of course, this is a fictional data set
frequency
• We are going to focus on how many people or observations fall into different categories, the ____
frequency o DO NOT use M because you are not finding the mean. Use n because you are finding the frequency.
Cronbach's alpha:High alpha is caused by [ ] variance. Q. high or low?
high
If the scale has a good reliability, this should be a _____ [ correlation.
high; positive *this correlation= split-half reliability
Cronbach's alpha: The [ ] between the items, the higher the reliability
higher correlations
• The goal of chi-square testing is to show ...
how well an observed frequency distribution fits an expected (or predicted) frequency (so, you want to see poor fit, instead of good fit, to reject your null hypothesis).
External validity addresses the question of whether the ..
independent-dependent variable relationship found in the experiment would hold in other contexts to which the results are being generalized.
Chi-square test for goodness of fit
independent: no, dependent: Categorical variable with two or more values
What's the potential disadvantage of using the inter-rater reliability?
individual differences among the raters. Systematic training may be necessary. Clear instruction will be needed. Sometimes, we do not have two or more raters, and yet, we need to demonstrate the reliability of the scale or manual.
If a question is "bad", this means it is not conforming with the rest of the test to measure the same basic factor (or construct). Then, the particular question is not "[__________________]" internally consisted
internally consistent
Interrater Reliability: We prepare the coding manual that defines optimist. Two researchers will code yes, they are optimistic (1) or No they are not (2), by using our ....
manual.
Power for chi-square tests for independence: Like all other hypothesis-testing situations, First Assumption for chi-square tests 1. The ___ participants there are in the study, the more power there will be
more
3. A chi-square test for independence uses (one/more than one) categorical variable, while the chi-square test for goodness of fit uses (one/more than one) categorical variable.
more than one; one
______Regression (Relationships from 2 or more predictors to outcome)
multiple
Suppose these 100 OU students had been equally likely to drive each type of cars.what kind of hypothesis?
null hypotheiss If that were the situation, then [__________] of OU students are expected to drive Chrysler, Ford, GM, or other car.
The chi-square table: Now, we want to know if this value exceeds the cutoff to reject the....(found in table)
null hypothesis.
o Chi-Square (Goodness of Fit): Difference between chi-square table and other tables: Sample size doesn't matter; the focus is the ___
number of categories in the chi-square table
The number in each cell is the _____ of people that have a combination of variables.
number or frequencies
• There are certain times when other types of tests (t tests, correlation/regression, ANOVA) cannot be used. why?
o The above-mentioned tests all require that the dependent variable is equal interval ("Scale" measurement in SPSS). These tests sometimes use categorical variables but as an independent variable, rather than a dependent variable. o Often, you use means in the formulation of these tests because our dependent variable was an equal interval variable.
• When you want to test the differences in level of satisfaction among three or more groups, what test would you use? one-way ANOVA o To run a two-way factorial ANOVA, add ...
one more dimension (ex: course and time)
• When you want to test the differences in level of satisfaction among three or more groups, what test would you use?
one-way ANOVA
High variance means you have a wide spread of scores, which means ...
participants are easier to differentiate
o Cutoff is always skewed to the right in chi-square. Therefore, the cutoff is always__
positive.
Power for Chi-Square Test for Independence: • The more participants in a study the more ____ there will be
power
The chi-square table: As always, we want to know the cutoff for a chi-square to be extreme enough to ...
reject the null hypothesis.
cronbach's alpha = The most widely used measure of...
reliability.
Are the findings true for other populations/contexts?
replication
The distributions are all skewed to the...
right
Cohen's conventions for the phi coefficient are as follows:
small effect size φ = .10 Medium effect size φ = .30 large effect size φ = .50
cohen's conventions for effect size for Cramer's phi depend on the degrees of freedom for the___
smaller side of the table
• Power for Chi-Square Test for Independence - To find the number of participants for 80% power, given a desired effect size, use Table 13-10 in the Appendix of the textbook o Example: 40% power means that if the research hypothesis is true, and there is a medium effect size, there is about an 80% chance that the ____
study will come out significant.
• When you want to test the differences in level of satisfaction between two groups, what test would you use?
t test for independent means
If you are interested in the differences in the level of satisfaction between people who took PSY2510 and people who took PSY2500, what statistics would you use?
t-test for independent means
The expected frequencies are just a different way of dividing up the column total; thus ...
the total should be the same
True or False: With Chi-square tests, we use nominal variables.
true
Test-retest reliability, We can use the measure with the same group of people ...
twice.
• Conducting Chi-Square for Independence; A contingency table is a two-dimensional chart showing frequencies in each combination of categories of ...
two (or more) nominal variables
Chi-square test for independence:
two nominal variables, each with several categories.
+.30 to +.70 =
weak positive association.
t-test for independent means: step 2
µm: 0
t-test for dependent means: μM = [____]
μM = 0
Chi-Square Tests in Research Articles
• As an intro sentence, state the results with the APA formatting given above • Then, include the frequencies or percentages in each category or cell, as well as df as a way of further describing the results o Example: The participants were categorized into two groups based on what type of drink (Beer, Pop, Water) they preferred and then they identified what type of sport they like to watch (Football or Baseball) • THEN: Break down the conclusion if the results are significant or insignificant (See slide titled Chi-Square tests in research articles for example) χ2(df, N = ...) = chi statistic, p < .05
Power for Chi-Square Test for Independence : • The more degrees of freedom there are (the more different categories are cross with each other) the ____ power there will be:
• The more degrees of freedom there are (the more different categories are cross with each other) the less power there will be: o It can sometimes be good to remove categories, or collapse/broaden a category to include several of the original sub-categories