Quiz #5

A Type 2 error means that a researcher has

falsely concluded that a treatment has no effect

Under what circumstances is a very small treatment effect still likely to be statistically significant?

if the sample size is large and the sample variance is small

For data that consist of two or more groups of scores with each score a numerical value measuring the same variable, the most common statistics are those that measure and evaluate the mean differences between groups

true

One concern with measuring split-half reliability is that many different correlations are possible depending on how you decide to split the test items in half

true

Without some correction, the percentage of agreement between two observers will tend to overestimate the true level of inter-rater reliability

true

Interpreting

difference b/w means; which mean is lager?; is there a change or growth in variables?

Matched Group

each participant in group 1 can be paired with a participant in group 2; groups are either directly measured or treatment 1 is given to group 1 and treatment 2 (or not treatment) is given to group 2

the __________ of factor A _________ depending on the _________ of factor B

effect; changes; level

A scatter plot shows data points that are widely scattered around a line that slopes down to the right. Which value would be closest to the correlation for these data??

-.40

Which of the following is an impossible value for FObt (fecal occult blood test)?

-2.97

Which statistics was developed to correct problems with inter-rater reliability??

Cohen's kappa

A Type 1 error means that the hypothesis test has failed to detect a real treatment effect

False

When a research report demonstrates a significant treatment effect at the .05 level, you can be more confident that the effect is real than if the researcher had reported a significant effect with an alpha level of only .01

False

With some correction, such as the Spearman-Brown formula, a correlation measuring split-half reliability will tend to overestimate the reliability for the full test

False

It is always possible that patterns observed in sample data do not accurately represent patterns in the population

True

A research reports a significant correlation using an alpha level of .01. The probability that the researcher is making a Type 1 error is p < .01

True

A researcher reports a significant correlation using an alpha level of .01. In this situation, the probability that the researcher is making a

Type 1 error is p < .01

For a research study comparing the average problem-solving ability for 5, 6 and 7 year old boys, the appropriate hypothesis test is

a single factor analysis of variance

Post Hoc

allows you to determine which groups differ after you reject Ho

What type of graph is used for interactions?

bar graph

What test is used to analyze the difference between the means of two independent samples?

between-groups t-test

For a study comparing the number of registered Democrats who approve or disapprove of the presidents' economic policy with the corresponding numbers for registered Republicans, the appropriate hypothesis test is

chi-square test for independence

Using ANOVA allows

consideration of more than 2 groups, control over Type 1 error rate, and researcher to answer much more complicated research questions

In correlational studies, a relation is typically measured and described by the numerical value obtained for a

correlation coefficient

ANOVA is used for

correlational, quasi-experimental and experimental designs

One category of data is obtained by measuring two different variables for each individual in a single group of participants. What statistics are most commonly used for data in this category?

correlations or chi-square tests

Matched Groups reduce

data variability, Type 2 error and error in outcome variable

Using an independent groups design and a t test analysis you are worried you made a Type 2 error. Which of the following might you do in the next experiment to reduce the probability of a Type 2 error?

increase N, change to a within-groups design and try to reduce the variance by exerting better experimental control

As the magnitude of r increases, the proportion of variability of Y that is accounted for by X

increases

What is true about the F disribution?

it has no negative values, it is positively skewed and there are a family of F curves uniquely determined by df (numerator) and df (denominator)

Independent Groups

no consistent basis on which the groups are matched;

Experimentwise Type 1 error

pairwise tests involving groups should retain the null hypothesis --> a series of these tests is then subject to this error

What would the scatter plot show for the data that produce a Pearson correlation of r = .88?

points clustered close to a line that slopes up to the right

Interactions in Factorial ANOVA

refer to particular patterns of effects in the outcome involving 2 or more factors

Factorial ANOVA

referred to by # of factors or design or factor names (2 x 3; 2 x 3 x 3; two way; 2 (sex) x 3(treatment) x 3(BMI)

Main Effect in Factorial ANOVA

refers to the effect of any given factor ignoring levels of the other factor(s)

For a within-subjects study evaluating the mean difference between two treatment conditions the appropriate hypothesis test is

repeated-measures t test

Statistical inference

retain or reject the null; look at p level

If a sample is selected from a population, the sample mean is essentially guaranteed to be different from the population mean. In general, the naturally occurring difference between a sample statistic and the corresponding population paramater is called

sampling error

Which statistic estimates the average of all the possible split-half correlations but can only be used when each question has only two response alternatives?

the Kuder-Richardson formula 20

In a hypothesis test, standard error measures

the amount of difference expected just by chance

Interaction

the effect of factor A (on the DV) changes depending on the level of factor B

What is a basic assumption for a chi-square hypothesis test?

the observations must be independent

If its assumptions are met, the ANOVA technique is appropriate when

the researcher varies two or more factors, several levels of a single independent variable are compared and when three or more groups are involved

What problem is Cohen's kappa intended to correct?

the simple percentage of agreement tends to overestimate the true level of agreement between two observers

Values such as Cohen's d or r2 (the percentage of a variance accounted for) are used to measure

the size of a treatment effect or the strength of a correlation

Values such as Cohen's d or r2 (the percentage of variance for) are used to measure

the size of a treatment effect or the strength of a correlation

The F ratio increases as

the variability between means increases relative to the variability within groups

When deciding which measure of correlation to employ with a specific set of data, you should consider

whether the relation is linear or nonlinear and the type of scale of measurement for each variable