Chapter 15 B Flashcards
Goal of inferential statistics
to use samples as basis for reaching general conclusion about the POPULATION from which the samples were collected
What stands in the way of inferential statistics’ goal?
sampling error
what is sampling error?
error due to discrepancy between the info available from a sample to the true situation that exists in the population. A NATURALLY EXISTING difference between a sample statistic and the coinciding population parameter.
A sample does not portray a perfectly accurate picture of the population
What is a hypothesis test?
a systematic and statistical procedure that determines whether the sample data provides convincing evidence to support the original research hypothesis. Ensures that the difference in sample means is CAUSED BY TREATMENT and NOT BY CHANCE
Hypothesis tests help ensure ____ validity
internal validity
Goal of a hypothesis
to make statements about a population using only data collected from a sample, and to rule out chance as a threat to internal validity or plausible explanation for treatment results.
5 basic elements of a hypothesis
1) the null hypothesis
2) the sample statistic
3) the standard error
4) the test statistic
5) the alpha level
the ______ is the hypothesis that is tested for rejection, and is a STATEMENT about the population indicating INITIAL BELIEF or status quo
the null hypothesis
the null hypothesis assumes there is __(yes/no)__ effect on change or relationship, or that there is ___ diffrerence between treatments.
the null hypothesis assumes there is no effect on change or relationship between the 2 variables, and there is no difference between treatments. IF there is a change between treatments, the null hypothesis relates this back to just random chance.
The alternative hypothesis is the ___ of the null hypothesis
the complement
Non directional vs directional alternative hypothesis
non-directional: “a difference” “a change”
directional: “increase” “decrease”
the goal of a study is to ____the null hypothesis and ___ the alternative hypothesis
the goal of a study is to reject the null hypothesis and accept the alternative hypothesis.
What is a sample statistic?
a basic element of a hypothesis test. A value calculated from the sample data that allows the researcher to make inferences about a given population (ex/ sample mean, sample deviation)
What is the standard error?
a basic element of a hypothesis test. the average size of the sample error/ the standard distance between a sampling statistic and the CORRESPONDING PARAMETER IN THE POPULATION.
what is the test statistic?
compares the samples statistic to the null hypothesis, using standard error as a baseline.
(sample statistic- parameter of the null hypothesis)/standard error
is the test statistic is near ___, this is an indication of NO treatment effect
if the test statistic is near 1. that means that the treatment sample statistic is similar to the chance results.
is real treatment occurs, the test statistic should be _____, indicating that actual results are different than chance results and thus the null hypothesis should be rejected
the test statistic should be bigger than 1
What is the alpha level?
The level of significance- the likely hood that the results would have occurred by chance.
if p<0.05, then
then there is less than 5% chance that the results obtained by the experiment is due to chance.
What’re the two main types of errors in hypothesis testing?
a type 1 error or type 2 error
Type 1 error
a REJECTION ERROR/ FALSE ALARM. You accidentally accepted the alternative hypothesis but there was actually no difference in your treatment conditions or effect on the population.
Type 2 error:
an ACCEPTANCE ERROR/”MISS.” The sample sata does not appear to show significance, but there is actually an effect in the population. You should have rejected the null hypothesis.
Factors that Affect the Outcomes of the Hypothesis test
1) the number of scores in the sample
2) the size of variance
T/F the larger the sample size, the more representative of the population
true
T/F the sample statistic differences or correlations that are obtained with HIGH VARIANCES are less convincing than if the same results were obtained with low variance
true, you want low variance if you want your argument to be convincing.
What is effect size and how is it measured?
an effect size provides information about the absolute SIZE of the TREATMENT EFFECT that is NOT INFLUENCED by outside factors, such as sample size. Measured using Cohens D, or as a measure of % variance.
D= sample mean difference/standard sample deviation
r^2 (when the % variance is measured for t tests
n^2 (when % variance is measured for ANOVA) or if the experiment utilizes more than one treatment mean.
A larger Cohen’s d, the ___ difference in treatment conditions
the larger the difference in treatment conditions (the largest effect)
formula for % variance
r^2= (t^2)/ (t^2+df), the LARGER the r^2 value, the LARGER the effect
Confidence Interval
a range of values created on either side of a sample statistic (ex/ Mean, correlation) that estimates the unknown population parameter is located within an interval around a known sample statistic.
A narrow confidence interval is _____, but less confident
more precise, but less confident that the interval contains the population parameter.
a larger error= a ____ confidence interval
a larger error=a wider confidence interval
Confidence Intervals are based on 3 factors, what are they?
1) the value of the statistic (mean, correlation)
2) the standard error for the sample statistic (larger error=wider confidence interval)
3) chosen level of confidence (increasing confidence=increasing interval)
a larger sample population would lead to a ____ in the width of a confidence interval
lead to a decrease in the width of a confidence interbal. Larger samples lead to smaller standard errors, which increase the likelihood of finding a significant results and decrease the width of CI.
Are confidence intervals a good substitute for Cohens D or R^2 value?
no, becaus ethe CI are influenced by sample sizes. They do not provide an unqualified measure of ABSOLUTE EFFECT SIZE.
if r^2 for percentage variance is 0.0441, this means:
that 4.41% of the variance of one variable is accounted for by the OTHER variable (In a relationship between 2 VARIABLES) (if r^2 is greater than 0.25, then there is a large effect that the variance seen is caused by the other variable.
example of measuring the effect size using percentage of variance
for example, studying the effects on a cholersterol drug on patients. it is expected that the patients who are givin the drug will have lower levels of cholersterol. When seeing the data, it is evident that there is variance, there are low scores mixed with high scores. however, some of these low scores can be attributed to the fact that there is a relationship between the drug and the levels of cholesterol in each participant. by knowing which patient is in the treatment group and by calculating the r^2 value, the researchers can determine how much variance in the samples is caused by the relationship between the drug and the levels of cholesterol
